APPENDICES

Auckland UniServices Limited

II

The University of Auckland in collaboration with:

Canterbury District Health Board, Hutt Valley District Health Board and

Waikato District Health Board

Wellington Masonic Villages Trust, Pegasus Health and

Presbyterian Support Northern

III

Research team Investigators

Named Investigator Dr Matthew Parsons Department GERAC, The School of Nursing Organisation The University of Auckland PO Box/Street number Private Bag 92019 Suburb Grafton City Auckland Telephone 09 373 7599, ext 83033 Fax 09 367 7158 Email m.parsons@auckland.ac.nz

Named Investigator Professor Craig Anderson Department Formerly of, The Clinical Trials Research Unit University/Organisation The University of Auckland

Named Investigator Hugh Senior Department The Clinical Trials Research Unit University/Organisation The University of Auckland

Named Investigator Xenia Chen Department The Clinical Trials Research Unit University/Organisation The University of Auckland

Named Investigator Associate Professor Ngaire Kerse Department Department of General Practice and Primary Care University/Organisation The University of Auckland

Named Investigator Diane Jorgensen Department School of Population Health University/Organisation The University of Auckland

Named Investigator Dr Paul Brown Department Health Systems Department University/Organisation The University of Auckland

Named Investigator Stephen Jacobs Department The School of Medicine University/Organisation The University of Auckland

Named Investigator Stephen Vanderhoorn Department The Clinical Trials Research Unit University/Organisation The University of Auckland

Named Investigator Associate Professor Judy Kilpatrick Department The School of Nursing University/Organisation The University of Auckland

Research associates

Maria Donaldson, Kylie Wright, Liz Bennett, Danielle Lamb, Diane Jorgensen, Lorraine Ritchie

IV

Acknowledgements

This project would not have been possible without the considerable support of the following individuals1 and organisations: Hon Ruth Dyson (Minister for Health of Older People, Associate Minister of Health) Deb Kerry, Judy Kavanagh, Stephen Lungley, Allan Bruce, Sally Stewart, Judy Glackin, Stephen Jacobs and Pam Fletcher (Ministry of Health) Christchurch

Kelly Maw (Pegasus Health) Carole Kerr, Gill Coe and Dr Nigel Miller (Canterbury District Health Board) Dr Sally Keeling (Christchurch Medical School, The University of Otago) Lower Hutt

Karina Kwai, Nicola Turner and Peter Glensor (Hutt Valley District Health Board) Fiona Begg and John Buttler (Sigma Trust) Warwick Dunn and Diane Jorgensen (Wellington Masonic Villages Trust) Hamilton

Dr Jan White, Jeff Bennett, Maree Pierce, Rob Hillman and Dr Sareth Fonseka (Waikato District Health Board) Jan White (Disability Support Link, Waikato District Health Board) Julie Martin and John Baird (Presbyterian Support Northern) The University of Auckland

Frank Tracey, Anjanette Baker, James Chal, Persees Anitia and Mark Coulstan (UniServices) Associate Professor Phillippa Poole, Kathy Peri, Phil Wood, Mark Jones and Mike Law (Faculty of Medical and Health Sciences) In particular, we would like to express our warmest appreciation to the Needs Assessment Service Coordination organisations and in particular the managers of those services involved, namely: Carole Kerr, Fiona Begg (latterly John Buttler) and Jan White.

1 ASPIRE commenced in 2003 and since that time, several individuals who have supported the project have relocated and no longer hold the positions stated here

V

Dedication

This project has involved the lives of many older people, their families and friends. They chose to participate in this project freely and in the full knowledge that there was no personal gain from involvement, but did so in order to improve the lives of countless older people in the generations to follow.

VI

Disclaimers The views expressed in this report are those of the authors and should not be taken to

represent the views or policy of The Ministry of Health or The Government. Although all reasonable steps have been taken to ensure the accuracy of the information, no

responsibility is accepted for the reliance by any person on any information contained in this report.

Competing interests The authors of this work are not aware of any competing interests that may impact on any

aspect of work.

Funding The ASPIRE trial was funded by the Ministry of Health

Cover note This report is presented in two parts, the first deals with the clinical effectiveness of the

Ageing-in-Place initiatives and the second focuses on the costs attributed to the initiatives. This current piece of work focuses on the former.

VII

Executive Summary

Introduction People are living longer. That is good news but, as baby boomers reach old age, the changes in demographics pose several challenges for society. Already, it is recognised that the majority of older people prefer to remain living at home rather than spend protracted periods of time in a residential facility. Indeed, it is anticipated that this trend will continue and strengthen as proceeding cohorts reach old age. To create the conditions to facilitate this concept in an affordable manner requires careful thought and skilled implementation.

The concept of ageing-in-place is already well integrated into several government strategies and concerns the ability for people to “make choices in later life about where to live, and receive the support to do so” (Positive Ageing Strategy, 2001). Despite this broad categorisation, ageing-in-place invariably refers to the ability of older people to remain dwelling in the community, including within retirement villages. Moreover, residential care in the form of either rest homes or private hospitals is specifically excluded.

Unfortunately, disability in later life is very much a reality for many and the choice around whether to remain at home or not is often down to the availability of appropriate services. Programmes that facilitate older people to age-in-place are in their infancy and this is despite the recent considerable increase in home-based funding. For several years, District Health Boards and The Ministry of Health have been supporting the development of ageing-in-place services and this project, ASPIRE (Assessment of Services Promoting Independence and Recovery in Elders) is an evaluation of three of the more significant programmes: Coordination of Services for Elderly (COSE) in Christchurch; The Promoting Independence Programme (PIP) in Lower Hutt and; Community FIRST (Flexible Integrated Restorative Support Team) in Hamilton.

VIII

The ageing-in-place initiatives COSE was established in 2000 through a collaboration of Canterbury DHB and Pegasus Health and is a community-based needs assessment and service co-ordination initiative. It was established with the aim of avoiding duplication in service provision. A key worker (COSE) is based in primary health care and is assigned to several general practice (GP) teams, though works independently of the practices. The model allows the COSE worker to identify resources and opportunities within communities, both funded and non-funded. This offers older people a greater choice of service support, facilitating their remaining safely in the community as long as they wish to. There is a strong evidence base for COSE, which is an example of case management. Studies that have evaluated community-based case management of older people have been conducted in several countries such as the United States, the United Kingdom and Italy. They have been found to reduce hospital admissions, the length of hospital stay, mortality, emergency department visits and admission to long-term facilities as well as costs of care. The COSE worker undertakes comprehensive assessments of the older person and liaises with the GPs and practice nurses ensuring that there is recognition of and a quick response to any change in an older person’s circumstances, thereby allowing the level of care that is required for safe continuous ageing-in-place. As a component of providing an appropriate level of care, the COSE worker co-ordinates the appropriate community services, informal networks and medical care based on assessed need and GP liaison. In essence, COSE is an evolution of the current Needs Assessment Service Coordination (NASC) service that is operating across New Zealand. NASC is invariably hospital based with an extensive outreach component and provides an assessment and service brokerage facility for people requiring access to disability services.

The Promoting Independence Programme (PIP) is for older people who would not be able to maximise their potential for recovery within the average hospital stay. The initiative was developed by The Wellington Masonic Villages Trust in collaboration with Hutt Valley District Health Board and is operates currently both in Woburn, Hutt Valley as well as Horowhenua, Levin. Referrals to the programme are made through Medical Consultants (Private and Public), General Practitioners, Sigma (NASC) or other like referral agencies. A key worker is assigned to each older person and their role is to initiate and co-ordinate that person’s pathway through the rehabilitation process. The team includes Rehabilitation

IX

Co-ordinator who has overall responsibility for the team and is the first point of contact for client, their family / whanau and outside agencies. The team consists of Registered Nurses, Occupational Therapists, Physiotherapists, a Speech and language therapist, Social Worker, Podiatrist, Dietitian, Kaiawhina, designated Caregivers, Rehabilitation Assistants and a Rehabilitation Specialist / Geriatrician. Older people are able to receive up to 12 weeks of facility based rehabilitation in the Promoting Independence Programme (not offered through the Masonic facility) if they are assessed as having high needs and are at risk of residential care or long-term hospitalisation. Clients who are assessed as having high or very high needs but are able to receive rehabilitation services in the community, or clients who have been discharged to the community from a residential facility may receive a monitored amount of input up to a maximum of one year from the health event. On completion of the residential home based rehabilitation programme, the team undertakes a comprehensive handover to designated home care providers that allows for an individually tailored education programme to be delivered to the formal and informal caregivers. The Promoting Independence Programme does not replace current NASC as COSE has, more it aims to integrate with current practice and overall case management remains with NASC.

Community FIRST offers a different approach and was established in 2002 in Hamilton through collaboration between Presbyterian Support Northern, Waikato District Health Board and The Ministry of Health. Essentially, Community FIRST was the first example of restorative home support for older people with high and complex needs in New Zealand. Restorative home support invariably involves the integration of physical activity into the day-to-day delivery of services. The model relies on a multi-disciplinary team (primarily registered nurse, physiotherapist and occupational therapist) providing an in-depth support plan, which is delivered by well-trained support workers / therapy aids under the close supervision of the multi-disciplinary team. Contact by support workers is up to four times a day and by registered nurses, a minimum of once every two weeks. The delivery of restorative home based support services can be divided into several levels according to the needs of the older person. Currently, the only provider delivering restorative home based support to older people of all needs level is Presbyterian Support Northern under the Enliven label. Older people accessing the Community FIRST service in Hamilton require a needs assessment and are eligible if they have high and complex needs. The funding

X

structure has been tailored to the unique needs of this group and a bulk funding arrangement is in place. The funding arrangement allows the older person in collaboration with the team coordinator to negotiate service provision, including frequency of visits, day centres attendance and caregiver residential respite. The service is based on core values such as care management, comprehensive geriatric assessment and functional and repetitive ADL training. All support programmes are orientated around the meaningful and invariably socially integrated goals of the older person, which are translated into support and exercise plans that ensure higher compliance and high quality. Community FIRST offers a replacement for current home care provision as opposed to the Promoting Independence Programme that provides enhanced oversight to traditional home care and residential care.

ASPIRE Assessment of Services Promoting Independence and

Recovery in Elders

ASPIRE is a meta-analysis of randomised controlled trials of the three ageing-in-place initiatives, COSE, PIP and Community FIRST, which means that the information arising from the three separate evaluations are pooled together to provide a greater level of confidence in the results. In essence, across Hamilton and Hutt, older people when assessed by NASC of having high and complex needs (i.e. at risk of entry to residential care) were randomly assigned to either usual care or the new ageing-in-place initiative. In Christchurch, GP practices were randomised to either usual NASC or the new COSE model and therefore when an older person was assessed by NASC as having high and complex needs, they were assigned on the basis of their GP (called cluster randomisation) to either NASC or COSE. A total of 55 GPs were assigned to one of the two groups and all older people within each GP practice received the same intervention. Ethical approval was granted from the lead ethics committee (Auckland) in July 2003 (Reference No: AKX/03/07/177). Interviews with the older person were undertaken before randomisation with a trained health professional research associate at the older person’s home. The initial interview involved an informed consent process and assessment of function (i.e. the independence levels of the older person), quality of life, how involved the older person was in the community, their use of health and social care services as well as assessment

XI

around mood. In addition, the caregiver (when present) was interviewed to assess their satisfaction with being a caregiver as well as questions around the impact on caring on their quality of life and employment opportunities. Adverse events such as falls, injuries and hospitalisations were recorded. The interRAI MDS-HC assessment was used as the research tool to collect the above information and where areas were missing, were supplemented by other assessment measures.

Recruitment began in November 2003 and lasted for 12 months in Hutt and Christchurch, but due to the difficulty in recruiting the required number of older people, recruitment lasted 18 months in Hamilton. Interviews were repeated at three months, six months and every six months to an average of 18 months and data collection ended in November 2005.

ASPIRE had several key objectives:

1 To assess the effectiveness of ageing-in-place initiative, as compared to usual care in preventing (or delaying) the time before a community-based older person requires permanent residential care

2 To assess the effectiveness of ageing-in-place initiative in improving survival in community-based older people compared to conventional care

3 To determine the impact of the ageing-in-place initiative on an older person’s independence and health-related quality of life compared to similar measures in those receiving conventional care

4 To establish the degree of correlation between the expected improvement in the health-related quality of life of informal caregivers attributable to ageing-in-place initiative, in comparison to those receiving conventional care

5 To determine the cost effectiveness of ageing-in-place initiative to the client, family, providers and funding agency in relation to the conventional care model

6 To assess the sustainability of ageing-in-place initiative to improve outcomes and cost changes over a two-year period

7 To identify the key elements of the ageing-in-place initiative healthcare models of community-based service delivery that lead to beneficial outcomes

Findings In total, 569 older people were randomised in the ASPIRE trial. Of these, 113 older people participated from the Hamilton region of which, 57 participants received usual care, and the remainder received Community FIRST. A total of 53 received usual care and 52 received PIP in Lower Hutt. In Christchurch, 182 received usual care and 169 received COSE. The

XII

level of baseline disability observed in each of the three sites varied considerably with older people in Christchurch being of far lower disability than those assessed using the same NASC criteria in Hutt and to a much greater extent in Hamilton.

The primary analysis revealed a statistically significant difference in the results for COSE compared to usual care in Christchurch for both the combined outcome (mortality and residential care admission) as well as for residential admissions (demonstrated by a 43% reduction in the risk of residential home placement) alone, though not mortality alone. Further, there was a reduction in residential home admission in Hamilton (33% risk reduction) and Hutt (16% risk reduction) and mortality in Hamilton (28% risk reduction) and Hutt (16% risk reduction) in the AIPI, though with the small sample size, the results were not statistically significant. The high impact of Community FIRST on residential home placement in Hamilton is particularly positive given the high baseline disability of older people in that region. The pooled analyses across centres (i.e. the Meta analysis) also showed a statistically significant treatment effect in delaying permanent residential home admission and combined primary outcome for the AIPI, which is around 30% lower than usual care with 95% confidence interval around 8% to 49%.

A trend for improvement in activities of daily living was observed in older people in the Community FIRST service compared to usual care. No trends were observed in either the Masonic PIP service or COSE. There were few differences in quality of life of the older person, though when the data from older people who entered a residential facility was removed, there was a trend for lower rates of depression in the Community FIRST participants. Importantly, the new initiatives did not appear to increase caregiver stress.

Predictive modelling was undertaken in the ASPIRE study using the 30 MDS-HC Home Care Quality Indicators as well as the EuroQoL Visual Analogue Scale and Caregiver Reaction Assessment. Hazard Risk ratios were calculated for these variables using hospitalisation and residential home admission as primary endpoints.

No medication review, negative mood and previous hospitalisation were correlated with increasing the risk of hospitalisation. Where as, inadequate meals, dehydration, ADL/rehab potential with no therapies, failure to improve/incidence of decline in ADL, social isolation, caregiver stress (CRA), negative mood and delirium were all correlated with an increased risk of residential home placement. Interestingly, when there is a failure

XIII

to improve or prevent a decline in ADL in an older person, there is an 11 times higher risk of the older person being admitted to a residential home. Such a result is highly pertinent for the increasing interest in restorative home support. These findings are useful in the development of an evidence based service specification for ageing-in-place initiatives.

OPERA (Older People Entering Residential Accommodation), a sub-study of ASPIRE provides the in-depth qualitative analysis around the ASPIRE study. It is very clear that few quality of life indicators are appropriate for older people in New Zealand and therefore qualitative findings and the conclusions drawn from such are highly pertinent for this population. The findings indicate that there are a number of factors that were highly important to the enrolled population of older people such as: coping, support, decisions and place of residence themes. The study also explored the process around decision-making in relation to placement and there appeared to be some disagreement around who made the decision for residential home placement, with the older person feeling that they were the main decision maker, though both family and NASC felt that the family was. Of the 131 older people interviewed (who were also enrolled in the ASPIRE study), of those that had relocated to a residential home nearly half were sad or very sad around the decision to move, whereas 75% of people living in their own home were happy or very happy with their decision to remain living at home.

Conclusions ASPIRE has provided highly valuable information around the relative successes of the ageing-in-place initiatives, it will allow informed decision-making around the evolution of ageing-in-place services. The results presented here must be viewed in light of the cost-effectiveness of the relative ageing-in-place initiatives, which is available in ASPIRE (Report II). Also of note are the very different approaches each initiative took to facilitate ageing-in-place. Where as there are clear benefits in exploring multiple means to support older people to age-in-place, there is a tendency to compare the ageing-in-place initiatives evaluated here and the relative success each achieved. In actuality, the strength of ASPIRE is to isolate those factors that are effective in facilitating ageing-in-place to allow new and existing services to evolve and develop.

XIV

The COSE project without doubt was highly successful in preventing and delaying entry of older people to residential care in Christchurch. As part of the OPERA study, NASC, reported that the main decision-makers in relation to residential care admission for older people were primarily family and indeed their own role was fairly minimal. However, the highly significant role of NASC in supporting older people to remain at home can not be underestimated and if nothing else, ASPIRE has clearly demonstrated this through the evaluation of COSE, which is in essence a modified community based NASC service. For the first time in New Zealand, ASPIRE has enabled a very thorough comparison of ‘high’ and ‘very high’ needs across three regions and it appeared that the actual disability level of older people assessed within these supposedly similar funding bands varied greatly across the three regions.

The level of effectiveness observed through COSE was so statistically significant, it is highly likely that implementation in other DHBs would result in either comparable or at the very least, positive changes. COSE was dependent on usual services to deliver packages of care and therefore, it is not surprising that there was no impact on the secondary outcomes, such as function, quality of life or depression. For the COSE service to be integrated and to maximise effectiveness, one would anticipate linking the initiative with a care delivery model such as either Community FIRST in Hamilton or the Promoting Independence Programme in Hutt.

Both Community FIRST and the Promoting Independence Programme provided very different solutions. Both had unique features and although neither service statistically significantly reduced residential home admission alone, both caused a reduction in risk, Community FIRST in the region of 33% and Promoting Independence Programme, 16%. It is not possible to know whether with an optimal sample size these figures would have been statistically significant. However, what was clear was that older people with a level of disability that would have normally required residential home admission were being maintained with no increase in adverse events, such as falls, hospitalisations, GP visits, or indeed an increase in caregiver stress in their own home. Further, in the case of Community FIRST there appeared to be a trend for an improvement in function and a reduction in depression.

XV

Key findings ASPIRE had several key objectives. The main objectives were to assess the effectiveness of ageing-in-place initiatives, as compared to usual care in:

Preventing (or delaying the time before) a community-based older person requires permanent residential care and;

In improving survival in community-based older people compared to usual care.

The research also sought to:

Determine the impact of the ageing-in-place initiatives on an older person’s independence and health-related quality of life compared to similar measures in those receiving conventional care. A wide range of indicators were used for this purpose and;

Establish the degree of correlation between the expected improvements in the health-related quality of life of informal caregivers attributable to ageing-in-place initiative, in comparison to those receiving conventional care.

In total, there were 569 participants randomised in the ASPIRE trial. Of these, 113 older people in the Hamilton region participated; 57 received usual care, and the rest received Community FIRST. In Lower Hutt, a total of 105 people participated, of which 53 received usual care and 52 received PIP. In Christchurch, 351 participated, 182 received usual care and 169 received COSE.

The sample sizes in the primarily Hamilton and Hutt were smaller than anticipated, for a number of reasons. This had some impact on the ability to determine statistically significant results. However, clear trends are apparent in all the results both between the services, and in comparison with the usual services. This gives confidence that the results are strongly indicative.

The key findings of ASPIRE are, which are not in any order of priority:

Older people with high and complex needs, who would otherwise be admitted to residential care can remain living at home with no apparent increased risk of harm.

The current Support Needs Level Assessment and categorisation system used by the NASC services to determine allocation of funding was highly variable across the three District Health Boards under investigation. It appeared that older people in Christchurch were assessed as being able to enter Residential care with a lower level of disability than those living in Hamilton and Lower Hutt. The interRAI MDS-HC

XVI

assessment tool used by the research team appeared to provide a more rigorous and standardised method of assessment. This variation is probably a factor in variations between the key outcomes achieved in different services, such as reducing mortality or admission to residential care.

All three services appeared to reduce the risk of mortality compared with usual services. This varied from 28% in Community FIRST, 14% in PIP and 10% in COSE in comparison to older people in usual care. Although these figures are not statistically significant they reflect a clear trend in each service and are consistent with other results in the trial.

COSE reduced the risk of entry of older people to residential care in comparison with the usual care NASC services by 43% (reduction).

Community FIRST (appeared to result) in a reduction of risk of entry to residential care by 33%, in comparison to usual care, though given the lower sample size this was not statistically significant.

The Promoting Independence Programme appeared to reduce risk of entry to residential care by 16% in comparison to usual care. Given the lower sample size this was not statistically significant.

Caregiver stress levels did not appear to increase in the intervention groups in comparison to usual care, despite the higher number of older people with high and complex needs remaining living at home.

An improvement in the independence levels of older people (Activities of Daily Living) within Community FIRST was noted, in comparison to usual care. No change was noted in function in the COSE or PIP initiatives in comparison to usual care.

Predictive modelling of the likelihood of older people being hospitalised or entering residential care was carried out using all the older people in the sample. This produced interesting results consistent with much overseas research. • If a functional decline occurs in older people and the deterioration is not

stopped, the older person is 11 times more likely to enter residential care. • An older person is almost twice as likely to enter residential care if they are

socially isolated. • If an older person reports as having a negative mood, they are over twice as

likely to be admitted to residential care. • For every one unit increase on the Caregiver Reaction Assessment (which

measures caregiver stress), there is a 7% increased risk of residential care entry.

• When an older person experiences inadequate meals and dehydration, they are over twice and 1.7 times more likely to be admitted to residential care, respectively.

• Delirium is highly correlated with risk of admission to residential care; those older people with delirium are 3.6 times more likely to be institutionalised.

XVII

• A lack of medication review (almost twice as likely), negative mood (1.5 times more likely) and previous hospitalisation (1.8 times more likely) are all correlated with increased risk of hospitalisation.

A related piece of research reported in ASPIRE was the Older People Entering Residential Accommodation (OPERA) study, part of a PHD study using the same population. This research was based on in-depth interviews with a small sample (n=131) and is not part of the ASPIRE dataset. Given the shortage of widely accepted quality of life indicators appropriate for older people in New Zealand this study will contribute towards the development of these measures.

The findings show there are a number of factors highly important to the enrolled population of older people such as coping, support, decision making and place of residence.

The study also explored the process by which older people entered residential care. Whilst the majority of older people often felt they had made the decision (to enter residential care) in most cases both the family and the NACS services thought that the family had been the main decision-makers.

Nearly half of those who had entered residential care were sad or very sad about the decision. By contrast three quarters of those living in their own homes were happy or very happy with their decision to remain living at home.

XVIII

Table of contents

Chapter I: Ageing in place 1.1 Introduction............................................................................................................. 1 1.2 Ageing-in-place initiatives ....................................................................................... 3 1.2.1 Rationale for selecting Ageing-in-place initiatives................................................... 3 1.2.2 Coordination of Services for Elderly (COSE) .......................................................... 4 1.2.3 Restorative and habilitative focussed community home support ............................ 7 1.2.4 Masonic Promoting Independence Programme.................................................... 10 1.2.5 Summary .............................................................................................................. 11 Chapter II: The ASPIRE Evaluation 2.1 Introduction........................................................................................................... 13 2.2 Ageing-in-place initiatives to be evaluated in this study........................................ 14 2.2.1 The usual care (Control) group............................................................................. 15 2.3 Study aims............................................................................................................ 16 2.4 Study objectives ................................................................................................... 16 2.5 Study design......................................................................................................... 16 2.5.1 Participating centres ............................................................................................. 16 2.5.2 Criteria for eligibility .............................................................................................. 17 2.5.3 Study interventions ............................................................................................... 18 2.5.4 Recruitment strategies.......................................................................................... 19 2.5.5 Randomisation assignment .................................................................................. 23 2.5.5 Randomisation assignment .................................................................................. 24 2.6 Outcome measures .............................................................................................. 25 2.6.1 Primary end-points ............................................................................................... 25 2.6.2 Secondary end-points........................................................................................... 25 2.6.3 Outcome measures of the older person................................................................ 26 2.6.4 Outcome measures of the primary informal caregiver .......................................... 30 2.7 Determination of the cost-effectiveness................................................................ 33 2.8 Older People Entering Residential Accommodation (OPERA) ............................ 33 2.8.1 Pilot study ............................................................................................................. 34 2.8.2 Main study ............................................................................................................ 34 2.9 Ethical considerations........................................................................................... 35 2.10 Adverse event reporting........................................................................................ 35 2.10.1 ASPIRE interim reporting...................................................................................... 36 2.11 Study definitions ................................................................................................... 37 2.11.1 Institutionalised-free survival ................................................................................ 37 2.11.2 Participant withdrawal and lost to follow-up .......................................................... 37 2.11.3 Censoring date ..................................................................................................... 38 2.11.4 Changes from Baseline ........................................................................................ 38 2.12 Statistical issues ................................................................................................... 38 2.12.1 Sample size and cluster size calculations............................................................. 38 2.12.2 Amendments to the protocol................................................................................. 39

XIX

2.12.3 Statistical analysis ................................................................................................ 40 2.12.4 Reporting of intra-cluster correlation coefficient (ICC) .......................................... 40 2.12.5 Primary endpoint .................................................................................................. 41 2.12.6 Secondary endpoints ............................................................................................ 41 2.12.7 Subgroup analysis ................................................................................................ 43 2.12.6 Per-protocol data set ............................................................................................ 43 2.12.9 Missing values and outliers................................................................................... 43 2.12.10 Predictive modelling analyses .............................................................................. 43 2.13 Summary .............................................................................................................. 46 Chapter III: Findings 3.0 Introduction........................................................................................................... 48 Section 1: Sample characteristics 3.1 Introduction........................................................................................................... 49 3.2 Recruitment .......................................................................................................... 49 3.2.1 Protocol violations ................................................................................................ 50 3.2.2 Completeness of information and treatment allocation ......................................... 51 3.2.3 Older person demographics ................................................................................. 53 3.2.4 Caregiver demographics....................................................................................... 65 3.3 Summary .............................................................................................................. 67 Section 2: Primary results 3.4 Introduction........................................................................................................... 69 3.5 Design effect......................................................................................................... 70 3.6 Permanent residential home placement and mortality ......................................... 71 3.7 Permanent residential home admission................................................................ 72 3.8 Mortality ................................................................................................................ 74 3.9 Combined primary endpoint.................................................................................. 76 3.10 Summary .............................................................................................................. 79 Section 3: Secondary results 3.11 Introduction........................................................................................................... 80 3.12 Older person......................................................................................................... 81 3.12.1 Presentation of Instrumental and Activities of daily living findings ........................ 81 3.12.2 Presentation of VAS, CPS, DRS, CHESS and Pain findings................................ 89 3.12.3 Presentation of GP visit findings........................................................................... 96 3.12.4 Sub-group analysis............................................................................................... 97 3.13 Informal caregiver................................................................................................. 97 3.14 Summary ............................................................................................................ 105 Section 4: Analysis of risks of entry to residential care and hospitalisation 3.15 Introduction......................................................................................................... 106 3.16 The risk of hospitalisation ................................................................................... 107 3.17 The risk of residential home admission............................................................... 109

XX

Section 5: The impact of moving into a residential facility 3.18 Introduction......................................................................................................... 111 3.19 Findings.............................................................................................................. 112 3.19.1 Change............................................................................................................... 112 3.19.2 Control ................................................................................................................ 113 3.19.3 Placement........................................................................................................... 113 3.19.4 Coping ................................................................................................................ 114 3.18.5 Support ............................................................................................................... 116 3.19.6 Decisions theme ................................................................................................. 117 3.19.7 Residence........................................................................................................... 120 3.20 Summary ............................................................................................................ 122 Chapter IV: Key findings 4.1 Introduction......................................................................................................... 125 4.2 Conclusions ........................................................................................................ 127 Chapter V: Study limitations 5.1 Introduction..............................................................Error! Bookmark not defined. 5.2 Recruitment ........................................................................................................ 128 5.3 Design issues ..................................................................................................... 129 5.4 Selection bias ..................................................................................................... 129 5.5 Maturation........................................................................................................... 130 5.6 Testing................................................................................................................ 130 5.7 Hawthorne effect ................................................................................................ 131 5.8 Experimenter effects........................................................................................... 131 Appendices Appendix 1: Statistical analysis ....................................................................................... 134 Appendix 2: The interRAI MDS-HC ................................................................................. 137 Appendix 3: Survival plots ............................................................................................... 146 Appendix 4: Secondary outcome data (Excluding residential home) ............................... 152 Appendix 5: Secondary outcome data (Including residential home) ................................ 198 Appendix 6: Adverse events ............................................................................................ 244 Appendix 7: Questionnaires............................................................................................. 247 References .................................................................................................................... 236

XXI

List of tables and figures

Chapter I: Ageing-in-place Table 1-1: Ageing-in-place initiatives (Key features) ..............................................................11 Chapter II: The ASPIRE Evaluation Table 2-2: Investigation schedule...........................................................................................26 Table 2-3: Summary of MDS-HC outcome scales validation..................................................29 Table 2-4: Caregiver Reaction Assessment Instrument .........................................................32 Table 2-5: Inter-RAI Home Care Quality Indicators ................................................................45 Chapter III: Findings Table 3-6: Study recruitment ..................................................................................................50 Table 3-7: ASPIRE Protocol violations ...................................................................................51 Table 3-8: Recruitment and enrolment over time ...................................................................51 Table 3-9: Treatment allocation and enrolment over time across all three sites.....................52 Table 3-10: Older person demographics across the each of the three sites ............................54 Table 3-11: Older person demographics across the three sites in total....................................55 Table 3-12: Older person demographics across each of the three sites ..................................56 Table 3-13: Older person demographics across the three sites in total....................................58 Table 3-14: Baseline older person demographics medical history, across the three sites .......59 Table 3-15: Baseline older person demographics com. services, across the three sites .........60 Table 3-16: Baseline older person demographics com. services, across the three sites .........62 Table 3-17: Baseline older person demographics scale, across the three sites.......................63 Table 3-18: ADL Long Form Scale scoring responses .............................................................65 Table 3-19: Baseline caregiver demographics across the three sites ......................................66 Table 3-20: Baseline caregiver demographics scale measurements across all three sites......67 Table 3-21: Hazard ratio on primary outcome – residential care .................................73 Table 3-22: Hazard ratio on Primary outcome – death.............................................................75 Table 3-23: Hazard ratio on Combined primary outcome – residential and mortality ...............77 Table 3-24: Treatment - ADL and IADL (Excl. residential home data) in Hamilton...................83 Table 3-25: Treatment - ADL and IADL (Incl. residential home data) in Hamilton....................84 Table 3-26: Treatment - ADL and IADL (Excl. residential home data) in Hutt .........................85 Table 3-27: Treatment - ADL and IADL (Incl. residential home data) in Hutt ..........................86 Table 3-28: Treatment - ADL and IADL (Excl. residential home data) in Christchurch............87 Table 3-29: Treatment - ADL and IADL (Incl. residential home data) in Christchurch.............88 Table 3-30: Treatment - VAS, CPS, DRS, CHESS and Pain (Excl. res data) in Hamilton .......90 Table 3-31: Treatment - VAS, CPS, DRS, CHESS and Pain (Incl. res data) in Hamilton.........91 Table 3-32: Treatment - VAS, CPS, DRS, CHESS and Pain (Excl. res data) in Hutt ...............92 Table 3-33: Treatment - VAS, CPS, DRS, CHESS and Pain (Incl. res data) in Hutt ...............93 Table 3-34: Treatment - VAS, CPS, DRS, CHESS and Pain (Excl. res data) in Christchurch .94

XXII

Table 3-35: Treatment - VAS, CPS, DRS, CHESS and Pain (Incl. res data) in Christchurch...95 Table 3-36: Number of GP visits at each follow-up ..................................................................96 Table 3-37: Treatment - Caregiver SF36, Caregiver Reaction Assessment and EuroQoL

perception of older person (Excluding residential home data) in Hamilton...........99 Table 3-38: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at

each visit for Caregiver SF36, Caregiver Reaction Assessment and EuroQoL perception of older person (Including residential home data) in Hamilton..........100

Table 3-39: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for Caregiver SF36, Caregiver Reaction Assessment and EuroQoL perception of older person (Excluding residential home data) in Hutt ................101

Table 3-40: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for Caregiver SF36, Caregiver Reaction Assessment and EuroQoL perception of older person (Including residential home data) in Hutt .................102

Table 3-41: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for Caregiver SF36, Caregiver Reaction Assessment and EuroQoL perception of older person (Excluding residential home data) in Christchurch...103

Table 3-42: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for Caregiver SF36, Caregiver Reaction Assessment and EuroQoL perception of older person (Including residential home data) in Hamilton..........104

Table 3-43: Hazard Ratio estimates Hospitalisation as the primary end-point .......................108 Table 3-44: Hazard Ratio estimates Residential Care Entry as the primary end-point...........110 Table 3-45: Older people’s report on decision makers ...........................................................118 Table 3-46: Caregivers’ report on decision makers ................................................................118 Table 3-47: NASC report on major decision makers ..............................................................118 Table 3-48: Percentage of decision makers for the older person to return home..................119 Table 3-49: Percentage of decision makers for the older person to enter residential care.....119 Table 3-50: Older people’s feelings about their residential placement decision ......................121 Appendices Table 5-51: ADL self-performance hierarchy scale ................................................................138 Table 5-52: CPS rating scale..................................................................................................140 Table 5-53: The Depression Rating Scale (DRS)...................................................................142 Table 5-54: IADL difficulty scale .............................................................................................142 Table 5-55: IADL Involvement Scale ......................................................................................143 Table 5-56: The CHESS scale ...............................................................................................144 Table 5-57: Adverse events count (including RH data) for Hamilton ......................................244 Table 5-58: Adverse events count (including RH data) for Hutt..............................................245 Table 5-59: Adverse events count (including RH data) for Christchurch ................................246

XXIII

Chapter III: Findings Figure 3-6: Adjusted overall treatment effect estimate (residential home)...............................74 Figure 3-7: Adjusted overall treatment effect estimate (death) ................................................76 Figure 3-8: Adjusted overall treatment effect estimate (residential home and death) ..............78 Figure 3-10: Older people's feelings about their decision for residential placement ................122 Appendices Figure 5-14: Survival in Hamilton region (using Death as Primary endpoint) ..........................147 Figure 5-15: Survival in Lower Hutt region (using Death as Primary endpoint) ......................147 Figure 5-16: Survival in Christchurch region (using Death as Primary endpoint) ....................148 Figure 5-17: Suriuval in Hamilton region (using Residential Care entry) .................................148 Figure 5-18: Surival in Lower Hutt region (using Residential Care entry) ..............................149 Figure 5-19: Surival in Christchurch region (using Residential Care entry) .............................149 Figure 5-20: Surival in Hamilton region (using combined primary outcome) ...........................150 Figure 5-21: Surival in Lower Hutt region (using combined primary outcome) .......................150 Figure 5-22: Surival in Christchurch region (using combined primary outcome) .....................151 Figure 5-23: Repeated measures of ADL short form scale in Hamilton...................................153 Figure 5-24: Repeated measures of ADL short form scale in Lower Hutt................................154 Figure 5-25: Repeated measures of ADL short form scale in Christchurch.............................155 Figure 5-26: Repeated measures of ADL self-performance scale in Hamilton ........................156 Figure 5-27: Repeated measures of ADL self-performance scale in Lower Hutt .....................157 Figure 5-28: Repeated measures of ADL self-performance scale in Christchurch ..................158 Figure 5-29: Repeated measures of ADL Long form scale in Hamilton...................................159 Figure 5-30: Repeated measures of ADL Long form scale in Lower Hutt................................160 Figure 5-31: Repeated measures of ADL Long form scale in Christchurch.............................161 Figure 5-32: Repeated measures of IADL difficulty scale in Hamilton .....................................162 Figure 5-33: Repeated measures of IADL difficulty scale in Lower Hutt..................................163 Figure 5-34: Repeated measures of IADL difficulty scale in Christchurch ...............................164 Figure 5-35: Repeated measures of IADL Involvement scale in Hamilton...............................165 Figure 5-36: Repeated measures of IADL Involvement scale in Lower Hutt ...........................166 Figure 5-37: Repeated measures of IADL Involvement scale in Christchurch.........................167 Figure 5-38: Repeated measure of IADL Summary scale in Hamilton ....................................168 Figure 5-39: Repeated measures of IADL Summary scale in Lower Hutt ...............................169 Figure 5-40: Repeated measures of IADL Summary scale in Christchurch.............................170 Figure 5-41: Repeated measures of CPS scale in Hamilton ...................................................171 Figure 5-42: Repeated measures of CPS scale in Lower Hutt ................................................172 Figure 5-43: Repeated measures of CPS scale in Christchurch..............................................173 Figure 5-44: Repeated measures of DRS scale in Hamilton ...................................................174 Figure 5-45: Repeated measures of DRS scale in Lower Hutt ................................................175 Figure 5-46: Repeated measures of DRS scale in Christchurch .............................................176 Figure 5-47: Repeated measures of CHESS scale in Hamilton...............................................177 Figure 5-48: Repeated measures of CHESS scale in Lower Hutt ...........................................178 Figure 5-49: Repeated measures of CHESS scale in Christchurch.........................................179 Figure 5-50: Repeated measures of Pain scale in Hamilton....................................................180 Figure 5-51: Repeated measures of Pain scale in Lower Hutt ................................................181

XXIV

Figure 5-52: Repeated measures of Pain scale in Christchurch..............................................182 Figure 5-53: Repeated measures of SF36 PCS in Hamilton ...................................................183 Figure 5-54: Repeated measures of SF36 PCS in Lower Hutt ................................................184 Figure 5-55: Repeated measures of SF36 PCS in Christchurch .............................................185 Figure 5-56: Repeated measures of SF36 MCS in Hamilton...................................................186 Figure 5-57: Repeated measures of SF36 MCS in Lower Hutt................................................187 Figure 5-58: Repeated measures of SF36 MCS in Christchurch.............................................188 Figure 5-59: Repeated measures of CRA in Hamilton.............................................................189 Figure 5-60: Repeated measures of CRA in Lower Hutt .........................................................190 Figure 5-61: Repeated measures of CRA in Christchurch.......................................................191 Figure 5-62: Repeated measures of EQ VAS of OP in Hamilton.............................................192 Figure 5-63: Repeated measures of EQ VAS of OP in Lower Hutt..........................................193 Figure 5-64: Repeated measures of EQ VAS of OP in Christchurch.......................................194 Figure 5-65: Repeated measures of EQ VAS of OP from Caregiver in Hamilton ....................195 Figure 5-66: Repeated measures of EQ VAS of OP from Caregiver in Lower Hutt .................196 Figure 5-67: Repeated measures of EQ VAS of OP from Caregiver in Christchurch ..............197 Figure 5-68: Repeated measures of ADL short form scale in Hamilton...................................199 Figure 5-69: Repeated measures of ADL short form scale in Lower Hutt................................200 Figure 5-70: Repeated measures of ADL short form scale in Christchurch.............................201 Figure 5-71: Repeated measures of ADL self-performance scale in Hamilton ........................202 Figure 5-72: Repeated measures of ADL self-performance scale in Lower Hutt .....................203 Figure 5-73: Repeated measures of ADL self-performance scale in Christchurch ..................204 Figure 5-74: Repeated measures of ADL Long form scale in Hamilton...................................205 Figure 5-75: Repeated measures of ADL Long form scale in Lower Hutt................................206 Figure 5-76: Repeated measures of ADL Long form scale in Christchurch.............................207 Figure 5-77: Repeated measures of IADL difficulty scale in Hamilton .....................................208 Figure 5-78: Repeated measures of IADL difficulty scale in Lower Hutt..................................209 Figure 5-79: Repeated measures of IADL difficulty scale in Christchurch ...............................210 Figure 5-80: Repeated measures of IADL Involvement scale in Hamilton...............................211 Figure 5-81: Repeated measures of IADL Involvement scale in Lower Hutt ...........................212 Figure 5-82: Repeated measures of IADL Involvement scale in Christchurch.........................213 Figure 5-83: Repeated measure of IADL Summary scale in Hamilton ....................................214 Figure 5-84: Repeated measures of IADL Summary scale in Lower Hutt ...............................215 Figure 5-85: Repeated measures of IADL Summary scale in Christchurch.............................216 Figure 5-86: Repeated measures of CPS scale in Hamilton ...................................................217 Figure 5-87: Repeated measures of CPS scale in Lower Hutt ................................................218 Figure 5-88: Repeated measures of CPS scale in Christchurch..............................................219 Figure 5-89: Repeated measures of DRS scale in Hamilton ...................................................220 Figure 5-90: Repeated measures of DRS scale in Lower Hutt ................................................221 Figure 5-91: Repeated measures of DRS scale in Christchurch .............................................222 Figure 5-92: Repeated measures of CHESS scale in Hamilton...............................................223 Figure 5-93: Repeated measures of CHESS scale in Lower Hutt ...........................................224 Figure 5-94: Repeated measures of CHESS scale in Christchurch.........................................225 Figure 5-95: Repeated measures of Pain scale in Hamilton....................................................226 Figure 5-96: Repeated measures of Pain scale in Lower Hutt ................................................227 Figure 5-97: Repeated measures of Pain scale in Christchurch..............................................228

XXV

Figure 5-98: Repeated measures of SF36 PCS in Hamilton ...................................................229 Figure 5-99: Repeated measures of SF36 PCS in Lower Hutt ................................................230 Figure 5-100: Repeated measures of SF36 PCS in Christchurch .............................................231 Figure 5-101: Repeated measures of SF36 MCS in Hamilton...................................................232 Figure 5-102: Repeated measures of SF36 MCS in Lower Hutt................................................233 Figure 5-103: Repeated measures of SF36 MCS in Christchurch.............................................234 Figure 5-104: Repeated measures of Caregiver Reaction Assessment in Hamilton .................235 Figure 5-105: Repeated measures of Caregiver Reaction Assessment in Lower Hutt ..............236 Figure 5-106: Repeated measures of Caregiver Reaction Assessment in Christchurch ...........237 Figure 5-107: Repeated measures of VAS of Older Person in Hamilton ...................................238 Figure 5-108: Repeated measures of VAS of Older Person in Lower Hutt................................239 Figure 5-109: Repeated measures of VAS of Older Person in Christchurch .............................240 Figure 5-110: Repeated measures of VAS of Older Person from Caregiver in Hamilton ..........241 Figure 5-111: Repeated measures of VAS of Older Person from Caregiver in Lower Hutt ......242 Figure 5-112: Repeated measures of VAS of Older Person from Caregiver in Christchurch ...243

XXVI

INTRODUCTION

PAGE 1

Chapter I: Ageing-in-place

1.1 Introduction

Currently, older people (65+) make up 12.4% of the population. This is anticipated to rise to 25% by 2050. Of more significance however, is the fourfold increase in 75+ year olds predicted to occur over the next 20 years. Given that 75+ year olds utilise three times the health care resources of other age groups, the impact on health and social resources will be considerable. The predicted demographic changes in the Mäori population are even more pressing (Kokiri, 1996). From 1998 to 2010, a fourfold increase in 75+ Mäori will be observed and given the incidence of age-related conditions occurring at younger ages, there are concerns around the impact of such an increase on the whanau. Advancing age is associated with declines in physiological reserve and physical functioning and a higher risk of disability and dependency. The looming socio-economic impact of providing services for the increasing number of older people in the population is of major concern to politicians, health care providers and policymakers alike. Merely adding years to life, rather than life to years, should no longer be acceptable and interventions are required to be in place which aim to prevent the emergence of disability or alleviate existing disability in later life. Both the Positive Ageing Strategy and Health of Older People Strategy provide guidance to District Health Boards (DHB) around the development of services for older people. However, without appropriate evidence to inform the direction of services, there is a distinct risk that services that aim to prevent or modify existing disability in old age develop in an ad hoc and disparate manner.

Internationally, the concept of ageing-in-place has been gaining currency in policy for many years and has undergone shifts in its interpretation during this period. The first major advance occurred in 1994 when OECD ministers reached a consensus that people should be able to continue living in their own place of residence in their later years. In the event that this is no longer possible, the alternative would be for older people to live in a ‘sheltered and supportive environment which is as close to their community as possible, in both the social and geographical sense’ (Organisation for Economic Co-operation and Development, 1994). Within New Zealand, ageing-in-place is defined as the ability of

INTRODUCTION

PAGE 2

people to “make choices in later life about where to live, and receive the support to do so” (Dalziel 2001, pg. 10). Despite this broad categorisation, ageing-in-place invariably refers to the ability of older people to remain dwelling in the community, including within retirement villages (MSD, 2006) and moreover, residential care in the form of either rest homes or hospitals is specifically excluded. In this, the wider definition is consistent with the approach taken in the New Zealand report which was part of the International Year of Older Persons activities, Factors affecting the ability of older people to live independently

(Dwyer, Gray & Renwick, 2000). The focus on remaining in the community has also been promoted repeatedly within the New Zealand policy arena (Richmond et al. 1995; Ministry of Health 2002;, 2004; Davey et al. 2004).

Ageing-in-place as a concept and policy direction has clearly developed as a result of the desire of most older people, even those with considerable disability to remain living at home rather than enter residential care (Salvage, Jones et al. 1989). Consequently, community support has acquired increasing relevance and therefore funding (Hing and Bloom 1991; Steel 1991; Coleman 1995; Kane and Kane 1995; Stuck, Aronow et al. 1995). Such a shift is in accordance with the key concept of the New Zealand Health of Older People Strategy (MoH, 2002), that services will be established that allow older people to age-in-place, providing them with the ability to make choices about where to live and to receive the support they require to do so. Certainly, remaining at home allows the older person to maintain social networks and a quality of life, and to continue integration with the community. However, although it is widely recognised that older people wish to age in place, it is often inherently difficult to develop and deliver services that appropriately facilitate this desire. It is clear that most health care costs occur in the last few years of life, arising in the main from the strong association between rising disability and increasing age. There exists a paradox therefore on the one hand of DHBs and other healthcare organisations wishing to promote ageing-in-place but on the other, few viable means to achieve this effectively. This is confounded by several factors. Increasingly, there is a growing awareness that what matters most to older people is not just health in its narrowest interpretation, but moreover older people are concerned about choice, autonomy, independence and community integration amongst many others. Given that current service models are invariably delivered with a traditional bio-medical focus, it is not surprising that these areas which are of high importance to the older person are omitted.

INTRODUCTION

PAGE 3

As a consequence of the somewhat limited scope of home based service delivery models currently available through District Health Boards (DHB), a number of organisations have developed new methods of delivering services that promote ageing-in-place in it’s fullest context.

1.2 Ageing-in-place initiatives

Currently, older people receive support services from a DHB following assessment and service co-ordination by a Needs Assessment and Service Co-ordination agency (NASC). When high and complex needs are identified, the older person is offered either (a) a package of care consisting of a combination of family and community resources and services to facilitate the older person remaining at home or; (b) if a package of care cannot assist them to stay safely at home, then they enter residential care. Guided by the Health of Older People Strategy (MoH, 2002), the DHBs commenced the development and implementation of numerous ageing-in-place initiatives (AIPI) in various parts of the country as alternative models of care to the existing NASC co-ordinated home services and residential care. It was envisaged that these services would be integrated and continuous, based on a case-management model whilst maintaining cost-effectiveness. However, due to the heterogeneity of older people in conjunction with geographical variance in the delivery of health services within New Zealand, there is always a distinct risk that there will be an increase in the regional disparity between service access and delivery. Further, evidence is lacking around whether the AIPI will result in an improvement in quality of life and independence for the older person while maintaining cost-effectiveness for the provider compared to existing services. To be cost-effective, the goal of AIPI is to produce the best clinical and social outcomes at the lowest cost.

1.2.1 Rationale for selecting Ageing-in-place initiatives

Of the AIPI services that were proposed and developed nationally, three were identified as the most significant in terms of guiding future policy direction. These were: The Coordinator of Services for Elderly in Christchurch (Canterbury DHB), The Promoting Independence Programme in Lower Hutt (Wellington Masonic Villages Trust in partnership with Hutt Valley DHB) and Community F.I.R.S.T. (Flexible Integrated Restorative Support

INTRODUCTION

PAGE 4

Team) in Hamilton (Presbyterian Support Northern in partnership with Waikato DHB). The rationale for the inclusion of these three services was that they broadly represented a range of services offered to older people in the community. Coordination of Services for Elderly (COSE) are NASC workers who have been relocated from a hospital location to the community and linked to a number of General Practice surgeries. The Promoting Independence Programme is a residential transition service for older people requiring a period of slow stream facility based rehabilitation following discharge from hospital and prior to returning home. Community FIRST is an intensive home based support service with a ‘restorative’ or promoting independence focus. Individually, each service provides a unique approach to ageing-in-place, though collectively, the three services represent the main aspects of disability support services for older people who are at risk of permanent residential care. It is for these reasons, that COSE, PIP and Community FIRST were selected by The Ministry of Health for in evaluation in the ASPIRE project. It is useful to explore the attributes of each of the services in more detail.

1.2.2 Coordination of Services for Elderly (COSE)

The role of NASC has traditionally been one of case management. However, for one reason or another, it has often been difficult to fully realise the full potential of the model. The evolution of COSE in Canterbury arose out of the desire to improve on the existing NASC model. The following is a brief description and underlying philosophy of COSE.

Components of case management

It is well recognised that coordinating services or care packages in response to the assessed needs of older people and their carers is a core part of delivering integrated, person-centred care. Good, comprehensive assessment and care planning; undertaken in a way that properly engages with the older person and their carer and involves them in decisions about their care plan is crucial in ensuring that the most appropriate services are provided. Co-ordinating these processes and services has the potential to avoid unnecessary duplication and promote good continuity of care. This facilitates older people’s independence by preventing deterioration in their health and home situation and by managing crises, as Challis describes “The impact of services upon well-being is much

INTRODUCTION

PAGE 5

greater when those services are planned and co-ordinated in an integrated fashion” (Challis et al, 2002). Ultimately, the aim of case management and indeed COSE is to tailor services to the individual older person in order to improve the quality of their life, taking into account the wishes and needs of their carer, be they a partner, relative or friend. The challenge for case management is that it takes place at the level of service provision at which needs and resources, scarcity and choice have to be balanced. Care management is no panacea but rather a mechanism which, if effectively implemented can offer one way to manage the tension between social objectives and economic constraints in long-term care services (Challis, 2003).

Traditionally case managers may be nurses, social workers, physiotherapists or other professionals. Bergen (1992; 1994; 2003) suggests that case management is operationally divided into phases of case finding, assessment and need identification, design and implementation of care packages, monitoring, evaluation or reassessment, that lead to the final phase of case closure or repetition of the cycle. However, within this broad framework, there is considerable variation especially with respect to caseload, how services are provided, and provision of specific services (e.g., home visits, education, counselling) (Pacala, Boult et al. 1995).

Case management aims at a controlled balance between quality and cost. Its goals are to:

(a) improve the quality of patient care through emphasising the importance of health restoration and maintenance and increased continuity of care;

(b) decrease the cost of care through empowering patients and their family to maximise self-care capabilities and prevent unnecessary or lengthy admissions; and

(c) improve patient, nurse and physician satisfaction and professional development through promotion of multi-disciplinary collaborative practice and coordinated care (McKenzie, Torkelson et al. 1989; Giuliano and Poirier 1991; Bryan, Dickerson et al. 1994; Crawley 1994; Gibson, Martin et al. 1994).

The evolution from NASC to COSE

Many different models of case management can be seen across New Zealand. One of the more prevalent forms is Needs Assessment Service coordination (NASC). NASC provides an assessment and service brokerage facility for people requiring access to disability services. NASC can be broadly separated into two aspects; the MoH funded under 65 and

INTRODUCTION

PAGE 6

the DHB funded over 65. One of the inherent criticisms of such a model is access to services. The development of COSE has come about from this issue and also the need to firmly embed such a service in a distinct primary care location.

COSE was established in 2000 and is a community-based needs assessment and service co-ordination initiative funded by MoH, DHB and ACC. It was established with the aim of avoiding duplication in service provision. A key worker (COSE) is based in primary health care and is assigned to several general practice teams. The model allows the COSE to identify resources and opportunities within communities, both funded and non-funded. This offers older people a greater choice of service support, enabling them to remain safely in the community as long as they wish to. There is a strong evidence base for COSE. Studies that have evaluated community-based case management of older people (or nurse home visitation to older people) have been conducted in several countries such as the United States (Eggert, Zimmer et al. 1991; Rogers, Riordan et al. 1991; Fitzgerald, Smith et al. 1994; Swindle, Weyant et al. 1994; Stuck, Aronow et al. 1995), the United Kingdom (Vetter, Jones et al. 1984; Pathy, Bayer et al. 1992; Dunn, Guy et al. 1995; Runciman, Currie et al. 1996), and Italy (Bernabei, Landi et al. 1998; Landi, Gambassi et al. 1999). They have been found to reduce hospital admissions (Hendriksen, Lund et al. 1984; Rogers, Riordan et al. 1991; Swindle, Weyant et al. 1994; Rantz, Mehr et al. 2000), the length of stay in hospital (Eggert, Zimmer et al. 1991; Rogers, Riordan et al. 1991; Pathy, Bayer et al. 1992; Swindle, Weyant et al. 1994), mortality (Hendriksen, Lund et al. 1984; Vetter, Jones et al. 1984; Pathy, Bayer et al. 1992), emergency department visits (Rogers, Riordan et al. 1991; Pathy, Bayer et al. 1992), admission to long-term facilities (Stuck, Aronow et al. 1995) and costs of care (Hendriksen, Lund et al. 1984; Eggert, Zimmer et al. 1991; Rogers, Riordan et al. 1991; Swindle, Weyant et al. 1994; Bernabei, Landi et al. 1998; Landi, Gambassi et al. 1999).

What is COSE?

COSE workers are drawn from multiple professional groups (typically: Nursing, Occupational Therapy, Physiotherapy and Social Work). A COSE worker is assigned to a cluster of designated practices of General Practitioners (GP), but works independently of the practices. Importantly, the COSE worker is physically located in the community,

INTRODUCTION

PAGE 7

invariably in locations central to the cluster of ‘attached’ GPs. The COSE worker undertakes comprehensive assessments of the older person, assessments that are currently taken by NASC staff, practice nurses and home care providers. The COSE worker is a case manager, liaising with the GPs and practice nurses ensuring that there is recognition of and a quick response to any change in an older person’s circumstances thereby allowing the level of care that is required for safe continuous ageing-in-place. As a component of providing an appropriate level of care, the COSE worker co-ordinates the appropriate community services, informal networks and medical care based on assessed need and GP liaison.

COSE has been established in East Canterbury since October 2000 and was extended to North, West and South Canterbury in 2003. It is these latter sites that were evaluated.

1.2.3 Restorative and habilitative focussed community home support

There is growing recognition that physical inactivity and disuse plays a major role in the well-reported age-related conditions, such as diabetes, sarcopenia (muscle loss) and heart disease. Research highlights the linear reduction in muscle mass over time, to a point where a woman in her mid-80s would often have borderline sufficient strength to stand from a chair unaided. Many researchers and clinicians write of the harm associated with ‘wrapping older people in cotton wool’ and much of this deterioration is linked to deconditioning and disuse (McMurdo, 1999). The recognition that old age is often associated with poor fitness and deconditioning forms the basis of restorative home support. There is a strong belief that older people have considerable ongoing potential to recover fitness and therefore restorative home support invariably involves the integration of physical activity into the day-to-day delivery of services. Recent work by de Vreede et al (2005) indicates that repetitive task based or ADL (activities of daily living) exercises for older people, as seen in restorative home support services are more effective than resistance exercises employed by traditional physiotherapy techniques.

The Quality and Safety project identified many of the issues facing the home support sector. These issues were later corroborated by The University of Auckland undertaking research under contract for the MoH, such as poor morale, high staff turnover, inefficient

INTRODUCTION

PAGE 8

service models and funding structures, as well as unresponsive and unwieldy staffing systems, poor training and ultimately poor quality (Parsons et al, 2003; 2004a; 2004b; 2004c). The concept of restorative home support with an associated shift in the funding structure was noted as having the potential to effectively address many if not all of these issues. The National Select Committee on Home Support recommended the Community FIRST type approach (a restorative home support model for older people with high and very high needs) as being a viable method of addressing these issues and moreover maximising the potential for older people to ‘age in place’.

The model relies on a multi-disciplinary team (primarily registered nurse, physiotherapist and occupational therapist) providing an in-depth support plan, which is delivered by well-trained support workers / therapy aids under the close supervision of the multi-disciplinary team. Contact by support workers is up to four times a day and by registered nurses, a minimum of once every two weeks. The delivery of restorative home based support services can be divided into several levels according to the needs of the older person. Currently, the only provider delivering restorative home based support to older people of all needs level is Presbyterian Support Northern under the Enliven label.

Community FIRST; an example of high intensive restorative home support

Although restorative home support can be configured to deliver services to older people irrespective of need, the Presbyterian Support home care services in Hamilton, Rotorua, Tauranga, Timaru, Dunedin and shortly Hawkes Bay are funded through DHBs to deliver community based restorative home support for older people assessed as requiring residential placement. Older people accessing the service require a needs assessment and are eligible if they are over 65 years of age and have high and complex needs. The funding structure has been tailored to the unique needs of this group and a suitable funding arrangement is in place.

The service is based on core values such as care management, comprehensive geriatric assessment and functional and repetitive ADL training. All support programmes are orientated around the meaningful and invariably socially integrated goals of the older person, which are translated into support and exercise plans that ensure higher compliance and high quality. The service model itself represents a philosophy that has

INTRODUCTION

PAGE 9

arisen from several pieces of research. The Community FIRST model was based on the supported discharge team established in South London and evaluated through a randomised controlled trial design (Martin et al, 1994). The evaluation demonstrated that input from the team resulted in a reduction in readmission to hospital and reduction in admission to residential home over a year period.

A key concept of the service is to base a support programme around the goals and aspirations of the older person. The Health of Older People Strategy states the importance of using a holistic person-centred approach that promotes wellness and active participation in the decisions about service user care. One of the major issues identified in the Assessment Processes for Older People Guidelines relates to examples where there can be an apparent lack of agreement between the health professional and older person regarding the main priorities for health care. The report recommends that respect for the older person’s concerns and the creation of equality in the decision-making process encourages concordance. Therefore, the process of goal-setting should be considered as an important mechanism. The goal facilitation process involves TARGET (Towards Achieving Realistic Goals in Elders Tool), the e-based training around which is currently being developed for Counties Manukau DHB and Presbyterian Support National and will be available for dissemination later this year. A key component to the service is the ability of the service coordinators to undertake a comprehensive geriatric assessment (CGA) and put in place an appropriate care package. The meta-analysis by Stuck et al (1993) indicates that when a comprehensive geriatric assessment (CGA) is linked to a strong care package, there is a decrease in admission to residential homes, decrease in mortality, reduction in falls and improvement in ADL function over time.

The Community FIRST model was initially developed by Presbyterian Support Nationally and the regions are working towards managing service developments nationally in order to ensure a cohesive approach to delivery, staff training and development. Therefore, irrespective of region, the service model and criteria are similar though are modified geographically to achieve maximum benefits for the older person, such as relationships with key organisations. The Hamilton site which provides services for older people with high and complex needs was evaluated as part of the ASPIRE study.

INTRODUCTION

PAGE 10

1.2.4 Masonic Promoting Independence Programme

The Masonic Promoting Independence Programme (PIP) is for older people who would not be able to maximise their potential for recovery within the average hospital stay. The Masonic Programme which was initially evaluated through the ASPIRE trial was Slow Stream Rehabilitation, However, approximately six months into the trial, there was a renegotiation of service specification between Hutt DHB and the Masonic Trust, the redeveloped service was re-titled the Promoting Independence Programme. Referrals could be generated from Medical Consultants (Private and Public), General Practitioners, Sigma (NASC) or other like referral agencies.

A key worker is assigned to each older person and their role is to initiate and co-ordinate that person’s pathway through the rehabilitation process. The team includes a Rehabilitation Co-ordinator who has overall responsibility for the team and is the first point of contact for client, their family / whanau and outside agencies. The team consists of Registered Nurses, Occupational Therapists, Physiotherapists, Speech therapist, Social Worker, Podiatrist, Dietician, Kaiawhina, designated Caregivers, Rehabilitation Assistants and a Rehabilitation Specialist / Geriatrician.

Older people are able to receive up to 12 weeks of facility based rehabilitation in the Promoting Independence Programme (not offered through the Masonic facility) if they are assessed as having high needs and are at risk of residential care or long-term hospitalisation. Clients who are assessed as having high or very high needs but are able to receive rehabilitation services in the community, or clients who have been discharged to the community from a residential facility may receive a monitored amount of input up to a maximum of one year from the health event. Alongside the rehabilitation programme offered by the Masonic, the team undertakes a comprehensive handover to designated home care providers that allows for an individually tailored education programme to be delivered to the formal and informal caregivers.

INTRODUCTION

PAGE 11

1.2.5 Summary

Clearly, there are common and unique features across all three initiatives. Table 1-1 below highlights the key areas of similarity.

Table 1-1: Ageing-in-place initiatives (Key features)

COSE, Canterbury

Masonic PIP, Lower Hutt

Community FIRST, Hamilton

Comprehensive assessment Yes Yes Yes

Led by: DHB NGO NGO

Site of delivery Person’s home Residential home Person’s home

Case Management Yes Yes Yes

Relationship with NASC Replaces NASC NASC controls

referrals NASC controls

referrals

Team constitution

One COSE, working alongside a number of

GP practices

Rehabilitation coordinator and MDT

(SLT, OT, PT and SW)

Coordinator (typically RN) OT, PT and Key

Support Workers

Relationship with Home Care

Refers to traditional Home care, does not deliver any disability

support services

Refers to traditional home care, but may provide education to the support workers

around the client

Fully replaces home care with a restorative home support service

Intensity of input

Initial assessment and ongoing episodic

review (for example 12 months and as

required)

Residential home placement for up to 12 weeks and following discharge, gradual

reduction in input with occasional contact for

up to 12 months

Up to four visits a day from key support

workers, face-to-face visit by Health professional

coordinator every two weeks until discharge (either from service, to

residential care or death)

Nature of Rehabilitation None

Slow stream residential

rehabilitation

Restorative (client goals leading a

functional rehabilitation approach

Type of client need

Generally older people, all SNL levels (but only level 4 & 5 in

the ASPIRE trial)

Generally older people,

SNL 4 & 5 Older people,

SNL 4 & 5

INTRODUCTION

PAGE 12

METHODS

PAGE 13

Chapter II: The ASPIRE evaluation

2.1 Introduction

Ageing-in-place as a model aims is to facilitate older people remaining living in the community for as long as possible, with residential care entry being delayed until the older person’s needs can no longer be safely met in the community. The individuals in the ASPIRE study are older people who have been assessed as having high (Support Needs Level 4) or very high and complex needs (Support Needs Level 5) which put them at risk of residential care entry. The individuals receiving community support packages from the ageing-in-place Initiatives (AIPI) will be compared to older people receiving NASC determined services or residential care. To demonstrate the effectiveness of the AIPI, both the quality of care and the cost of care must be examined. If the cost of care is decreased, but the level of quality in care delivered is less, then the AIPI is not a viable alternative for long-term care delivery. If the level of quality increases and the cost of care is significantly higher in the AIPI, then the AIPI may not be an affordable option for long-term care for funders to consider. It is also important for the study to consider the effects of AIPI on the primary informal carer, normally a spouse.

A randomised control trial is the best approach for measuring the effectiveness and cost-effectiveness of ageing-in-place initiatives for the following reasons.

1 Allow direct comparison between the AIPI and conventional health care delivery 2 Produce study groups that are comparable with respect to known and unknown

confounding factors 3 Remove investigator (selection) bias in assignment of older people to AIPI or

conventional health services 4 Guarantee that statistical tests will have valid significance levels (Weinberger,

Nagle et al. 1994)

A randomised controlled trial generates evidence typically classified as Level II as illustrated in Table 2-1. However, ASPIRE is a pooling of the randomised controlled trial evaluation of three ageing-in-place initiatives, or in other words, a meta-analysis. Because meta-analyses draw on different datasets, they generate highly reliable results and as a

METHODS

PAGE 14

consequence, the findings arising from ASPIRE can be considered as a minimum of Level II evidence.

Table 2-1: Level of evidence (modified from NHMRC, 2000)

Level of evidence

Study design

I Evidence obtained from a systematic review (meta analysis) of all relevant randomised controlled trials

II Evidence obtained from at least one properly-designed randomised controlled trial

III-1 Evidence obtained from well-designed pseudo randomised controlled trials (alternate allocation or some other method)

III-2 Evidence obtained from comparative studies (including systematic reviews of such studies) with concurrent controls and allocation not randomised, cohort studies, case-control studies, or interrupted time series with a control group

III-3 Evidence obtained from comparative studies with historical control, two or more single arm studies, or interrupted time series without a parallel control group

IV Evidence obtained from descriptive studies e.g. case series, either post-test or pre-test/post-test designs

The study is administered and managed by Auckland UniServices Ltd a wholly owned company of The University of Auckland; the conduct is supervised by the ASPIRE operations team, a group of researchers experienced in clinical trials and medical research.

2.2 Ageing-in-place initiatives to be evaluated in this study

The combined and individual effectiveness were examined in the following services:

The Coordinator of Services for the Elderly (COSE) model in Christchurch. The Presbyterian Support Northern (PSN) Community F.I.R.S.T. (Flexible

Integrated Rehabilitation Support Team) model in Hamilton The Wellington Masonic Villages Promoting Independence Programme (PIP) in

Lower Hutt

METHODS

PAGE 15

2.2.1 The usual care (Control) group

Services available to older people assessed with high and complex needs are numerous and include primary care (e.g. GP, Practice Nurses, Pharmacists), Disability Support Services (e.g. NASC, Home Care and Residential Care, Respite, Caregiver support), Health services (e.g. Geriatrician, Therapy, District Nursing), Community services (e.g. Lions, Rotary), Consumer groups (e.g. Age Concern, RSA) Regional council services (Citizen’s Advice Bureau, Housing), ACC (Tai Chi and Otago Exercise Programme), WINZ (Household support and funding).

Services such as these vary considerably according to geographical location and therefore usual care similarly varied across the three DHB regions. However, although there was considerable commonality, the unique features of the AIPI meant that what was delivered in usual care differed in certain key areas according to what services the AIPI offered or replaced.

COSE, in Canterbury as already discussed is an evolution of traditional NASC services and replaces NASC. Therefore, the usual care group includes NASC and is in direct comparison with COSE (The AIPI). However, both groups share all other services as described above. For instance older people randomised to the AIPI group would receive the same home care services or residential services as those randomised to the usual care group.

The Masonic PIP model is an ‘additional’ service to what is normally offered to older people and does not replace any services. PIP represents a slow stream rehabilitation service offered through residential care for a limited period of time and upon discharge home, the team provide education to the traditional home care service and enhanced case management. Older people in the usual care group however, would have been discharged directly to either their home or permanently to a residential home. The same NASC workers operate across the intervention and usual care group.

Community FIRST is an example of restorative home support and replaces traditional home care for older people and offers an alternative to residential care for older people choosing to remain living at home. NASC services are the same across usual care and AIPI and as the Community FIRST is funded through a bulk funding arrangement, other

METHODS

PAGE 16

aspects that would normally be funded through NASC are covered by this arrangement, such as caregiver residential respite and day care.

2.3 Study aims

The aim of this study is to assess the effectiveness and cost-effectiveness of ageing-in-place initiatives in assisting older people with high and very high needs to remain living in the community, preventing or delaying entry into permanent residential care and reducing mortality. This project seeks to evaluate the effectiveness and cost-utility of the following AIPI; the Presbyterian Support Community FIRST model operating in Hamilton, the Masonic PIP model in Lower Hutt and the COSE model in Christchurch.

2.4 Study objectives

1 To assess the effectiveness of AIPI, as compared to conventional home-based care in preventing (or delaying) the time before a community-based older person requires permanent residential care

2 To assess the effectiveness of AIPI in improving survival in community-based older people compared to conventional care

3 To determine the impact of the AIPI on an older person’s independence and health-related quality of life compared to similar measures in those receiving conventional care

4 To establish the degree of correlation between the expected improvement in the health-related quality of life of informal caregivers attributable to AIPI, in comparison to those receiving conventional care

5 To determine the cost-effectiveness of AIPI to the client, family, providers and funding agency in relation to the conventional care model

6 To identify the key elements of the AIPI healthcare models of community-based service delivery that lead to beneficial outcomes

2.5 Study design

2.5.1 Participating centres

AIPI in Hamilton, Lower Hutt and Christchurch are being evaluated using three separate randomised controlled clinical trials. Hamilton and Lower Hutt are models to improve the functional status (or independence) of older people and Christchurch is a case

METHODS

PAGE 17

management intervention. In Hamilton and Lower Hutt, the study population were recruited from older people who were being referred to a needs assessment and service co-ordinator (NASC). Randomisation occurred at the individual level, or in other words, every time an older person consented to enter the trial they were randomised to either AIPI or usual care. In Christchurch, General Practitioner practices were randomised to either receive a COSE case manager or usual care (standard hospital based NASC). The service that the older person received was dependent on which arm of the trial their GP was randomised to; COSE or usual care (standard NASC). On completion of the study, the three groups were pooled for analysis.

2.5.2 Criteria for eligibility

All older people who were referred to a NASC co-ordinator in Hamilton, Lower Hutt and Christchurch and who were under the care of a General Practitioner or a Geriatrician, were potentially eligible for inclusion in this study if they met the following criteria: These criteria were inclusive of the criteria for the AIPI being evaluated. Or in other words, the ASPIRE criteria matched the normal service eligibility criteria. The purpose for this is to ensure that ASPIRE mimics normal every day service activities rather than introducing a ‘true’ experiment and in doing so allows the findings to be more readily generalised to the wider population of older people.

METHODS

PAGE 18

Inclusion criteria

1 Males and females aged 65 or greater years on the day of baseline examination or 55+ Maori or Pacific Island or classified by NASC as ‘like age and interest’

2 All participants are assessed by the NASC co-ordinators or hospital clinicians as meeting the classification of High2 or very High3 (complex) health and disability needs

3 All participants or their main caregiver must provide and be capable of providing a declaration of ‘Informed Consent’

4 All participants must be English-speaking or provide a family member who can act as an interpreter

Exclusion criteria

If participation in the programme, in the opinion of the multi-disciplinary team, meant that there was an unacceptably high risk to the older person remaining at home, with multi-disciplinary care, such as high risk of injury from falls.

2.5.3 Study interventions

There are three models of AIPI that are being evaluated, namely;

The Presbyterian Support Northern (PSN) Community F.I.R.S.T. (Flexible Integrated Rehabilitation Support Team) model in Hamilton

The Masonic Promoting Independence Programme in Lower Hutt The Coordinator of Services for the Elderly (COSE) model in Christchurch.

2 High needs: The older person’s ability to remain in their environment is compromised due to significant safety issues and complex support needs. Outcome: The older person has access to safe environment and effective support; the carer has access to meaningful and practical support, enabling them to maintain their life roles; specialised assessment and treatment services are accessed; the carer is valued and supported. (Ministry of Health, 2002)

3 Very High Needs: Due to rapid deterioration the older person‘s support needs have significantly increased; current support is no longer effective; the safety of the older person and carer is at risk. Outcome: The older person is sustained by a support package; the carer is sustained by a support package; areas requiring specialist attention are addressed, i.e. reversibility and rehabilitation; longer term planning is underway (Ministry of Health, 2002)

METHODS

PAGE 19

2.5.4 Recruitment strategies

Community FIRST, Hamilton and Masonic PIP, Lower Hutt

1 Following NASC assessment and confirmation of entry criteria and needs level being confirmed as high or very high, the older person was asked whether they were interested in a person from The University of Auckland coming to speak to them about an evaluation of health services

2 If yes, NASC contacted the local research associate with older person contact details

3 If no, NASC collected demographic information

4 The research associate contacted and visited the older person and their caregiver to provide information and obtain consent

5 If yes to consent, research associate undertook baseline assessment of older person. The research associate left the questionnaires for the primary informal caregiver to complete for baseline assessment

6 If no, research associate requested demographics

7 The research associate undertook randomisation on their laptop computer

8 The research associate contacted NASC within a maximum of 24 hours of first contact to inform them which service the older person had been allocated to (new service or usual care)

Figures 2-1 and 2-2 highlight the recruitment process employed in the ASPIRE study for Hamilton and Hutt respectively.

METHODS

PAGE 20

Figure 2-1: Hamilton recruitment process

METHODS

PAGE 21

Figure 2-2: Hutt Valley recruitment process

METHODS

PAGE 22

The COSE initiative, Christchurch4

1 General Practices were randomised to receive either a COSE worker (1 COSE worker per 3000 older people (Greater than or equal to 65 years) living in the area or to receive NASC co-ordinated services

2 Any older person assessed by NASC or COSE as meeting entry criteria was asked if they were interested in a person from The University of Auckland speaking to them about an evaluation of health services

3 If yes, NASC or COSE contacted the local research associate with the older person’s contact details

4 If no, NASC or COSE collected demographic information

5 Research associate contacted and visited older person and caregiver to provide information and obtain consent

6 If yes to consent, the research associate undertook baseline assessment of older person. The research associate left the questionnaires for the primary informal caregiver to complete for baseline assessment.

7 If no, research associate requested demographics

4 The older person continued to receive NASC or COSE services if they refused participation in the ASPIRE study

METHODS

PAGE 23

Figure 2-3: Christchurch recruitment process

METHODS

PAGE 24

2.5.5 Randomisation assignment

Randomisation with equal probability to each of the trial arms was undertaken. A minimisation design was employed in Hamilton and Hutt whereby older people were stratified according to:

Disability (high or very high health and disability needs); age (below 74 years, 75 years and above); gender (male or female) and; living situation (living alone or with others).

This process meant that an equal balance of factors (variables) known to influence outcomes were balanced between randomised groups, thereby increasing the reliability of the study detecting a true effect of the AIPI services on outcome. For example, there were very similar numbers of males in the Masonic PIP programme in Lower Hutt as there were in the usual care group. This method is often used in studies that have smaller sample sizes as seen in the Hamilton and Lower Hutt sites; as if this process was not followed, an imbalance of certain variables such as gender may occur.

Christchurch employed a more complex randomisation process (block randomisation) as it was not possible to randomise at the point of NASC assessment as COSE workers were attached to G.P. practices. Therefore, the older person’s place of residence informed which group they would be assigned to. Consequently, two randomisation processes were utilised as described below.

Community First, Waikato and the Masonic Slow Stream Rehabilitation, Lower Hutt

Every older person was individually randomised by the research associate through a software programme installed on laptop computers. Minimisation methods were achieved through daily uplink to the Clinical Trials Research Unit server system.

Coordinator of Services for Elderly (COSE), Christchurch

COSE workers were assigned to G.P. practices according to a stratified cluster randomisation process and individual older people had their randomisation status pre-

METHODS

PAGE 25

assigned according to whether or not they were registered with a COSE-G.P. (active) or NASC-G.P. (control). Consent to participate as an active or control participant meant consent to participate in an evaluation of health outcomes over two years.

2.6 Outcome measures

The follow-up period lasted an average of 18 months. Data were collected at baseline, three, six, 12, 18 and 24 months post-randomisation. Please see Table 2-2 for the investigation schedule. The data were collected by research associates within each site.

2.6.1 Primary end-points

Combination of either death or permanent residential home admission

2.6.2 Secondary end-points

Survival Permanent residential home admission

Disability Quality of life Number of acute hospitalisations Number of falls Social support network Health-related quality of life of the primary informal caregiver Experience of the primary informal caregiver Costs (direct and indirect)

Tertiary end-points

Predictive modelling

METHODS

PAGE 26

Table 2-2: Investigation schedule

Time (month) T0 T3 T6 T12 T18

Older person

Demographics

MDS-HC

Social Networks

EuroQoL 5D

Caregiver

CRA

SF36

EuroQoL 5D5

2.6.3 Outcome measures of the older person

Survival is defined as the amount of time a participant lives, from registration until the time of death. The date of death was collected from the health service provider by the research associate. Administrative staff in each site also surveyed local newspapers for death notices and in addition, further confirmation was ascertained by access to various national datasets on completion of the study.

The entry of the older person into permanent residential care required the approval from the DHB and the notification of the NASC co-ordinator. Entry of the older person into permanent residential care was confirmed by the research associate contacting the NASC co-ordinator on a regular basis and further confirmation was ascertained using the national CCPS dataset on completion of the study.

5 The EuroQoL 5D is completed by the caregiver, but is the caregiver’s perception of the quality of life of that older person at that point in time. It is not a reflection of the caregiver themselves

METHODS

PAGE 27

The Minimum Data Set – Home Care (MDS-HC)

InterRAI, a United States (US) based not-for-profit organisation with membership from notable international gerontologists and clinicians, has developed and implemented a range of assessment tools for older people. The tools are now being widely utilised in the US, the United Kingdom (UK) and Canada as well as six DHBs across New Zealand6. Following a systematic review of the literature on comprehensive assessment instruments as part of the NZ Guidelines Group’s development of Guidelines for Assessment Processes for Older People, consideration is being given to the adoption of the InterRAI assessment tools within New Zealand. The first MDS assessment instrument to be developed was the MDS-RAI instrument for nursing homes. The MDS-RAI was developed in the USA and InterRAI has since developed other instruments for the assessment of older people in a variety of settings:

MDS-Acute Care (MDS-AC) MDS-Post Acute Care (MDS-PAC) MDS-Palliative Care (MDS-PC) MDS-Mental Health (MDS-MH) MDS-Home Care (MDS-HC) MDS-Self-reliance screener (MDS-Screener)

InterRAI (Carpenter et al, 2002) reports that all the tools have significant benefits in that;

(a) they facilitate the consistent and comprehensive assessment of older people; (b) the use of the tools support assessors to consider the whole person; (c) that care is based on accurate, reliable information; (d) the results of the assessment assist clinicians in identifying problems and potential

for improvement and; (e) that interdisciplinary staff involvement in assessment and care planning is

improved. Further, the data arising from the assessment facilitates the monitoring of indicators of quality of care, which in turn allows for evaluation of impact on case mix and resource management, and clinical effectiveness.

6 Waikato, Bay of Plenty, Hutt Valley, Capital and Coast and Canterbury DHBs are all using the MDS-HC as part of a trial implementation of the tool in New Zealand. The University of Auckland are involved in a MoH funded evaluation.

METHODS

PAGE 28

Services such as those being evaluated under ASPIRE often have numerous impacts on older people with complex needs and it was identified as extremely important that as much information as possible would be collected from the older person and their caregiver. Normal research practice until recently has been to utilise multiple tools to assess function, cognition, quality of life, satisfaction as well as a host of other areas. Even with this approach, there remains a distinct risk that important areas will be missed. Therefore, for the purposes of the ASPIRE study, the MDS-HC7 was adopted as the main assessment outcome tool. All MDS-HC assessments in the ASPIRE study were undertaken by a research associate, of whom all were experienced health professionals, trained specifically in the use of the MDS-HC. The MDS-HC is a comprehensive assessment instrument which allows assessment of multiple key domains such as function, health, social support and service use (Morris, 1996). Summary scales have been derived from the MDS items and validated in comparison with widely-accepted instruments (Mor et al. 1997; Hawes et al. 1995; Hartmaier et al. 1995). The scales explore each older person’s performance in basic and instrumental activities of daily living and in cognitive function (cognitive performance scale). The validity & reliability of MDS-HC is well-documented and is included in table 2-3 (Morris, Fries et al. 1997). Some of the studies that validate the scales against established scales use the scales within the MDS-RAI (Residential Assessment Instrument) and not the MDS-HC. However, the MDS-HC was developed as an extension of the MDS-RAI and the items have been shown to be equally applicable to the home care environment (Morris, Fries et al. 1997).

All research associates were trained in the use of the assessment tool and inter-rater reliability of assessments were established.

7 The ASPIRE study utilised the MDS-HC (version 2.03) measurement tool (UK Version)

METHODS

PAGE 29

Table 2-3: Summary of MDS-HC outcome scales validation

Outcome Scales Validated Against Activities of Daily Living scales (short & long forms)

Reliability (Morris, Fries et al. 1999) Barthel Index Scale (Landi, Tua et al. 2000)

MDS ADL Self-Performance Hierarchy Scale

Reliability (Morris, Fries et al. 1999)

The Cognitive Performance Scale

MMSE (Morris, Fries et al. 1994; Hartmaier, Sloane et al. 1995; Landi, Tua et al. 2000), Test for Severe Impairment (Morris, Fries et al. 1994; Hartmaier, Sloane et al. 1995)

The MDS Depression Rating Scale

15-item Geriatric Depression Scale (Burrows, Morris et al. 2000)

IADL Difficulty Scale and IADL Involvement Scale

Not validated

Changes in Health, End-stage disease and Signs and Symptoms (CHESS)

Compared CHESS score against survival time (Hirdes, Frijters et al. 2003)

Pain Scale Visual Analogue Scale (Fries, Simon et al. 2001)

The MDS-HC has various scales embedded within as listed below and described in full in Appendix II.

ADL Self-Performance Hierarchy Scale ADL Short Form Scale ADL Long Form Scale Instrumental activities of daily living (IADL) scales IADL involvement Scale The Cognitive Performance Scale (CPS) Depression Rating Scale (DRS) (Burrows et al, 2000) Changes in Health, End-stage disease and Signs and Symptoms (CHESS Pain Scale for the Minimum Data Set Index of Social Engagement (ISE)

METHODS

PAGE 30

Social support network

Social support network and network type of the older person is being ascertained by the use of the Practitioner Assessment of Network Type (PANT) instrument (Wenger 1994). Wenger has defined five different support network types and the differences between them are associated with the presence and availability of local close family, the frequency of interaction within the networks and the degree of involvement within the community (Wenger 1991). There is no evidence of validation or repeatability for this instrument. Wenger is currently validating the PANT instrument in a large study. The existence of the network typology constructed by Wenger has been independently verified by Litwin (2001).

Hospital admissions

The number of hospital admissions including visits to the emergency department will be obtained from the AIPI service provider and the DHB for 6 months before and 12 months after study entry. Encrypted NHI numbers will be used to access costs. The number of falls will be recorded into the MDS-HC by the research associate.

2.6.4 Outcome measures of the primary informal caregiver

The primary informal caregiver is defined as an adult family member, or another individual, designated by the older person as taking responsibility for their Health and Safety. Questionnaires for the primary informal caregiver are left with the caregiver at the older person’s residence to complete and to return by post in a prepaid envelope to The University of Auckland.

Medical Study Short Form (SF36®)

The Health related quality of life of the primary informal caregiver is being measured by the Medical Study Short Form (SF-36®) (Ware and Sherbourne 1992). This questionnaire is a single multi-item scale that assesses eight health concepts: physical limitations caused by health problems, limitations in social activities caused by physical or emotional problems, role limitations caused by physical health problems, and emotional problems, bodily pain, general mental health, vitality and general health perceptions. The SF-36 has been tested

METHODS

PAGE 31

Figure 2-4: SF36 Scales Measure Physical and Mental Components of health

extensively for reliability and validity (McHorney, Ware et al. 1992; Ware and Sherbourne 1992; McHorney, Ware et al. 1993; McHorney, Ware et al. 1994).

Figure 2-4 illustrates the taxonomy of items and concepts underlying the construction of the SF-36 scales and summary measures. The eight scales are hypothesized to form two distinct higher-ordered clusters due to the physical and mental health variance that they have in common.

Caregiver Reaction Assessment (CRA) The experience of the primary informal caregiver with health services, including perceived carer stress, is being measured by the Caregiver Reaction Assessment (CRA) Instrument (Given, Given et al. 1992). The CRA assesses specific aspects of the care giving situation, including both negative and positive dimensions of care giving reactions. The CRA contains 24 items and factor analysis provides five subscales, namely, caregiver’s esteem, lack of family support, impact on finances, impact on schedule and impact on health. Respondents use a five-point Likert scale. CRA has been used and validated with caregivers of older people with physical impairments, and Alzheimer’s disease and with caregivers and partners of cancer patients (Given et al. 1992; Nijboer, Triemstra et al. 1999). The CRA was chosen for this study, as it has been utilised in caregivers of older

METHODS

PAGE 32

people with both physical and mental impairments and addresses the positive aspect of care-giving. Table 2-4 highlights this tool.

Table 2-4: Caregiver Reaction Assessment Instrument

Item Question 1. *I feel privileged to care for ___. 2. Others have dumped caring for ___ onto me. 3. *My financial resources are adequate to pay for things that are required for care giving. 4. My activities are centred around care for ___. 5. Since caring for ___, it seems like I'm tired all of the time. 6. It is very difficult to get help from my family in taking care of ___. 7. I resent having to take care of ___. 8. I have to stop in the middle of work. 9. *I really want to care for ___. 10. My health has gotten worse since I've been caring for ___. 11. I visit family and friends less since I have been caring for ___. 12. *I will never be able to do enough care giving to repay ___. 13. *My family works together at caring for ___. 14. I have eliminated things from my schedule since caring for ___. 15. *I have enough physical strength to care for ___. 16. Since caring for ___, I feel my family has abandoned me. 17. *Caring for ___ makes me feel good. 18. The constant interruptions make it difficult to find time for relaxation. 19. *I am healthy enough to care for ___. 20. *Caring for ___ is important to me. 21. Caring for ___ has put a financial strain on the family. 22. My family (brothers, sisters, and children) left me alone to care for ___. 23. *I enjoy caring for ___. 24. It's difficult to pay for ___’s health needs and services.

Coding Scheme: 1. = Strongly disagree 2. = Disagree 3. = Neither agree nor disagree 4. = Agree 5. = Strongly agree * These questions are reverse scored.

METHODS

PAGE 33

2.7 Determination of the cost-effectiveness of AIPI and conventional health services

In addition to determining whether the AIPI has resulted in improved health status among the targeted populations, the study also seeks to ascertain whether the initiatives represent good value for money. This is an essential component of the evaluation as it speaks to the issue of whether the initiatives are sustainable in the long run. There are several possible outcomes. For instance, the initiatives could result in improved health outcomes and reduced costs. This would be the best possible outcome as it would suggest the initiatives are both technically and allocatively efficient. Another possibility is that the initiatives result in better health outcomes, but at an additional cost to one or more of the parties (e.g., health services, families or caregivers). In this event, the health gains must be placed in some context that enables decision and policy-makers to ascertain whether the gains justify the additional expense.

The cost-effectiveness analysis examines both the cost and health outcomes from the AIPI and conventional services. The work will include:

Identifying the average and marginal cost of providing the services at the locations, Identifying the direct and indirect costs associated with the initiatives and with

standard care, Identifying the cost-effectiveness of the initiatives.

NOTE. The results of this analysis will form the basis of a further report and will not be included in this document.

2.8 Older People Entering Residential Accommodation (OPERA)

OPERA is a sub-study of ASPIRE and focuses on the qualitative aspects of residential home admission. The objectives of the OPERA study were to explore, describe and interpret:

the factors which led the older person to enter residential care; who were the decision-makers for the residence of the older person; the older person’s feelings about the residence decision (either residential care, or

their own home).

METHODS

PAGE 34

The study was conducted in two phases: pilot and the main study. Both involved semi-structured interviews of older people (and their caregivers where possible) who had recently been assessed as needing high or very high levels of support. This level of support was consistent with the requirements for entry to residential care.

2.8.1 Pilot study

The pilot, consisting of 13 older people and who were not part of the ASPIRE study included interviews with six people who had recently been admitted to residential care, six primary caregivers, a Needs Assessment Service Co-ordination team (N.A.S.C.), and a multi disciplinary hospital based team (MDT). Data gathering was undertaken through interview using a set of open-ended questions with only pre-arranged prompts being given when necessary, which decreased the likelihood of any potential biases. The interview with the older person and caregiver lasted approximately 45 minutes and were related to the older person’s residence decision, the processes and the support available to them. The interview questions with the NASC, and MDT were general questions concerning who made the decisions regarding the residence of an older person and their processes for the decision. Also included were questions regarding residence decisions in regard to some specific older people.

2.8.2 Main study

The main study, consisting of 131 older people who were all enrolled in the ASPIRE study, included a total of 28 participants who were interviewed in residential care, with the remainder being interviewed at home. The participants where possible, had two interviews at an interval of six months. Also interviewed were 24 caregivers and an NASC team who discussed 12 of the older people recently admitted to residential care. The main study participants were invited to participate from the ASPIRE trial in the three centres, Hamilton, Lower Hutt and Christchurch. The interviews were carried out face-to-face in residential care, and in the older person’s home, or by telephone. A sequential mixed methods study for data collection was selected for this phase because of the multiple approaches to data collection, analysis and inferences employed in the sequence of events. Data arising from the audio-recorded interviews were transcribed verbatim to ensure accuracy of intent.

METHODS

PAGE 35

A general inductive approach methodology was used throughout with the intention of:

(a) finding common themes from the pilot, to use in the development of the questionnaire for the main study and;

(b) forming an understanding of the factors and decision makers for the residence decision, and the older person’s ultimate satisfaction with that decision.

The qualitative data analysis programme QSR NVivo was used to unitise and categorise the data. This process facilitated the development of ‘meta-themes’ representing the range of experiences described within the study. Inter-rater reliability was tested by transcripts being given to two independent senior researchers to code. The codes were compared and discussed. Four different triangulation types were used to strengthen this study as follows:

1. data triangulation; data from a variety of different people, 2. investigator triangulation, where several different researchers were used, 3. theory triangulation, where the data was looked at from multiple perspectives

within the ASPIRE trial and the OPERA study; and 4. methodological triangulation using both qualitative and quantitative methods

2.9 Ethical considerations

Ethical approval was granted from the lead ethics committee (Auckland) and was submitted to the ethics committees in Hamilton, Lower Hutt and Canterbury in July 2003 (Reference No: AKX/03/07/177) and for OPERA as an amendment to the ASPIRE trial by the lead Ethics Committee, Auckland on the 14th October 2004 (AKX/03/07/177 PIS/Con V#2,5/09/03). Approval for the trial was also gained from the managers of NASC and providers. Informed consent was gained by a signed consent from the older people and caregivers. Confidentiality and anonymity was maintained at all times.

2.10 Adverse event reporting

All of the older people enrolled in this study were under the clinical care of their G.P. and the health services being evaluated. The trial was evaluating health services; therefore the intervention does not directly increase the risk of adverse events. If the research associate found during the collection of data that the older person required medical or social assistance, then she notified the ASPIRE project manager who immediately notified the

METHODS

PAGE 36

health service. In emergency situations, the research associate called an ambulance and notified the ASPIRE project manager. However, adverse events may result due to the health service itself, and as such under Adverse Effect (AE) and Serious Adverse Effect (SAE) definitions, processes have been adopted to ensure that such events are monitored in an appropriate manner.

2.10.1 ASPIRE interim reporting

At six monthly intervals throughout ASPIRE, the study statistician and Associate Professor Philippa Poole, Faculty of Medical and Health Sciences, The University of Auckland reviewed the un-blinded adverse event data from the study. Professor Poole is an independent medical clinician with extensive experience in geriatrics. Adverse events which were reviewed included:

Death Life threatening events Permanent or substantial injury Unplanned hospitalisation or prolongation of hospitalisation Falls with serious injury resulting in medical attendance by a GP or Hospital Medically important illnesses were coded as infectious, cardiovascular,

orthopaedic, neuro-psychiatric or other

All treatment comparisons were performed by intention to treat8 and by trial and then pooled in a meta-analysis. Any differences between treatment groups for these outcomes were deemed to reach statistical significance if the p-value was less than 0.003 or 3 standard deviations. Additional data that was monitored included the rate of the primary end-point of institutional-free survival in the control group and the coefficient of variation of true proportions between clusters within each group (k) in the Christchurch study. These two measures have been estimated to be 0.35 and 0.3, respectively, in the sample size

8 a type of analysis of clinical trial data in which all patients are included in the analysis based on their original assignment to intervention or control groups, regardless of whether participants failed to fully participate in the trial for any reason, including whether they actually received their allocated treatment, dropped out of the trial, or crossed over to another group.

METHODS

PAGE 37

calculation. Consideration has been given to increasing the sample size if the control rate of institutional-free survival is lower than 0.35 or if k is greater than 0.3.

On the basis of the results of each interim analysis, Professor Poole had the ability under the respective Ethical Committees to advise the trial operations committee to continue on as previous for the proceeding six months, or contact the DHB to modify the health service or consider stopping the delivery of the health service, or modify the study protocol or stop the evaluation. During the course of the study, there was no reason for Professor Poole to contact the Operations Committee to modify the service.

2.11 Study definitions

2.11.1 Institutionalised-free survival

Institutionalised-free survival will be defined as participants without permanent admission into residential care or death from the Date of Assessment. Therefore institutionalised-free survival is defined as the number of days from the randomisation date until either of permanent residential care admission or death, whichever is come first.

2.11.2 Participant withdrawal and lost to follow-up

A withdrawal is defined as a withdrawal of consent to participate post-randomisation. Lost to follow-up is defined as participants who were alive at their last contact, but have became unavailable for further contacts. The Research Associates will request the opportunity to ask some questions on a six-monthly basis to ascertain if they are alive and where they are living. The older person or primary informal caregiver has the right of refusal. Participants who withdrawal / lost to follow-up from the study will be assumed to be censored unless the death of participants are confirmed.

Time to Death or permanent residential care = (Date of the Death, Date of permanent residential care admission minus Date of

Randomisation)

METHODS

PAGE 38

2.11.3 Censoring date

If the participants are still alive at the end of study then the censoring date is the date that study ends. If the subjects are withdrawal or lost to follow-up the censoring date is the date of last contact. Censoring date is not applicable to subjects who died or entered into permanent residential care. Censoring days (only for subjects who are event free) is defined as:

2.11.4 Changes from Baseline

Absolute change from Baseline is defined as:

2.12 Statistical issues

2.12.1 Sample size and cluster size calculations

The ASPIRE meta-RCT design provides an efficient way to evaluate the effects of the AIPI models of care on a substantive outcome that would not be possible in a single RCT. Consideration was given to running a single, multi-centre trial; however this was not feasible for two main reasons. Firstly, the interventions in each of the centres are similar but not the same. Secondly, it is not feasible to randomise individuals in Christchurch hence GP practices will form the units of randomisation in Christchurch. Standard criteria and measures will apply across all trials so it is anticipated that heterogeneity will be kept to a minimum. The following sample size calculation is based on the meta-RCT.

A total sample size of 830 patients provides a power of 90% with a two-sided α = 0.05 to

detect a 30% relative risk reduction in the primary end-point of institutional-free survival between groups, assuming 35% frequency of death or residential care in controls. A total sample size of 830 would also provide almost 80% power to detect a 25% relative risk reduction. This sample size is sufficient to allow analysis of effects of the different AIPI on secondary and tertiary endpoints as well as the primary endpoint. 830 subjects from three

(Date of last contact or Date of the study ends Date of Randomisation)

(Post Baseline Value – Baseline Value)

METHODS

PAGE 39

trials equates to 277 from each trial. However, the 277 subjects required in Christchurch needs to be increased to take account of the clustering. To do this some assumptions are required.

Since Christchurch centre had a randomisation allocation at the GP level, 55 GPs were randomly assigned to one of the 2 intervention groups and all older people within each GP received the same intervention. Typically, individuals within a Practice are more similar than those from different Practices. In statistical terms, observations within a cluster are correlated and this lack of independence must be taken into account when it comes to analysing the data. This is commonly referred to as a design effect (DEFF) and will almost always be greater than 1 indicating that more patients are required to effectively achieve the same sample size as that from the standard trial, which randomises at the patient level. As explained above, the DEFF increases as the within-cluster correlation and the number of participants within a cluster increase. For example when Design Effect equal 2, has double the variance and half the effective sample size compared to the study without clustering.

For the purpose of the setting the total sample size to allow 90% Power when all 3 centres are combined, we have assumed that the DEFF is equal to 2 (cluster size=11, ICC=0.1). This gives us the anticipated total sample size required of 1109 patients.

2.12.2 Amendments to the protocol

Sample size was revised several times, on the basis of event rates also revised due to inappropriate assumptions on initial statistical tests. Throughout the study, the event rate was periodically monitored in order to revise the sample size for the study. Survival models (exponential and weibull) were applied to time-to-event data to estimate the 18 month institutionalised-free survival rates.

The estimated parameters were then used to predict event rate at the end of study. The general study design, and analysis approaches remained unchanged. The schedule of visits was changed to Baseline, 3-months, 6-months (short version), 12-months and 18-months. As a consequence of time and financial considerations, it was decided to end recruitment short of the target sample size.

METHODS

PAGE 40

2.12.3 Statistical analysis

Procedures of the statistical analysis system SAS (SAS Institute Inc. Cary NC) SPLUS and Acluster software was used in all analysis. All statistical tests were two-tailed and a 5% significance level maintained throughout the analyses. It is generally considered that a 5% level of significance is of statistical interest and was employed in ASPIRE. However, more pertinent to the real life evaluations is a level that can describe a clinical difference (e.g. moving from a participant requiring supervision with meal preparation ‘score 2’ on the ADL scale to setup help only ‘score 1’). Despite this, clinical significance is often difficult to delineate. Stratford et al (1996) stress that clinical significance can be likened to “beauty” in that “clinical significance is often unique to the beholder” (pg. 1119). McMurdo and Rennie (1994) when investigating the influence of exercise on institutionalised frail older people concluded that a 10% improvement in flexibility or handgrip strength is of clinical interest. Nonetheless, researchers would agree (Stratford et al, 1996) that patients, practitioners and economics all feasibly play a defining role. This pertinent issue will be discussed further in the chapters III and IV.

All treatment evaluations were performed on the principle of ‘Intention to Treat’. No adjustments for multiplicity were undertaken for the secondary endpoints, adverse events, or other endpoints. The study design of ASPIRE was to combine three trials by using Meta-analysis to give an over treatment effect across centres. The study was powered based on the Meta-analysis, and not on any of the individual trials alone. However, analysis proceeded initially by exploring each centre separately and then pooling the results together. The principle of Meta-analysis is to increase power to detect an overall treatment effect as well as investigate the amount of variability between studies.

2.12.4 Reporting of intra-cluster correlation coefficient (ICC)

As the Intra-cluster correlation coefficient plays a key role in any clustered randomised trial, it is essential to estimate the ICC, to allow for clustering in the analysis, as well as report as outlined in the extension to the consort statement. There are many methods available to estimate ICC, however, for the purposes of this study, ANOVA was employed to estimate ICC (Shrout and Fleiss, 1979). In addition, this result was confirmed by utilising the

METHODS

PAGE 41

Acluster software, which is the software for design and analysis of clustered randomised trials (Donner, 2000) (See Appendix 1 for more information).

2.12.5 Primary endpoint

As defined above, institutional-free survival was the primary outcome for this analysis. Participants could only experience the primary outcome once. The impact of treatment on binary time-to-event outcomes was determined using standard log rank survival tests and adjusted analyses were performed when imbalances in important baseline characteristics were observed. As there were several variables that were believed to contribute to survival, a Multivariate Cox Proportional Hazards Model approach was used. When the Hazard Ratio (HR) is greater than 1, it means the hazard of new service is greater than the usual care. Or in other words, if the HR = 1.50, then the hazard of the new service is 1.5 times higher compared to the usual care. Or one can say that the chance of reaching the primary endpoint (residential home admission or death) is 50% higher in the new service compared to the usual care. Hazard Ratio analysis was also employed to explore the relationship between various variables and residential home admission and death.

Stratification Factors (age, sex, needs level and home alone) was also included in the incorporate model together with treatment effect. A Meta-analysis was then used to pool the Hazard ratios from three centres to provide an overall estimate.

2.12.6 Secondary endpoints

Ninety-five percent confidence intervals with no adjustment for multiple comparisons were calculated and will be presented for all secondary outcomes. Confidence intervals provide a range in which the true value is likely to occur 95% of the time, which is better than reporting P values that relate to a single test. Secondary outcomes such as MDS Scale measurements, Quality of Life, Acute Hospitalisation and falls are measured at the baseline, 3-month, 6-month, 12-month and 18-month. Analysis on secondary outcomes will be undertaken according to two scenarios:

1. Partial dataset (removing data after an older person has entered residential care):

This allows for an exploration of the impact of Community FIRST and PIP in

METHODS

PAGE 42

comparison to usual home based services. It has disadvantages in that if one group has more observation points than the other (i.e. if there is an imbalance in the admission to permanent residential care between the groups) then the results are based on a smaller sample size.

2. Complete dataset (including data after an older person has entered residential care)

This approach draws on more data as all subjects are included and therefore results are more statistically sound.

Both methods have a role in interpreting the findings and both will be presented and scenario 2 is used for subgroup analysis. Summary statistics (i.e. mean, standard deviation, n, or median, inter quartile range [IQR]) on baseline data and follow-ups will be showed in findings and in Appendices where relevant.

The total number of follow-up assessments was dependent on when an older person was registered into the study, whether the person died, the date of their last visit, and whether they were lost to follow-up before the end of the study. Therefore, the number of visits with measurements on the secondary endpoints varied across participants. As well as for many participants, data were missing for other reasons, e.g. incomplete assessment made. Thus, to properly carry out an Intention to Treat (ITT) analysis, we will need to deal with missing data. A mixed model with repeated measures will be used to allow every participant to be included in the analysis, and allow estimation of the treatment effect on each of the secondary endpoints.

To allow for correlation between measurements at different time periods on the same participant, covariance pattern models will be considered. There are 8 common Covariance Pattern includes Unstructured, First Order Autoregressive, Compound Symmetry, Toeplitz, Heterogeneous Uncorrelated, Heterogeneous Compound Symmetry, Heterogeneous First Order Autoregressive and Heterogeneous Toeplitz. However, only the first four covariance patterns will be tested in the analysis. Different covariance patterns will be tested to choose the appropriate covariance pattern. Likelihood ratio tests or AIC can be used for statistical comparisons between models.

METHODS

PAGE 43

2.12.7 Subgroup analysis

Subgroup analyses was conducted on data categorised according to Needs Level (High or Very High), CHESS score, age group (65-75, 75-85, 85+), Sex, Ethnicity, presence or absence of dementia and have/not have caregiver if data is permitting. Centre analysis was performed at the subgroup level and then pooled by the Meta-analysis as described above.

2.12.6 Per-protocol data set

A per-protocol analysis was also performed in order to check the robustness of the results if more than 5% cross-over cases in the study was identified.

2.12.9 Missing values and outliers

Missing data on primary outcome should not be permitted, as the information on primary end-points are obtainable by using hot-pursuit case finding method (i.e. Death Certificate, Institutionalisation records of residential care). Outlying observations in the data set were identified and clear identification of an outlying observation was made on the basis of clinical as well as statistical grounds. A sensitivity analysis was performed in order to assess the impact of outlying observations. Handling of missing data for the secondary endpoints was handled via the use of mixed modelling techniques as described above.

2.12.10 Predictive modelling analyses

Hazard Risk Ratios estimates were calculated and adjusted for age, sex, ethnicity and treatment effect with 95% Confidence Intervals and p-values for both Residential Care Entry and Hospitalisation as the Primary end-points. The technique is described above. A risk with predictive modelling is to introduce error by excessive searching for risk factors. Bland and Altman (1995, pg. 170) argue that a large number of statistical tests are difficult to interpret

“…because if we go on testing long enough we will inevitably find something which is significant. We must be aware of attaching too much importance to a

METHODS

PAGE 44

lone significant result among a mass of non-significant ones. It may be the one in 20 which we would expect by chance alone.”

Thus, a more sensible approach is to use literature to inform the search analysis pathway. Hirdes et al (2004) developed the Home Care Quality Indicators (HCQI) as a means to assess quality of home based organisations and further to introduce the potential to benchmark services. The HCQI are drawn from pertinent scales and questions in the MDS-HC and utilise CAPS in order to gather benchmarking information around the service. In short, the HCQI are new tools providing a first step along the path of quality improvement for home care. These indicators can provide high-quality evidence on performance at the agency level and on a regional basis. The HCQI (Table 2-5) were used for predictive modelling analysis in ASPIRE and in addition, the EuroQoL VAS and Caregiver Reaction Assessment summative scores were used.

METHODS

PAGE 45

Table 2-5: Inter-RAI Home Care Quality Indicators for Minimum Data Set—Home Care Version 2.0

HCQI Numerator Denominator Prevalence HCQIs Inadequate meals Clients who ate 1 meals in 2 of last 3 days All clients

Weight loss Clients with unintended weight loss All clients, excluding clients with end-stage disease on initial assessment

Dehydration Insufficient fluid intake All clients

No medication review by MD

Number of clients whose medications have not been reviewed by a physician within last 180 days

Clients who are taking at least 2 medications

Difficulty in locomotion and no assistive device

Clients with impaired locomotion who are not using assistive device

All clients with impaired locomotion on most recent assessment (excludes clients for whom indoor locomotion did not occur)

ADL / rehabilitation potential and no therapies

Clients are not receiving OT, PT, or exercise therapy

Clients who trigger CAP for ADL/rehab potential

Falls Number of clients who record a fall on follow-up assessment

All clients not completely dependent in bed, mobility on previous assessment

Social isolation with distress

Clients who are alone for long periods of time or always and they also report feeling lonely or clients who are distressed by declining social activity

All clients

Delirium

Clients with sudden or new onset/change in mental function or clients who have become agitated or disoriented such that their safety is endangered or client requires protection by others

All clients

Negative mood Any client with sad mood on most recent assessment and at least 2 symptoms of functional depression are exhibited up to 5 days/week or daily or almost daily

All clients

Disruptive / intense daily pain

Clients having daily pain and intense pain or pain disrupts activities All clients

Inadequate pain control

Clients who have pain and are receiving inadequate pain control

All clients having pain on most recent assessment

Neglect or abuse

Clients who have unexplained injuries, have been abused or neglected All clients

Any injuries Clients with fractures or unexplained injuries All clients No flu vaccination

Clients who have not received influenza vaccination within the last 2 years

All clients excluding clients receiving chemotherapy/radiation therapy

Hospitalisation Clients who have been hospitalized, visited hospital emergency department, or received emergent care since last assessment

All clients

METHODS

PAGE 46

Table 2-5: Inter-RAI Home Care Quality Indicators for Minimum Data Set—Home Care Version 2.0 (continued)

HCQI Numerator Denominator Failure to improve/incidence HCQIs

Bladder incontinence

Clients who have experienced decline in bladder continence performance between previous and most recent assessment or clients who have developed new bladder continence problem

All clients with at least one reassessment

Skin ulcers Clients with ulcer on previous assessment who did not improve or clients with new ulcer on follow-up

All clients with at least one reassessment

ADL impairment

Clients with some impairment on ADL Long Form who failed to improve between previous and most recent assessment or clients who have new ADL impairment based on ADL Long Form

All clients with at least one reassessment who are not palliative on initial assessment

Impaired locomotion in home

Clients who fail to improve in locomotion in home or clients who have new impairment in locomotion in home

All clients with at least one reassessment who are not palliative on initial assessment

Cognitive function

Clients who have experienced decline in cognitive performance between previous and most recent assessment or clients who experience new cognitive impairment

All clients with at least one reassessment

Difficulty in communication

Clients with both failure to improve in communication/making self understood and failure to improve in ability to understand others or clients with new difficulties in making self understood or understanding others

All clients with at least one reassessment

2.13 Summary

ASPIRE is a large and complex study aimed to fully explore the impact of the new ageing-in-place initiatives on a host of outcomes for the older person and their informal caregiver. In essence, ASPIRE consists of three smaller evaluations of which when drawn together (meta-analysis) becomes a very powerful means to fully explore the impact of services and indeed ageing in a group of frail older people.

METHODS

PAGE 47

FINDINGS

PAGE 48

Chapter III: Findings

3.0 Introduction

This chapter has five distinct sections:

Section 1 relates to the study sample characteristics within the usual care and intervention groups for each of the three sites;

Section 2 highlights the primary results, which is the impact that the new ageing-in-place initiatives have on mortality and residential home admissions of the older people in the study;

Section 3 explores the secondary findings results, namely the ADL, IADL scales as well as pain, frailty, quality of life and depression of the older person. In addition, the quality of life and perceived burden of care is explored for the caregiver;

Section 4 presents the findings from the predictive modelling exercise, whereby the MDS-HC Home Care Quality Indicators are assessed in relation to their impact on residential home placement and hospitalisation;

Section 5 presents the qualitative results of OPERA (Older People Entering Residential Accommodation).

Findings will be presented using a combination of tables, plots and commentary. Where further commentary is required, this will be presented in the final Chapter of this document. It is noted that for simplicity, not all available data has been presented in this report and the authors acknowledge that considerable more analysis of the very large dataset is possible. In particular, the analysis around informal caregivers has not been presented in this document.

SAMPLE CHARACTERISTICS FINDINGS

PAGE 49

Section 1: Sample characteristics

3.1 Introduction

A significant amount of information was collected for both the older person and their informal caregiver at baseline (the first assessment point). This section highlights this data for multiple domains, obtained in the main through the MDS-HC. Data for older people in both the usual care and the new intervention groups are illustrated for each site and are presented though total numbers, percentages, medians, means and standard deviations (SD)9.

3.2 Recruitment

Table 3-6 illustrates the numbers of older people recruited to the study. Please note that 569 participants were registered and randomised into the ASPIRE trials, but data were

9 SD is the measure of dispersion around the mean. In a normal distribution, 68% of cases fall within one SD of the mean and 95% of cases fall within 2 SD. For example, if the mean age is 45, with a standard deviation of 10, 95% of the cases would be between 25 and 65 in a normal distribution

In total, there were 569 participants randomised in the ASPIRE trial.

113 older people participated from the Hamilton region of which, 57

participants received usual care, and the remainder received

Community FIRST. A total of 53 received usual care and 52 received

PIP in Lower Hutt. In Christchurch, 182 received usual care and 169

received COSE.

The level of baseline disability observed in each of the three sites

varied considerably with older people in Christchurch being of far

lower disability than those assessed using the same NASC criteria in

Hutt and to a much greater extent in Hamilton.

SAMPLE CHARACTERISTICS FINDINGS

PAGE 50

collected on only 567, as two participants reached the primary endpoint of residential care placement after randomisation, but prior to the collection of baseline data.

Table 3-6: Study recruitment Total Hamilton Lower Hutt Christchurch

N % N % 3 N % 3 N % 3

Number referred from NASC 1343 254 275 817

Screened1 985 73.34 251 98.82 275 100 459 56.18 Declined1 416 30.98 138 54.33 170 61.82 108 13.22

Registered2 569 113 105 351 Usual Care 2 292 51.32 57 50.44 53 50.48 182 51.85 New Service2 277 48.68 56 49.56 52 49.52 169 48.15

% of total enrolled 100 19.86 18.45 61.69

1 Percentages are calculated base on total number of Referrals 2 Percentages are calculated based on total number of Registered 3 Percentages are calculated based on total number of Registered participants in each centre

The numbers of older people who entered the study were fewer than anticipated. There were several reasons for this. Firstly, although the number of refusals (31% across all sites) was within normal parameters for a study of this kind; the numbers of older people refusing to participate in Hamilton and to a greater extent, Hutt were particularly high. These refusals occurred in the main at the point at which the NASC team were requesting whether the older person would agree to their name being passed to the researchers. A considerable amount of effort was undertaken by the ASPIRE team to facilitate this process. A further point is the dominance of Christchurch (61.7%) in the sample and further, the low sample size in Hamilton and Hutt means that it will always be difficult to establish statistically significant effects of the services.

3.2.1 Protocol violations

A protocol violation is a serious non-compliance, which may lead to exclusion of subjects from eligibility analysis and / or their discontinuation from the study (EFCCP, 2001). The

SAMPLE CHARACTERISTICS FINDINGS

PAGE 51

table below presents the cause and frequency of protocol violations in ASPIRE. The numbers are low and all were withdrawn from the sample but were included in the analysis as intention to treat (See Section 2.12)

Table 3-7: ASPIRE Protocol violations Number %

Receiving different treatment rather than the assigned treatment 7 1.23 Re-randomised 1 0.18 Become non eligible 1 0.18 No informed consent prior randomisation 1 0.18 Randomised twice 1 0.18 TOTAL 12 2.13

3.2.2 Completeness of information and treatment allocation

Table 3-8: Recruitment and enrolment over time

Usual Care AIPI Total Forms completed without outstanding queries N* % N* % N* % Older Person (OP) Questionnaire

Baseline 291 100 276 100 567 100

3-months follow-up 244 83.85 248 89.86 492 86.77

6-months follow-up 217 74.57 218 78.99 435 76.72

12-months follow-up 165 56.70 177 64.13 342 60.32

18-months follow-up 68 23.37 84 30.43 152 26.81

Caregiver (CG) Questionnaire

Baseline 152 100 131 100 283 100

3-months follow-up 122 80.26 116 88.55 238 84.10

6-months follow-up 101 66.45 96 73.28 197 69.61

12-months follow-up 70 46.05 68 51.91 138 48.76

18-months follow-up 27 17.76 34 25.95 61 21.55

Number of Missing Forms (OP) 5 0.51 3 0.30 8 0.40

Number of Missing Forms (CG) 2 0.42 7 1.57 9 0.98

SAMPLE CHARACTERISTICS FINDINGS

PAGE 52

The recruitment into ASPIRE was staggered, that is with every new referral to NASC for assessment became a potential candidate and as this is a real life experiment, it took 12 months for Christchurch and Hutt to recruit this sample and Hamilton a further six months. Consequently, as the study time period was 24 months, the numbers of older people who had 18 month follow-up were relatively few and as a result most analysis has been undertaken on the 12 month follow-up data.

The table below illustrates the same information though by site and it is very apparent that the 18 month follow-up data is very minimal in Hamilton and Hutt. A further point concerning Hamilton, as the recruitment into the trial proved difficult, the majority of referrals came about in the last few months of the 12-month recruitment phase, which explains the higher number of subjects at the 3-month follow up point as opposed to 12 months and 18 months. Clearly, this has ramifications, which are discussed later.

Table 3-9: Number of participants where secondary data was collected at each time point

All Hamilton Lower Hutt Christchurch

All Hamilton Lower Hutt Christchurch Usual

Care New

Service Usual Care

Com FIRST

Usual Care PIP Usual

Care COSE

Baseline 291 273 57 54 53 50 181 169

3 months 229 233 34 45 35 37 160 151

6 months 203 207 25 32 30 32 148 143

12 months 148 157 12 15 19 26 117 116

18 months 57 76 5 7 6 7 46 62

Over time, there is a clear reduction in the numbers of older people enrolled in the study. Table 3-9 excludes those who have entered residential care or have died. It is possible to crudely observe the impact of the services on the older people enrolled in the AIPI from this table.

SAMPLE CHARACTERISTICS FINDINGS

PAGE 53

3.2.3 Older person demographics

This section relates to the extensive information gained from the older person at baseline. It serves several purposes, firstly to allow readers to explore the relative generalisability of the findings to their own DHB and for a description of the sample within each location (Table 3-10) and as a whole (Table 3-11).

Interestingly, there appears to be differences between sites, with older people in Christchurch being younger, female and more likely to be living in a house (rented or owned), where as almost half of the Lower Hutt PIP participants were male. Hamilton participants tended to be more ethnically diverse, were older and were less likely to own their own homes and were overall more highly educated, followed by Christchurch and Hutt.

SAMPLE CHARACTERISTICS FINDINGS

PAGE 54

Table 3-10: Older person demographics across each of the three sites Hamilton Lower Hutt Christchurch

Usual Care

New Service

Usual Care

New Service

Usual Care

New Service

N* %2 N* % 2 N* % 2 N* % 2 N* % 2 N* % 2

Age 55-64 65-69 4 7.0 3 5.5 2 3.8 1 1.9 9 4.97 10 5.9 70-74 4 7.0 4 7.3 4 7.6 8 15.4 22 12.2 27 16.0 75-79 6 10.5 10 18.2 8 15.1 13 25.0 47 26.0 34 20.1 80-84 16 28.1 10 18.2 11 20.8 11 21.2 45 24.9 46 27.2 85+ 27 47.4 28 50.9 28 52.8 19 36.5 58 32.0 52 30.8 N* 57 55 53 52 181 169 Mean* 83.5 82.6 83.6 81.9 81.1 80.9 SD* 7.59 7.29 6.9 6.8 6.70 7.1 Median* 84 85 85 82.5 81 81

Education Primary school 22 38.6 17 30.9 23 43.4 13 25.0 57 31.5 47 27.8 High school 25 43.9 25 45.5 23 43.4 34 65.4 93 51.4 90 53.3 Tertiary 10 17.5 12 21.8 7 13.2 3 5.8 31 17.1 32 18.9 Unknown3 - - 1 1.8 - - 2 3.9 - - - -

Gender Female 35 61.4 33 60.0 30 56.6 27 51.9 123 68.0 120 71.0

Ethnicity NZ European 43 75.4 43 78.2 49 92.5 47 90.4 159 87.9 154 91.1 NZ Maori 2 3.5 - - 2 3.8 3 5.8 1 0.6 2 1.2 Pacific - - 1 1.8 1 1.9 1 1.9 - - - - Other 12 21.1 11 20.0 1 1.9 1 1.9 21 11.6 13 7.7

Marital Status Married / Defacto 56 98.3 54 98.2 52 98.1 48 92.3 179 98.9 161 95.3 Not Married 1 1.8 - - 1 1.9 2 3.9 2 1.1 8 4.7 Unknown3 - - 1 1.8 - - 2 3.9 - - - -

Living arrangement Own home 39 68.4 41 74.6 43 81.1 40 77.0 160 88.4 152 89.9 Family members home 8 14.0 7 12.7 8 15.1 4 7.7 14 7.7 10 5.9 Retirement unit/village 5 8.77 3 5.5 2 3.8 4 7.7 7 3.8 6 3.6 Community residential home 5 8.77 3 5.5 - - 2 3.9 - - 1 0.6 Rest home - - - - - - - - - - - - Private hospital - - - - - - - - - - - - Unknown3 - - 1 1.8 - - 2 3.9 - - - -

1 Percentages are calculated based on total number of Registered in the study 2 Percentages are calculated based on total number of Registered in the study in each centre 3 Unknown data is data that is missing as the older person refused or were unable to answer * Summary statistics will be presented instead of N.

SAMPLE CHARACTERISTICS FINDINGS

PAGE 55

Table 3-11: Baseline Older person demographics for the three sites in total All

Usual Care New Service

N* %2 N* % 2 Age 55-64 65-69 15 5.15 14 5.0770-74 30 10.31 39 14.1375-79 61 20.96 57 20.6580-84 72 24.74 67 24.2885+ 113 38.83 99 35.87N* 291 276 Mean* 82.01 81.39 SD* 7.00 7.11 Median* 82 82

Education Primary school 102 35.05 77 27.90High school 141 48.45 149 53.99Tertiary 48 16.49 47 17.03Unknown - - 3 1.09

Gender Female 188 64.60 180 65.22

Ethnicity NZ European 251 86.25 244 88.41NZ Maori 5 1.72 5 1.81Pacific 1 0.34 2 0.72Other 34 11.68 25 9.06

Marital Status Married / Defacto 287 98.63 263 95.29Not Married 4 1.37 10 3.62Unknown - - 3 1.09

Living arrangement Own home 242 83.16 233 84.42Family members home 30 10.31 21 7.61Retirement unit/village 14 4.81 13 4.71Community residential home 5 1.72 6 2.17Rest home - - - -Private hospital - - - -Unknown - - 3 1.09

1 Percentages are calculated based on total number of Registered in the study 2 Percentages are calculated based on total number of Registered in the study in each centre * Summary statistics will be presented instead of N.

The following tables provides information around disability and morbidity

SAMPLE CHARACTERISTICS FINDINGS

PAGE 56

Table 3-12: Baseline Older person demographics for each of the three sites Hamilton Lower Hutt Christchurch

Usual Care

New Service

Usual Care

New Service

Usual Care

New Service

N* %2 N* % 2 N* % 2 N* % 2 N* % 2 N* % 2 Health and Disability needs High 51 89.5 49 89.1 44 83.0 45 86.5 144 79.6 134 79.3 Very High 6 10.5 6 10.9 9 17.0 7 13.5 37 20.4 35 20.7

Vision Problems 54 94.7 52 94.5 29 54.7 30 57.7 118 65.2 120 71.0 Unknown - - 1 1.8 - - 2 3.8 - - - -

Hearing Problems 27 47.4 28 50.9 18 34.0 21 40.4 101 55.8 96 56.8 Unknown - - 1 1.8 - - 2 3.8 - - - -

Memory Problems10 45 78.9 40 72.7 21 39.6 27 51.9 122 67.4 102 60.4

Unknown - - 1 1.8 - - 2 3.8 - - - -

Communication Problems 33 57.9 24 43.6 16 30.2 17 32.7 21 11.6 14 8.3 Unknown - - 1 1.8 - - 2 3.8 - - - -

Falls 37 64.9 40 72.7 29 54.7 27 51.9 57 31.5 54 32.0 Unknown - - 1 1.8 - - 2 3.8 - - - - Hospital admission in past 12 months 33 57.9 30 54.5 34 64.2 36 69.2 79 43.6 80 47.3 Unknown - - 1 1.8 - - 2 3.8 - - - -

Home alone 21 36.8 21 38.2 24 45.3 24 46.2 85 47.0 94 55.6 Has a Caregiver 50 87.7 46 83.6 38 71.7 29 55.8 65 35.9 56 33.1

Require help for everyday activities 57 100.0 55 100.0 53 100.0 52 100.0 172 95.0 166 98.2 Using aids/devices for indoor mobility 55 96.5 54 98.2 53 100.0 51 98.1 179 98.9 169 100.0

1 Percentages are calculated based on total number of Registered in the study 2 Percentages are calculated based on total number of Registered in the study in each centre * Summary statistics will be presented instead of N.

Interestingly, the percentage of high and very needs allocation differs across the three sites, with Hamilton having the lowest percentage of very high needs (10.5 and 10.9 for

10 An accurate diagnosis of dementia is difficult to define in this population group and requires corroboration from the subject’s GP and medical specialist. At this present point, a combination of self-reported memory problems as presented here or scoring on the MDS-HC Cognitive Performance Scale is illustrated in the document.

SAMPLE CHARACTERISTICS FINDINGS

PAGE 57

usual care and Community FIRST respectively) and Christchurch having the highest percentage (20.4 and 20.7 for usual care and COSE respectively). The balance between groups (usual care vs. intervention) for this variable should be equal as this formed one of the stratification factors.

There appears to be various differences between the three sites, the large majority of older people in Hamilton tended to have visual problems, tended to have more communication problems, higher levels of cognitive impairment, were falling over more frequently and were more likely to have an informal caregiver. These findings are at baseline and so no treatment effect has occurred and yet it appears clear that the Hamilton group are more disabled despite having the lowest number of ‘very high needs’ older people out of the three sites. This can be explored in more detail with the MDS-HC outcome tools, such as the ADL and IADL scales. The table below illustrates the characteristics for subjects as a total and is useful in demonstrating the prevalence of disability for the meta-analysis

SAMPLE CHARACTERISTICS FINDINGS

PAGE 58

Table 3-13: Baseline Older person demographics across the three sites in total

All Usual Care New Service

N* % 1 N* % 1 Health and Disability needs High 239 82.1 228 82.6 Very High 52 17.9 48 17.4 Vision Problems 201 69.1 202 73.2 Unknown - - 3 1.1 Hearing Problems 146 50.2 145 52.5 Unknown - - 3 1.1 Memory Problems 188 64.6 169 61.2 Unknown - - 3 1.1 Communication Problems 70 24.1 55 19.9 Unknown - - 3 1.1 Falls 123 42.3 121 43.8 Unknown - - 3 1.1 Hospital admission in past 12 months 146 50.2 146 52.9 Unknown - - 3 1.1 Home alone 130 44.7 139 50.4 Has a Caregiver 153 52.6 131 47.5 Require help for everyday activities 282 96.9 273 98.9 Using aids/devices for indoor mobility 287 98.6 274 99.3

1 Percentages are calculated based on total number of Registered in the study 2 Percentages are calculated based on total number of Registered in the study in each centre * Summary statistics will be presented instead of N.

SAMPLE CHARACTERISTICS FINDINGS

PAGE 59

Table 3-14: Baseline older person demographics medical history across the three sites

Hamilton Lower Hutt Christchurch Usual Care

New Service

Usual Care

New Service

Usual Care

New Service

N % 2 N % 2 N % 2 N % 2 N % 2 N % 2 Medical History Arthritis in joint 36 63.2 28 50.9 33 62.3 33 63.5 120 66.3 101 59.8 Hypertension 39 68.4 33 60.0 31 58.5 29 55.8 97 53.6 87 51.5 Cataracts 27 47.4 24 43.6 28 52.8 23 44.2 71 39.2 74 43.8 Angina 21 36.8 26 47.3 15 28.3 23 44.2 71 39.2 62 36.7 Stroke 25 43.9 20 36.4 25 47.2 29 55.8 50 27.6 50 29.6 Wrist/vertebral fracture 24 42.1 21 38.2 18 34.0 24 46.2 55 30.4 49 29.0 Irregular pulse 13 22.8 15 27.3 18 34.0 11 21.2 46 25.4 47 27.8 Asthma 15 26.3 10 18.2 13 24.5 13 25.0 52 28.7 44 26.0 Depression 13 22.8 12 21.8 13 24.5 18 34.6 52 28.7 37 21.9 Osteoporosis 16 28.1 9 16.4 20 37.7 19 36.5 35 19.3 42 24.9 Terminally ill 2 3.5 2 3.6 3 5.7 1 1.9 3 1.7 1 0.6

Fall in the last 3 months 37 64.9 40 72.7 29 54.7 27 51.9 57 31.5 54 32.0 Smoking 5 8.8 7 12.7 5 9.4 - - 13 7.2 17 10.1

Number of medications 0 4 7.0 2 3.6 - - - - 2 1.1 4 2.4 1 2 3.5 2 3.6 2 3.8 - - 4 2.2 1 0.6 2 7 12.3 2 3.6 1 1.9 - - 11 6.1 12 7.1 3 5 8.8 - - 2 3.8 1 1.9 17 9.4 9 5.3 4 4 7.0 7 12.7 5 9.4 2 3.9 12 6.6 21 12.4 5 6 10.5 3 5.5 2 3.8 6 11.5 18 9.9 18 10.7 6 5 8.8 11 20.0 6 11.3 10 19.2 21 11.6 21 12.4 7 9 15.8 11 20.0 7 13.2 7 13.5 27 14.9 20 11.8 8 8 14.0 4 7.3 6 11.3 7 13.5 21 11.6 13 7.7 9 7 12.3 12 21.8 22 41.5 17 32.7 48 26.5 50 29.6 missing - - 1 1.8 - - 2 1.9 - - - - 1 Percentages are calculated based on total number of Registered in the study 2 Percentages are calculated based on total number of Registered in the study in each centre

These findings are interesting in that again regional differences have been demonstrated, particular in light of the reported incidence of stroke, medication usage and depression. It must be remembered that for diseases, these are reported by the older person with confirmation from a caregiver if present and therefore the accuracy of these findings have not been verified from medical records. The medication containers are viewed and recorded by the research associates and may therefore have a greater level of reliability.

SAMPLE CHARACTERISTICS FINDINGS

PAGE 60

Table 3-15: Baseline older person demographics – community services across the three sites

Hamilton Lower Hutt Christchurch Usual Care

New Service

Usual Care

New Service

Usual Care

New Service

N % 2 N % 2 N % 2 N % 2 N % 2 N % 2 Social activities Not having visitors or visiting people 51 89.5 49 89.1 51 96.2 49 94.2 165 91.2 154 91.1 No decline in social activities in the last 3 months 19 33.3 17 30.9 15 28.3 12 23.1 127 70.2 112 66.3 Not seeing/speak to relatives 1 1.8 1 1.8 3 5.7 1 1.9 4 2.2 8 4.7 No interaction with friends 18 31.6 9 16.4 17 32.1 11 21.2 17 9.4 13 7.7 No interaction with neighbours 13 22.8 15 27.3 17 32.1 11 21.2 31 17.1 25 14.8 Not attending religious meetings 15 26.3 9 16.4 12 22.6 15 28.8 45 24.9 35 20.7 Not attend meetings of community / neighbourhood or social group 9 15.8 8 14.5 9 17.0 13 25.0 54 29.8 52 30.8 Feel lonely when alone 29 50.9 16 29.1 8 15.1 12 23.1 42 23.2 39 23.1

As can be seen later in this Chapter, the risk factors for admission to a residential home for an older person are not just related to their functional independence and morbidity, moreover it appears that low mood or social isolation are strong predictors. This table is useful in that it portrays the level of isolation and loneliness amongst the sample. At least 90% of the sample across the three sites report that they do not have visitors and do not visit people and half of the usual care group in Hamilton express that they feel lonely when alone. These findings are highly relevant to planning service intervention for older people. All three of the AIPI purport to address loneliness issues by re-integrating the older person back into their community.

Social connectedness and inclusion within social networks are an important part of everyday life. Networks may involve family/whanau members, friends, neighbours and people within the wider community. Patterns of family composition are changing. People can be socially isolated and lonely whether they are residing in their own home or residential care. Vulnerability factors for loneliness that have been identified in a recent

SAMPLE CHARACTERISTICS FINDINGS

PAGE 61

British study include marital status (whether partnered), a past history of loneliness, deteriorating mental health and physical health, and poorer health than anticipated in old age (Victor et al. 2005). Research shows that social participation mitigates loneliness and has a positive effect on quality of life (Victor et al. 2005; Netten et al. 2002).

Table 3-16 presents self-reported information around service input. These data were collected prior to the older person being randomised to an AIPI and is therefore of some note, this relates to service provision prior to the AIPI. That said, there appears to be a regional imbalance regarding home care provision, therapy input (Occupational Therapy [OT] and Physiotherapy [PT]) and Specialist Medical input with Christchurch participants appearing to receive less input than the other areas. This could be as a consequence of the lower disability observed in this group. Although the Christchurch participants were theoretically assessed at the same level of need (high and very high and complex needs), in reality it appears that they were a less disabled group and therefore would require less service input. The lower level of disability in Christchurch in relationship to the other two sites is very clearly demonstrated in Table 3-17 and in Figure 3-5.

SAMPLE CHARACTERISTICS FINDINGS

PAGE 62

Table 3-16: Baseline older person demographics – community services across the three sites

Hamilton Lower Hutt Christchurch Usual Care

New Service Usual Care

New Service Usual Care New Service

N % 2 N % 2 N % 2 N % 2 N % 2 N % 2 Has Home Care 27 47.4 29 52.7 35 66.0 37 71.2 82 45.3 80 47.3N 57 54 53 50 181 169 Mean 39 8.7 178.7 212.2 5.9 6.4 SD 139 20.5 286.6 304.4 8.7 10.1 Median 0 3.5 12 12 0 0 Has visiting nurses 22 38.6 28 50.9 9 17.0 11 21.2 25 13.8 27 16.0N 57 54 53 50 181 169 Mean 4.3 3.1 13.2 6.5 0.7 1.0 SD 22.4 6.5 92.2 40.7 2.8 3.5 Median 0 0.33 0 0 0 0 Has Home Help 35 61.4 33 60.0 34 64.2 34 65.4 157 86.7 151 89.3N 57 54 53 50 181 169 Mean 3.9 3.3 25.9 53.3 7.5 8.0 SD 7.7 4.3 111.7 171.7 4.4 4.6 Median 2 2 4 4 8 8 Has MOW (meal) 14 24.6 15 27.3 18 34.0 17 32.7 33 18.2 35 20.7Has PT 14 24.6 20 36.4 11 20.8 12 23.1 17 9.4 17 10.1N 57 54 53 50 181 169 Mean 0.8 7.3 0.5 0.4 0.3 0.4 SD 1.8 46.6 1.5 0.9 1.3 1.6 Median 0 0 0 0 0 0 OT 11 19.3 13 23.6 14 26.4 8 15.4 10 5.5 11 6.5N 57 54 53 50 181 169 Mean 0.3 0.5 0.4 0.6 0.1 0.2 SD 0.7 1.1 0.7 2.9 0.8 0.9 Median 0 0 0 0 0 0 Has PT or OT 17 29.8 24 43.6 19 35.8 15 28.8 21 11.6 21 12.4N 57 54 53 50 181 169 Mean 1.1 7.8 0.9 1.1 0.4 0.5 SD 2.2 46.6 1.9 3.0 1.8 2.2 Median 0 0 0 0 0 0 Number of Med specialist visits 19 33.3 26 47.3 22 41.5 24 46.2 40 22.1 27 16.0N 57 54 53 50 181 169 Mean 0.5 0.8 0.8 0.8 0.2 0.2 SD 1.0 1.3 1.1 1.0 0.4 0.5 Median 0 0 0 0 0 0 Number of Com Services 0 4 7.0 2 3.6 7 13.2 3 5.8 4 2.2 4 2.41 8 14.0 9 16.4 4 7.6 6 11.5 63 34.8 57 33.72 22 38.6 9 16.4 14 26.4 11 21.2 64 35.4 58 34.33 13 22.8 18 32.7 11 20.8 14 26.9 31 17.1 32 18.94 and above 10 17.5 16 29.1 17 32.1 16 30.8 19 10.5 18 10.7

SAMPLE CHARACTERISTICS FINDINGS

PAGE 63

Table 3-17: Baseline older person demographics – scale measurements across the three sites

All Hamilton Lower Hutt Christchurch Usual Care

New Service

Usual Care

New Service

Usual Care

New Service

Usual Care

New Service

ADL Long Form N 291 273 57 54 53 50 181 169 Mean 1.9 2.0 4.6 3.9 2.1 2.0 0.9 1.5 SD 4.3 4.3 6.6 5.0 4.0 3.6 2.9 4.1 Median 0.0 0.0 2.0 2.0 0.0 0.5 0.0 0.0IADL Summary N 291 273 57 54 53 50 181 169 Mean 10.0 10.4 14.8 14.7 12.3 12.0 7.8 8.6 SD 5.9 5.8 4.4 3.9 4.4 4.4 5.5 5.8 Median 9.0 10.0 15.0 15.0 12.0 12.0 7.0 8.0Cognitive Performance Scale (CPS) N 291 273 57 54 53 50 181 169 Mean 1.4 1.3 1.9 1.5 1.2 1.2 1.3 1.3 SD 1.2 1.2 1.4 1.1 1.0 1.1 1.2 1.3 Median 1.0 1.0 2.0 2.0 1.0 1.0 1.0 1.0Depression Rating Scale (DRS) N 291 273 57 54 53 50 181 169 Mean 3.1 3.1 6.3 5.1 3.0 2.9 2.1 2.5 SD 2.8 2.7 3.3 3.7 1.8 2.1 2.0 2.0 Median 3.0 2.0 6.0 5.0 3.0 2.0 2.0 2.0Changes in Health, End-stage disease and Signs and Symptoms (CHESS) N 291 273 57 54 53 50 181 169 Mean 2.5 2.4 3.1 3.0 3.0 3.0 2.1 2.0 SD 1.1 1.1 1.0 1.1 0.9 0.8 1.1 1.0 Median 3.0 2.0 3.0 3.0 3.0 3.0 2.0 2.0Pain Scale N 291 273 57 54 53 50 181 169 Mean 1.3 1.2 1.5 1.5 1.1 1.2 1.3 1.0 SD 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 Median 1.0 1.0 2.0 2.0 0.0 2.0 1.0 1.0EuroQoL Thermometer scale N 291 272 57 54 53 50 181 168 Mean 62.1 65.0 60.9 68.8 58.8 62.2 63.4 64.6 SD 21.6 18.9 26.6 22.3 18.3 17.1 20.8 18.2 Median 60.0 65.0 55.0 72.5 50.0 60.0 60.0 65.0

SAMPLE CHARACTERISTICS FINDINGS

PAGE 64

Figure 3-5: Mean Disability across the regions at Baseline11

Contrary to most other outcome scales for older people, MDS-HC scales differ in that complete independence is 0 with higher numbers depicting a higher level of disability. Figure 3-5 represents the mean values for both the intervention and usual care group for each of the three sites. The plot clearly illustrates the much higher levels of disability in Hamilton in relation to the other two sites and of some importance the very low levels of disability in Christchurch. Table 3-18 provides an indication of the meaning behind the relative scores for ADL measures and in general terms one could take from this that on average the older people assessed with high and very high needs in Christchurch have setup help only, where in Hamilton, the enrolled participants require extensive help to perform ADL tasks. This clearly has ramifications for the interpretation of findings presented in the next section.

11 These are mean measures of physical and cognitive disability at entry into the ASPIRE trials

4.3

2.1 1

12.1

8

1.7 1.2 13.1 3.0

2

14.8

0

2

4

6

8

10

12

14

16

Hamilton Wellington Christchurch

Mea

n sc

ores

for u

sual

car

e an

d A

IPI

ADL Long Form IADL Summary Cognitive Performance Scale CHESS

SAMPLE CHARACTERISTICS FINDINGS

PAGE 65

Table 3-18: ADL Long Form Scale scoring responses

Score Description

0. INDEPENDENT – No help, setup, or oversight – OR – Help, setup, oversight provided only 1 or 2 times during last 3 days (with any task or subtask)

1. SETUP HELP ONLY – Article or device provided within reach of client 3 or more times

2. SUPERVISION – Oversight, encouragement or cueing provided 3 or more times during last 3 days – OR – Supervision (1 or more times) plus physical assistance provided only 1 or 2 times during last 3 days (for a total of 3 or more episodes of help or supervision)

3. LIMITED ASSISTANCE – Client highly involved in activity, received physical help in guided manoeuvring of limbs or other non-weight bearing assistance 3 or more times – OR – combination of non-weight bearing help with more help provided only 1 or 2 times during period (for a total of 3 or more episodes of physical help)

4. EXTENSIVE ASSISTANCE – client performed part of activity on own (50% or more of subtasks), period, but help of following type(s) were provided 3 or more times: weight bearing support – OR – Full performance by another during part (but not all) of last 3 days

5. MAXIMAL ASSISTANCE – Client involved and completed less than 50% of subtasks on own (includes 2+ person assist), received weight bearing help or full performance of certain subtasks 3 or more times.

6. TOTAL DEPENDENCE Full performance of activity by another

8. ACTIVITY DID NOT OCCUR (regardless of ability)

3.2.4 Caregiver demographics

As previously discussed, the informal caregiver, where present was asked to complete a questionnaire at each of the time points. The information presented here pertains to the baseline demographics and once again there appears to be regional differences. The caregiver in Christchurch is much more likely to be living with the older person and is more likely to be a spouse, where as in Hamilton, the caregiver tended to be a child or child-in-law who was working in an employed position. In contrast, as seen in Table 3-19, Christchurch has a higher proportion of older people living alone than the other sites. Also of importance is the impact that caring has on the employment opportunities of the caregiver; nearly 1/5 of caregivers in Hamilton reported that they were required to reduce employment as a consequence of their caring role, where as in Christchurch, the impact of caring on employment was less pronounced and was probably linked to most caring being undertaken by a spouse.

SAMPLE CHARACTERISTICS FINDINGS

PAGE 66

Table 3-19: Baseline caregiver demographics across the three sites Hamilton Lower Hutt Christchurch

Usual Care

New Service

Usual Care

New Service

Usual Care

New Service

N* %2 N* % 2 N* % 2 N* % 2 N* % 2 N* % 2 Live with OP Yes 30 61.2 24 52.2 28 75.7 23 79.3 53 80.3 47 83.9 Unknown - - 2 4.3 4 10.8 2 6.9 2 3.0 2 3.6 Relationship with OP Child or child – in – law 23 46.9 26 56.5 17 45.9 8 27.6 25 37.9 19 33.9 Spouse 20 40.8 14 30.4 16 43.2 18 62.1 37 56.1 33 58.9 Other relative 5 10.2 2 4.3 - - 1 3.4 2 3.0 - - Friends / Neighbour 1 2.0 2 4.3 - - - - - - 2 3.6 Unknown - - 2 4.3 4 10.8 2 6.9 2 3.0 2 3.6

Employed 17 34.7 19 41.3 10 27.0 7 24.1 14 21.2 16 28.6 Unknown - - 2 4.3 4 10.8 2 6.9 2 3.0 2 3.6

Reduced employment No 8 16.3 11 23.9 6 16.2 6 20.7 8 12.1 13 23.2 Yes 9 18.4 8 17.4 4 10.8 1 3.5 6 9.1 3 5.4 Unknown 32 65.3 27 58.7 27 73.0 22 75.9 52 78.8 40 71.4

1 Percentages are calculated based on total number of Registered in the study 2 Percentages are calculated based on total number of Registered in the study in each centre * Summary statistics will be presented instead of N. Informal care from family and friends in the form of emotional and practical support facilitates the ability of people to remain living in the community, even with disabilities (Dwyer, Gray & Renwick 2000). Social support is not necessarily caregiving. Support may take many forms, such as simply ‘being there’ when needed or ‘in kind’ assistance, such as home and car maintenance, funding holidays, or providing meals. Those living alone are recorded as receiving considerably more ‘in kind’ support from outside the home than those in two-person households (Statistics New Zealand 2004). However, despite the major role that informal caregivers have in supporting an older person to remain living in their own home, it appears very evident from table 3-20 that the caregivers of those older

SAMPLE CHARACTERISTICS FINDINGS

PAGE 67

people enrolled in the study perceived a high level of dissatisfaction in their role as caregiver as recorded on the Caregiver Reaction Assessment12.

Table 3-20: Baseline caregiver demographics scale measurements across all three sites

All Hamilton Lower Hutt Christchurch Usual

Care New Service

Usual Care

New Service

Usual Care

New Service

Usual Care

New Service

SF-36 Physical Component Scale (PCS)

N 145 125 49 44 33 27 63 54 Mean 41.9 43.5 42.9 43.2 40.0 44.3 42.2 43.3 SD 12.3 11.3 11.9 10.8 10.5 10.2 13.4 12.3 Median 43.6 45.8 41.9 43.3 39.7 46.9 44.8 46.4

SF-36 Mental Component Scale (MCS)

N 145 125 49 44 33 27 63 54 Mean 45.8 44.6 44.0 41.0 43.8 44.9 48.1 47.3 SD 11.2 11.8 11.6 13.7 11.2 9.1 10.7 10.8 Median 48.2 47.5 44.2 40.4 42.1 44.7 50.5 48.9

Caregiver Reaction Assessment (CRA)

N 146 125 49 44 33 27 64 54 Mean 74.0 75.3 76.3 77.7 75.5 73.8 71.4 74.1 SD 9.0 9.2 7.4 9.9 9.1 7.5 9.4 9.2 Median 74.0 76.0 77.0 78.0 75.0 74.0 71.0 74.0

3.3 Summary

This section has highlighted the characteristics of both the enrolled older people and their caregivers and has thus identified an anomaly around the differing disability levels across the three regions. Given that all older people entering ASPIRE were assessed with the same NASC processes, one would assume broadly similar disability levels. However, it is apparent from this section that those older people living in Christchurch are far less disabled than those living in Hamilton or Hutt. These findings may have implications for the primary results presented next and the power of the study to identify any true treatment

12 CRA has 24 item questions, scored from 0 to 5, yielding scores from a minimum of 24 to a maximum of 120, the higher the score the greater the dissatisfaction with the caregiving role

SAMPLE CHARACTERISTICS FINDINGS

PAGE 68

effect as well as implications for new more rigorous assessment tools such as the MDS-HC currently being trialled across New Zealand.

PRIMARY RESULTS FINDINGS

PAGE 69

Section 2: Primary results

3.4 Introduction

Ageing-in-place is a complex notion to describe, as it encompasses multiple domains, including housing, community integration, family / whanau support as well as a host of other aspects. However, it is often simpler to explore and define what is not represented by ageing-in-place. As already discussed, ageing-in-place invariably refers to the ability of older people to remain dwelling in the community, including within retirement villages (MSD, 2006); and residential care in the form of either rest homes or hospitals is specifically excluded. However, it is overly simplistic to isolate ageing in place from residential care as for an older person to live at home with a high quality of life; they may receive considerable support from an informal caregiver such as a child (as seen in

The primary analysis revealed a statistically significant difference in

the results for COSE compared to usual care in Christchurch for both

the combined outcome (mortality and residential care admission) as

well as for residential admissions (demonstrated by a 43% reduction

in the risk of residential home placement) alone, though not mortality

alone.

Further, there was a reduction in residential home admission in

Hamilton (33% risk reduction) and Hutt (16% risk reduction) and

mortality in Hamilton (28% risk reduction) and Hutt (16% risk

reduction) in the AIPI, however possibly due to the small sample size,

the results were not statistically significant. The high impact of

Community FIRST on residential home placement in Hamilton

appears positive given the high baseline disability of older people in

that region.

The pooled analyses across centres (i.e. the Meta analysis) also

showed a statistically significant treatment effect in delaying

permanent residential home admission and combined primary

outcome for the AIPI, which is around 31% lower than usual care with

95% confidence interval around 9% to 47%.

PRIMARY RESULTS FINDINGS

PAGE 70

Hamilton) or a spouse (as seen more in Christchurch and Hutt). Therefore, a balance is required between the quality of life of the older person and that of the caregiver as a situation may develop where caring for an older person may detract from the quality of life of the caregiver and place an untenable burden on that individual. Similarly, situations frequently arise whereby the disability of the older person is so pronounced, that a placement in a private hospital13 becomes the only viable option and it may be possible, using the ASPIRE dataset to explore this ‘threshold’ in more depth. The AIPI under investigation in this study aimed not so much to fully exclude access to residential care but moreover to delay entry into residential care and further, if residential care was required, it would be at the level of private hospital. Therefore, the results of the analysis that have been presented in this section reflect this important distinction and do so using a combination of techniques such as Survival plots (Appendix 3) and Multivariate Cox Proportional Hazards Model approach (Section 2.12). The primary findings are structured through graphs, tables and accompanying narrative.

3.5 Design effect

As a consequence of the required cluster design in Christchurch, (as GP practices were the units of randomisation rather than individuals as in the other sites), further calculations were required to establish the extent to which the clusters were different from another. The original premise was to anticipate some level of heterogeneity and therefore a DEFF of 2 was considered. However, following analysis upon completion of the data collection, intra-cluster correlation coefficient (ICC) for the combined primary outcome was found to be at 0.0061 in Christchurch. ICC for the death and residential care endpoints separately were estimated to be zero. This very small clustering effect was taken into account when conducting statistical analysis for the combined endpoint.

13 As discussed, there are broadly three different forms of residential care: Rest homes are typically for older people assessed with Support Needs Level 4, Private Hospitals are for older people with Support Needs Level 5 and dementia units are for older people with significant cognitive impairment which is often accompanied with physical disability and behavioural changes that make it difficult for the older person to remain living in their own home

PRIMARY RESULTS FINDINGS

PAGE 71

3.6 Overall statistical findings – permanent residential home placement and mortality

Log-Rank Test was performed initially to access the treatment effect across groups in each centre. No treatment effects of statistical significance were found for mortality and permanent residential admissions, as well as for the combined primary outcome in Hamilton. Similar results were found for the Lower Hutt region. No treatment effect of statistical significance was found for death in Christchurch. However, statistically significant treatment effects were found in Christchurch for residential admissions (p value = 0.0295) and for the combined primary outcome (p value = 0.0311). Kaplan-Meier curves indicated similar results. There was a statistically significant treatment effect between groups in Christchurch (Appendix 3). This means that people receiving the AIPI had a delay in residential admissions or death, compared to the people who receiving Usual Service. Even though, the Log-Rank tests for time-to-death data indicated no statistically significant treatment effect, there was a better survival rate and delay in residential care placement for the AIPI from months 3 to 18 for all sites, as indicated in Survival plots (Appendix 3).

Cox-PH models were then conducted to estimate Hazard Ratios together with 95% confidence limits. Both unadjusted and adjusted analyses were performed, where the adjusted analyses included all stratification factors. Treatment effect was statistically significant for residential entry as well as the combined primary outcome for Christchurch participants.

Hazard Ratio for residential admissions in Christchurch for unadjusted analysis for people who were receiving COSE was 40% lower, compared to people who were receiving Usual Care with a 95% confidence interval of (5%, 63%). The adjusted analysis showed very similar results 43% lower compared to people who were receiving Usual Care with 95% confidence interval (8%, 65%).

Meta-analysis was performed to pool treatment effects from the three study centres to provide an overall treatment effect estimate. Homogeneity test was conducted while performing Meta-analysis. No heterogeneity was found between centres, therefore fixed effect Meta-analysis was performed for the overall treatment effect.

PRIMARY RESULTS FINDINGS

PAGE 72

The overall treatment effect for combined primary outcome for all three sites combined was 32% lower with a 95% confidence interval of (11%, 49%) for AIPI compared to the Usual Care for the unadjusted analysis. The adjusted analysis showed that the combined primary outcome is 31% lower for the AIPI compared to the Usual Care with a 95% confidence of 9% to 47%.

3.7 Permanent residential home admission

Older people access residential care either permanently or through a temporary arrangement for respite or sub-acute transitional care. For the purposes of ASPIRE, it was important to make this distinction and given that permanent residential home placement has funding implications, it was relatively simple to ascertain the date at which the older person permanently entered residential care. This data were gathered initially directly from NASC and later confirmed using the national datasets. There are three main forms of residential care and these are: Rest Home, Private Hospital and Dementia Unit and although it is possible to analyse the data by placement option, for the purposes of this report, all types of residential homes are classified as simply residential. As AIPI programmes develop, it is likely that older people with levels of disability that have traditionally been seen in rest homes would be supported in their own home. Therefore, there is a possibility that further analysis of the ASPIRE dataset may identify lower residential home usage overall and a higher proportion of private hospital facility usage.

Both Hamilton and Lower Hutt had 39 participants admitted to residential care, 22 from the Usual Care and 17 from the New Service (Community FIRST and PIP). Christchurch had 72 permanent residential admissions, 45 from the Usual Care and 27 from COSE. Table 3-21 highlights the impact of the AIPI on permanent residential care, both for individual sites as well as through the meta-analysis. Although all three AIPI demonstrated a reduction in permanent residential home usage, the decrease was only statistically significant in Christchurch. However, this is not surprising as the study was not powered for the primary endpoints in the individual sites, though despite this, Christchurch independently has demonstrated a statistically significant 43% decline (HR: 0.57). Further, the observed 33% decline in residential home admission in the Community FIRST group in relation to usual care (HR: 0.67) appears positive given the relatively high baseline

PRIMARY RESULTS FINDINGS

PAGE 73

disability of the participants (as determined from baseline data instead of assessed needs level).

Table 3-21: Hazard ratio on primary outcome, residential care (adjusted, design effect estimated inflating variance of 4%)

Usual Care New

Service

N* % N* % HR 95%CI p-

value Permanent Residential Care

Hamilton 22 56.41 17 43.59 0.67 (0.35,1.28) 0.2261Lower Hutt 22 56.41 17 43.59 0.84 (0.44,1.60) 0.6003Christchurch 45 62.50 27 37.50 0.57 (0.35,0.92) 0.0206

ICC in Christchurch* 0

Overall 89 59.33 61 40.67 0.66 (0.47,0.92) * ICC calculated for Cluster Randomised Trial as: ICC= for residential care

Figure 3-6 presents these findings as a Forest plot. The size of the block for each of the three sites is proportional to the percentage weighted contribution in terms of sample size. The horizontal line represents the confidence interval. The horizontal line at the bottom is the scale measuring the treatment effect. Here, towards the left, the scale is less than one, meaning all three AIPI have made people less likely to enter permanent residential care. The vertical line in the middle is where AIPI and usual care have the same effect, or in other words, there is no difference between the two. If the confidence interval crosses this line it indicates that the result is not statistically significant. The pooled analysis is given a diamond shape; the widest aspect is located on the point estimate and the horizontal width in the confidence interval, which illustrates that the meta-analysis for residential care was statistically significant. Simply put, the further the box is to the left, the greater the reduction in permanent residential home admission in the AIPI.

PRIMARY RESULTS FINDINGS

PAGE 74

Figure 3-6: Adjusted overall treatment effect estimate (using residential home placement as primary endpoint)

3.8 Mortality

Mortality data were gathered from the research associates in each site with later confirmation being ascertained from national datasets. Survival is an unambiguous measure of the impact of a health services on an individuals health status. A total of 113 older people died during the course of the study, of which, 29 participants were from Hamilton, 24 from Lower Hutt and 60 from the Christchurch region. In Hamilton, 16 older people were from the usual care group and 13 were from Community FIRST. In Lower Hutt, there were 14 deaths from the usual care group and 10 from the PIP service. Finally, in Christchurch, 33 deaths occurred in the usual care group compared to 27 from COSE.

PRIMARY RESULTS FINDINGS

PAGE 75

The table below highlights the total deaths for all three sites as well as the meta-analysis. Although, all three services reduced the risk of mortality, none were statistically significant. However, the results are all positive with the highest reduction in mortality occurring in Hamilton within the Community FIRST initiative in comparison to usual care. A Hazard Ratio (HR) of 0.72 means a reduction in mortality of almost 30%. The study was not powered for the primary endpoints in each of the individual sites and therefore it is not surprising that these results are not statistically significant. The HR of 0.90 or a 10% reduction for Christchurch is not surprising given the relatively low disability level of older people enrolled into ASPIRE within that region.

Table 3-22: Hazard ratio on Primary outcome, death (adjusted, design effect estimated inflating variance of 4%) Usual Care New Service N* % N* % HR 95%CI p-valueDeath

Hamilton 16 55.17 13 44.83 0.72 (0.34,1.51) 0.3789 Lower Hutt 14 58.33 10 41.67 0.84 (0.37,1.91) 0.6700 Christchurch 33 55.00 27 45.00 0.90 (0.54,1.50) 0.6754

ICC in Christchurch* 0

Overall 63 55.75 50 44.25 0.83 (0.57,1.22) * ICC calculated for Cluster Randomised Trial as: ICC= for death

The Forest plot on the following page illustrates the Hazard Ratios. As with the residential home risk, the further the box is to the left, the higher the reduction in mortality in the AIPI in relation to the usual care groups in each of the sites (rectangular boxes) and as a whole (diamond shape). The figure illustrates that none of the reductions were statistically significant and none of the individual sites were expected to as the study was powered for a difference across all sites in total (meta analysis).

PRIMARY RESULTS FINDINGS

PAGE 76

Death Adjusted

Hazard Ratio

Stud

y C

entre

0.40 0.63 1.00 1.58

Hamilton

Wellington

Christchurch

Overall

Figure 3-7: Adjusted overall treatment effect estimate (using death as primary endpoint)

3.9 Combined primary endpoint (permanent residential care placement and mortality)

An important aspect of analysis was to explore the impact that AIPI has on both permanent residential home placement and mortality and in particular to examine the number of deaths occurring in the community within both the AIPI and usual care groups. The study was powered to detect an effect of at least a 30% reduction in these two combined endpoints across all three sites, assuming an event rate of 35% in the control group (see section 2.12).

PRIMARY RESULTS FINDINGS

PAGE 77

In total, 220 participants had a combined primary outcome of death or residential admission: 58 were from the Hamilton region; 52 from Hutt and 110 from Christchurch. A total of 32 incidents of combined primary outcome were reported from Usual Care and 26 from Community FIRST in Hamilton. For Lower Hutt, 30 incidents came from the Usual Care and 22 from PIP. In Christchurch, 66 of the incidents were reported from Usual Care and 44 from COSE.

Table 3-23: Hazard ratio on Combined primary outcome – residential home placement and mortality (adjusted, design effect estimated inflating variance of 4%) Usual Care New Service N* % N* % HR 95%CI p-value Death or Residential Care

Hamilton 32 55.17 26 44.83 0.69 (0.40,1.16) 0.1630 Lower Hutt 30 57.69 22 42.31 0.76 (0.43,1.33) 0.3356 Christchurch 66 60.00 44 40.00 0.67 (0.45,0.99) 0.0394

ICC in Christchurch* 0

Overall 128 58.18 92 41.82 0.69 (0.53,0.91) ICC = for combined outcome (death + residential care placement)

The overall treatment effect for combined primary outcome was 32% lower with a 95% confidence interval of (11%, 49%) for AIPI compared to the Usual Care for the unadjusted analysis. The adjusted analysis showed that the combined primary outcome is 31% lower for the AIPI, compared to the Usual Care with a 95% confidence of 9% to 47%. These results are expressed in Figure 3-8 and demonstrate the large positive impact all AIPI are having with residential home placement and mortality.

PRIMARY RESULTS FINDINGS

PAGE 78

Figure 3-8: Adjusted overall treatment effect estimate (using combined primary outcome as primary endpoint)

Using the above values, it is possible to identify whether the older person was permanently residing in a residential care facility at the time of death. The data at this point do not reveal whether the death occurred in a public hospital, merely whether their ‘home’ was or was not a residential facility. The figure on the proceeding page highlights the results. Figure 3-9 describes where the older person was residing at the time of their death, either in permanent residential care or elsewhere.

PRIMARY RESULTS FINDINGS

PAGE 79

3.10 Summary

The original sample size calculations were based on the reasonable assumption of a 35% event rate (residential home placement and mortality) within the usual care group and a 30% difference between AIPI and usual care. Even with a low design effect, the ability to detect a statistical difference between AIPI and usual care would have been difficult with the 569 participants enrolled in the study. However, the magnitude of reduction in residential home placement and mortality across the three sites was such that even with the much lower sample size, statistically significant differences were detected.

Clearly, findings such as these need to be placed within a context and observed differences explained to allow for successful aspects of the service to be identified and if necessary replicated in other DHBs.

0

5

10

15

20

25

Hamiltonusual care

CommunityFIRST

Hutt usualcare

PIP Christchurchusual care

COSE

Region (AIPI and usual care)

Num

ber o

f dea

ths

0

10

20

30

40

50

60

70

80

90

100

Perc

enta

ge o

f dea

ths

ousi

de R

C

Deaths outside RC Deaths inside RC % of deaths outside RC

Figure 3-9: Place of permanent residence at time of death (Residential care is represented by RC)

SECONDARY RESULTS FINDINGS

PAGE 80

Section 3: Secondary results

3.11 Introduction

The residential home admission and mortality findings presented in the previous section can not be viewed in isolation and need to be understood in a broader context. The secondary findings presented here for the older person are: instrumental and activities of daily living, quality of life, pain, cognition, frailty and depression; as well as the perceived burden of care and quality of life for the caregiver. A presentation and discussion of such allows the primary results to be considered and to a greater or lesser extent explained.

As previously highlighted, unlike the primary analysis, there are several ways to interpret these results. One method is to exclude older person data if the individual reaches the primary endpoint of permanent residential home placement. The rationale for such is that

A statistically significant treatment effect for improvement in activities

of daily living was observed in older people in the Community FIRST

service compared to usual care. No trends were observed in either

the Hutt PIP service or COSE. A significant treatment effect was

detected on instrument activities of daily living in Christchurch, when

data from people in residential care was included in the dataset in

that although both COSE and usual care groups tended to deteriorate

over time, the decline was less apparent amongst older people in the

COSE group. There were no differences in quality of life of the older

person between AIPI and usual care, though when the data from

older people who entered a residential facility was removed, there

was a trend for lower rates of depression in the Community FIRST

participants.

Of particular importance was the finding that the AIPI initiatives did

not appear to increase caregiver stress.

SECONDARY RESULTS FINDINGS

PAGE 81

it allows a comparison of the new AIPI community based services14 with usual home care. However, to gain a reflection of changes that occur once the individual has entered residential care, this data needs to be included. For the purposes of completeness, both methods of analysis have been included, though there are few changes noted between the two approaches.

3.12 Older person

Data are presented as both tables and plots (Appendices 4 and 5) across all three sites, for both included residential home data as well as excluded. Changes from baseline to 3 months, 6 months, 12 months and 18 months are illustrated for usual and AIPI. The treatment effect for secondary outcomes at each time point is reported with its standard error (SE) and 95% confidence interval (95% CI). Where the P values are less than 0.05, this indicates if the treatment effect is consistent across time points. The plots in Appendix 4 provide clarity around the trends and are recommended viewing.

3.12.1 Presentation of Instrumental and Activities of Daily Living findings

The following tables present the secondary results for the MDS-HC scales concerning activities of daily living and instrumental activities of daily living. All scales have been discussed in the Chapter II. It is noted that a decrease in ADL and IADL scales represents an improvement (a score of 0 represents independence; therefore a minus value in the change from baseline indicates an improvement).

It is important to consider all Activity of Daily Living (ADL) scales and Instrumental Activities of Daily Living (IADL) scales collectively, and when doing so a clear trend for improvement is observed amongst participants of the Community FIRST trial in Hamilton in

14 This is particularly pertinent for the evaluations of Community FIRST and Masonic PIP, as Community FIRST is a restorative home support programme for older people with high and complex needs and Masonic PIP provides a combination of residential based slow stream rehabilitation with MDT oversight of standard home care once the older person has been discharged from the facility

SECONDARY RESULTS FINDINGS

PAGE 82

ADL, with a statistically significant result for ADL self-performance (P=0.049). The trend for improvement remains when the data includes those older people who have entered residential care,

No trends are apparent in participants of the Hutt PIP programme in terms of ADL or IADL and there appears to be little differences between the PIP service and usual care, indicating that the service had no effect on improving functional status of the older person in comparison to usual care.

In Christchurch, a statistically significant treatment effect is detected in the IADL involvement scale when residential care data is included, and a trend is also observed indicating that the older people receiving COSE are declining in IADL more slowly that those in usual care. The IADL summary scale indicate that the decline in the COSE group is statistically significantly less than the usual care group (P=0.047). This change is most likely due to an increased functional deterioration following admission to a residential facility and is observed in Christchurch as there are a statistically significantly larger group of older people who have entered a residential facility in the usual care group vis a vis

COSE.

SECONDARY RESULTS FINDINGS

PAGE 83

Table 3-24: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for ADL and IADL (Excluding residential home data) in Hamilton Hamilton

Usual Care Community

FIRST Treatment effect

(difference of differences)

Treatment * Visit

Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Older person ADL Short Form 0.12627

3 months -0.46 0.36 -0.40 0.32 -0.06 (-1.02,0.90) 6 months -0.03 0.41 -0.35 0.36 0.32 (-0.76,1.41) 12 months 0.73 0.52 -0.68 0.46 1.40 (0.03,2.78) 18 months 0.10 0.72 -1.60 0.61 1.70 (-0.19,3.59)

ADL Self-performance 0.04930

3 months -0.31 0.19 -0.47 0.16 0.16 (-0.33,0.66) 6 months -0.28 0.26 -0.44 0.23 0.16 (-0.53,0.85) 12 months 0.00 0.24 -0.79 0.21 0.80 (0.15,1.44) 18 months 0.27 0.45 -1.56 0.38 1.83 (0.54,3.12)

ADL Long Form 0.12722

3 months -0.37 0.55 -0.40 0.48 0.03 (-1.42,1.49) 6 months -0.12 0.70 -0.34 0.62 0.23 (-1.65,2.10) 12 months 2.31 1.19 -1.07 1.06 3.38 (0.06,6.69) 18 months 0.20 1.14 -2.15 1.00 2.35 (-0.76,5.46)

IADL Difficulty 0.70106

3 months -0.24 0.22 -0.30 0.19 0.06 (-0.51,0.64) 12 months -0.72 0.36 -0.56 0.33 -0.16 (-1.14,0.81) 18 months -1.11 0.57 -0.55 0.48 -0.56 (-2.02,0.90)

IADL Involvement 0.62842

3 months 0.31 0.28 0.58 0.24 -0.27 (-0.99,0.45) 6 months -0.15 0.31 0.01 0.28 -0.16 (-0.99,0.66) 12 months 0.21 0.42 -0.16 0.38 0.37 (-0.76,1.49) 18 months -0.52 0.62 -1.03 0.53 0.51 (-1.09,2.12)

IADL Summary 0.64385

3 months 1.22 0.59 0.98 0.52 0.23 (-1.31,1.78) 6 months 0.35 0.66 0.44 0.59 -0.08 (-1.84,1.67) 12 months 1.20 0.93 -0.03 0.83 1.22 (-1.23,3.68) 18 months 1.35 1.42 -0.78 1.21 2.12 (-1.55,5.80)

SECONDARY RESULTS FINDINGS

PAGE 84

Table 3-25: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for ADL and IADL (Including residential home data) in Hamilton Hamilton

Usual Care Community

FIRST Treatment effect

(difference of differences)

Treatment * Visit

Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Older person ADL Short Form 0.58485

3 months -0.30 0.38 0.11 0.36 -0.41 (-1.44,0.62) 6 months 0.35 0.42 0.09 0.39 0.26 (-0.88,1.39) 12 months 0.26 0.55 0.08 0.52 0.18 (-1.31,1.66) 18 months -0.27 0.80 -1.08 0.72 0.81 (-1.30,2.93)

ADL Self-performance 0.37687

3 months -0.28 0.19 -0.20 0.18 -0.08 (-0.60,0.44) 6 months -0.11 0.21 -0.25 0.20 0.14 (-0.43,0.71) 12 months -0.17 0.26 -0.46 0.25 0.29 (-0.42,1.00) 18 months -0.21 0.38 -1.04 0.35 0.83 (-0.19,1.85)

ADL Long Form 0.32491

3 months 0.03 0.65 0.44 0.61 -0.41 (-2.17,1.35) 6 months 0.86 0.72 0.23 0.67 0.63 (-1.31,2.57) 12 months 1.90 0.95 0.34 0.90 1.56 (-1.02,4.15) 18 months 1.14 1.41 -1.46 1.26 2.60 (-1.13,6.34)

IADL Difficulty 0.89420

3 months -0.08 0.19 -0.17 0.17 0.09 (-0.42,0.60) 12 months -0.32 0.28 -0.23 0.27 -0.09 (-0.87,0.68) 18 months -0.21 0.43 -0.12 0.38 -0.09 (-1.23,1.05)

IADL Involvement 0.43242

3 months 0.69 0.24 0.72 0.22 -0.02 (-0.66,0.62) 6 months 0.61 0.30 0.49 0.28 0.12 (-0.69,0.93) 12 months 1.11 0.40 0.30 0.38 0.81 (-0.30,1.91) 18 months -0.19 0.61 -0.32 0.57 0.13 (-1.56,1.82)

IADL Summary 0.30175

3 months 1.80 0.53 1.17 0.49 0.63 (-0.80,2.05) 6 months 1.50 0.57 1.12 0.54 0.38 (-1.17,1.92) 12 months 2.46 0.72 0.69 0.68 1.77 (-0.19,3.73) 18 months 3.03 1.02 0.51 0.92 2.52 (-0.19,5.23)

SECONDARY RESULTS FINDINGS

PAGE 85

Table 3-26: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for ADL and IADL (Excluding residential home data) in Hutt Hutt

Usual Care PIP

Treatment effect (difference of differences)

Treatment * Visit

Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Older person ADL Short Form 0.84030

3 months 0.15 0.28 0.03 0.27 0.12 (-0.64,0.89) 6 months 0.42 0.30 -0.02 0.29 0.43 (-0.39,1.25) 12 months 0.27 0.37 0.21 0.32 0.06 (-0.90,1.01) 18 months -0.14 0.63 0.09 0.58 -0.23 (-1.93,1.47)

ADL Self-performance 0.57200

3 months 0.12 0.17 -0.03 0.17 0.15 (-0.33,0.63) 6 months 0.32 0.18 -0.06 0.17 0.38 (-0.12,0.88) 12 months 0.11 0.21 0.08 0.19 0.03 (-0.53,0.59) 18 months 0.02 0.33 0.04 0.31 -0.01 (-0.91,0.88)

ADL Long Form 0.42791

3 months 0.02 0.55 0.16 0.53 -0.14 (-1.66,1.38) 6 months 0.93 0.51 0.01 0.50 0.92 (-0.51,2.35) 12 months -0.30 0.85 0.28 0.76 -0.58 (-2.90,1.75) 18 months -0.48 0.78 0.55 0.70 -1.04 (-3.82,1.75)

IADL Difficulty 0.45142

3 months 0.25 0.23 -0.05 0.23 0.31 (-0.34,0.95) 12 months -0.08 0.31 -0.59 0.27 0.51 (-0.30,1.32) 18 months -0.04 0.53 0.42 0.49 -0.46 (-1.88,0.97)

IADL Involvement 0.21092

3 months 0.12 0.33 -0.46 0.32 0.58 (-0.33,1.48) 6 months 0.60 0.35 -0.01 0.34 0.61 (-0.36,1.57) 12 months -0.43 0.43 -0.17 0.37 -0.27 (-1.38,0.85) 18 months -0.39 0.72 0.71 0.66 -1.10 (-3.03,0.83)

IADL Summary 0.46297

3 months 0.34 0.68 -0.71 0.67 1.05 (-0.84,2.94) 6 months 1.08 0.72 0.58 0.70 0.49 (-1.51,2.50) 12 months -1.01 0.86 -0.56 0.76 -0.45 (-2.72,1.82) 18 months -1.27 1.40 0.09 1.30 -1.37 (-5.14,2.41)

SECONDARY RESULTS FINDINGS

PAGE 86

Table 3-27: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for ADL and IADL (Including residential home data) in Hutt Hutt

Usual Care PIP

Treatment effect (difference of differences)

Treatment * Visit

Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Older person ADL Short Form 0.70926

3 months 0.15 0.28 0.03 0.27 0.12 (-0.64,0.89) 6 months 0.42 0.30 -0.02 0.29 0.43 (-0.39,1.25) 12 months 0.27 0.37 0.21 0.32 0.06 (-0.90,1.01) 18 months -0.14 0.63 0.09 0.58 -0.23 (-1.93,1.47)

ADL Self-performance 0.51423

3 months 0.12 0.17 -0.03 0.17 0.15 (-0.33,0.63) 6 months 0.32 0.18 -0.06 0.17 0.38 (-0.12,0.88) 12 months 0.11 0.21 0.08 0.19 0.03 (-0.53,0.59) 18 months 0.02 0.33 0.04 0.31 -0.01 (-0.91,0.88)

ADL Long Form 0.57408

3 months 0.02 0.55 0.16 0.53 -0.14 (-1.66,1.38) 6 months 0.93 0.51 0.01 0.50 0.92 (-0.51,2.35) 12 months -0.30 0.85 0.28 0.76 -0.58 (-2.90,1.75) 18 months -0.48 0.78 0.55 0.70 -1.04 (-3.82,1.75)

IADL Difficulty 0.18072

3 months 0.25 0.23 -0.05 0.23 0.31 (-0.34,0.95) 12 months -0.08 0.31 -0.59 0.27 0.51 (-0.30,1.32) 18 months -0.04 0.53 0.42 0.49 -0.46 (-1.88,0.97)

IADL Involvement 0.13879

3 months 0.12 0.33 -0.46 0.32 0.58 (-0.33,1.48) 6 months 0.60 0.35 -0.01 0.34 0.61 (-0.36,1.57) 12 months -0.43 0.43 -0.17 0.37 -0.27 (-1.38,0.85) 18 months -0.39 0.72 0.71 0.66 -1.10 (-3.03,0.83)

IADL Summary 0.42738

3 months 0.34 0.68 -0.71 0.67 1.05 (-0.84,2.94) 6 months 1.08 0.72 0.58 0.70 0.49 (-1.51,2.50) 12 months -1.01 0.86 -0.56 0.76 -0.45 (-2.72,1.82) 18 months -1.27 1.40 0.09 1.30 -1.37 (-5.14,2.41)

SECONDARY RESULTS FINDINGS

PAGE 87

Table 3-28: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for ADL and IADL (Excluding residential home data) in Christchurch Christchurch

Usual Care COSE

Treatment effect (difference of differences)

Treatment * Visit

Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Older person ADL Short Form 0.25236

3 months -0.02 0.09 0.01 0.09 -0.03 (-0.29,0.22) 6 months 0.00 0.10 0.24 0.10 -0.24 (-0.51,0.03) 12 months 0.37 0.17 0.31 0.17 0.06 (-0.40,0.52) 18 months 0.42 0.23 0.53 0.22 -0.11 (-0.73,0.50)

ADL Self-performance 0.62822

3 months -0.03 0.05 0.01 0.05 -0.04 (-0.17,0.10) 6 months 0.03 0.05 0.12 0.06 -0.08 (-0.24,0.07) 12 months 0.12 0.07 0.12 0.07 0.00 (-0.18,0.18) 18 months 0.18 0.10 0.20 0.09 -0.02 (-0.28,0.24)

ADL Long Form 0.18976

3 months 0.03 0.18 0.15 0.18 -0.12 (-0.62,0.38) 6 months 0.12 0.20 0.50 0.20 -0.38 (-0.93,0.17) 12 months 0.68 0.29 0.60 0.29 0.08 (-0.73,0.88) 18 months 0.68 0.38 1.15 0.37 -0.47 (-1.51,0.57)

IADL Difficulty 0.82688

3 months -0.05 0.09 -0.01 0.10 -0.05 (-0.31,0.22) 12 months 0.16 0.11 0.21 0.11 -0.05 (-0.35,0.24) 18 months 0.44 0.15 0.37 0.14 0.07 (-0.33,0.47)

IADL Involvement 0.06842

3 months 0.08 0.12 0.02 0.12 0.06 (-0.28,0.40) 6 months 0.36 0.14 0.23 0.14 0.12 (-0.27,0.51) 12 months 0.34 0.13 0.22 0.13 0.12 (-0.23,0.47) 18 months 0.92 0.22 0.18 0.20 0.74 (0.15,1.32)

IADL Summary 0.34774

3 months -0.14 0.18 -0.06 0.19 -0.08 (-0.60,0.44) 6 months 0.23 0.22 0.14 0.23 0.09 (-0.53,0.72) 12 months 0.49 0.23 0.38 0.23 0.11 (-0.53,0.75) 18 months 1.24 0.42 0.34 0.37 0.90 (-0.20,1.99)

SECONDARY RESULTS FINDINGS

PAGE 88

Table 3-29: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for ADL and IADL (Including residential home data) in Christchurch Christchurch

Usual Care COSE

Treatment effect (difference of differences)

Treatment * Visit

Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Older person ADL Short Form 0.36504

3 months 0.10 0.10 0.03 0.10 0.07 (-0.21,0.35) 6 months 0.14 0.11 0.27 0.11 -0.13 (-0.43,0.18) 12 months 0.54 0.17 0.43 0.18 0.11 (-0.38,0.60) 18 months 0.66 0.25 0.81 0.24 -0.14 (-0.83,0.54)

ADL Self-performance 0.59104

3 months 0.04 0.05 0.03 0.05 0.01 (-0.14,0.16) 6 months 0.12 0.06 0.13 0.06 -0.01 (-0.17,0.16) 12 months 0.25 0.07 0.15 0.08 0.11 (-0.10,0.32) 18 months 0.30 0.11 0.30 0.11 0.01 (-0.30,0.31)

ADL Long Form 0.23159

3 months 0.23 0.21 0.15 0.22 0.08 (-0.51,0.68) 6 months 0.33 0.22 0.52 0.23 -0.19 (-0.81,0.43) 12 months 1.03 0.31 0.76 0.32 0.27 (-0.60,1.15) 18 months 1.23 0.43 1.47 0.42 -0.25 (-1.44,0.94)

IADL Difficulty 0.30519

3 months 0.01 0.09 0.04 0.10 -0.03 (-0.30,0.24) 12 months 0.44 0.11 0.32 0.12 0.11 (-0.21,0.43) 18 months 0.75 0.17 0.42 0.16 0.33 (-0.13,0.79)

IADL Involvement 0.02797

3 months 0.14 0.11 0.09 0.12 0.06 (-0.27,0.38) 6 months 0.56 0.14 0.36 0.15 0.20 (-0.20,0.60) 12 months 0.75 0.15 0.50 0.15 0.24 (-0.18,0.67) 18 months 1.35 0.23 0.44 0.22 0.91 (0.28,1.54)

IADL Summary 0.04738

3 months 0.07 0.20 0.06 0.21 0.01 (-0.56,0.59) 6 months 0.68 0.25 0.38 0.26 0.30 (-0.42,1.02) 12 months 1.45 0.30 0.80 0.31 0.65 (-0.20,1.50) 18 months 2.27 0.45 0.68 0.42 1.59 (0.38,2.80)

SECONDARY RESULTS FINDINGS

PAGE 89

3.12.2 Presentation of EuroQoL Visual Analogue Scale (VAS), Cognitive Performance Scale (CPS), Depression Rating Scale (DRS), CHESS and Pain findings

The merits of the EuroQoL Visual Analogue Scale (VAS) have been previously discussed (Chapter II). However, as can be seen here and further, in the next section on predictive modelling, the tool when used in this setting appears to have little sensitivity to change. An improvement in VAS is indicated by an increase in the value which is different to the other scales presented here, whereby an improvement is represented by a decrease in the value.

There are little differences noted in any of the scales across any of the three sites, apart from a statistically significant reduction in screened depression over time in the Community FIRST group, when residential home data is excluded (P=0.008). Although the trend continues when residential home data is included, the statistical significance is lost.

Changes in the CHESS score occur over time in the Hutt dataset, when residential care is excluded. This is interesting as it presumably indicates that older people are less frail. However, a similar change is noted in the usual care group and no other factors such as ADL support this observation. Clinically, this is difficult to interpret, but may be influencing at some level the small reduction in mortality and residential home admission.

SECONDARY RESULTS FINDINGS

PAGE 90

Table 3-30: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for EuroQoL VAS, Cognitive Performance Scale (CPS), Depression Rating Scale (DRS) CHESS and Pain (Excluding residential home data) in Hamilton Hamilton

Usual Care New Service Treatment effect

(difference of differences) Treatment

* Visit Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Older person

EuroQoL Thermometer scale 0.73491 3 months -2.60 3.60 0.20 3.14 -2.80 (-12.30,6.69) 6 months -6.55 4.12 2.64 3.67 -9.19 (-20.10,1.73) 12 months -4.84 5.91 1.29 5.30 -6.14 (-21.88,9.61) 18 months -0.16 9.20 -0.57 7.70 0.41 (-23.25,24.07)

Cognitive Performance Scale (CPS) 0.65955

3 months 0.15 0.19 0.27 0.17 -0.12 (-0.62,0.39) 6 months 0.48 0.21 0.29 0.19 0.18 (-0.38,0.75) 12 months 0.34 0.27 0.25 0.24 0.09 (-0.64,0.81) 18 months 0.44 0.38 0.60 0.33 -0.16 (-1.16,0.83)

Depression Rating Scale (DRS) 0.00894

3 months -1.26 0.56 -0.95 0.47 -0.31 (-1.77,1.14) 6 months -0.82 0.63 -2.04 0.55 1.21 (-0.46,2.88) 12 months -2.23 0.88 -2.07 0.80 -0.16 (-2.55,2.23) 18 months 0.63 0.92 -1.69 0.80 2.32 (-0.20,4.85)

Changes in Health, End-stage disease and Signs and Symptoms (CHESS) 0.38148

3 months -0.30 0.16 -0.11 0.14 -0.19 (-0.61,0.22) 6 months -1.70 0.19 -1.69 0.16 -0.02 (-0.51,0.47) 12 months -0.78 0.25 -0.22 0.22 -0.57 (-1.23,0.09) 18 months -0.83 0.37 -0.16 0.32 -0.67 (-1.65,0.30)

Pain Scale 0.58082

3 months -0.15 0.19 0.12 0.16 -0.26 (-0.75,0.22) 6 months -0.25 0.22 -0.31 0.19 0.06 (-0.51,0.63) 12 months -0.28 0.31 -0.31 0.28 0.03 (-0.78,0.85) 18 months -0.37 0.47 -0.89 0.40 0.53 (-0.69,1.75)

SECONDARY RESULTS FINDINGS

PAGE 91

Table 3-31: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for EuroQoL VAS, Cognitive Performance Scale (CPS), CHESS and Pain (Including residential home data) in Hamilton Hamilton

Usual Care New Service Treatment effect

(difference of differences) Treatment

* Visit Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Older person

EuroQoL Thermometer scale 0.473983 months -2.19 3.15 0.90 2.93 -3.10 (-11.61,5.42) 6 months -3.76 3.44 3.08 3.25 -6.84 (-16.18,2.50) 12 months 2.64 4.71 -1.75 4.48 4.39 (-8.46,17.24) 18 months 0.51 7.28 3.55 6.45 -3.04 (-22.20,16.12)

Cognitive Performance Scale (CPS) 0.20873

3 months 0.26 0.19 0.29 0.17 -0.03 (-0.53,0.48) 6 months 0.72 0.20 0.40 0.19 0.32 (-0.22,0.87) 12 months 0.28 0.26 0.64 0.24 -0.36 (-1.06,0.33) 18 months 0.74 0.36 0.44 0.33 0.30 (-0.66,1.26)

Depression Rating Scale (DRS) 0.15063

3 months -1.14 0.47 -0.87 0.43 -0.27 (-1.53,1.00) 6 months -0.28 0.52 -1.77 0.48 1.49 (0.10,2.89) 12 months -2.18 0.69 -2.22 0.66 0.04 (-1.85,1.93) 18 months -0.44 1.03 -1.86 0.93 1.42 (-1.33,4.18)

Changes in Health, End-stage disease and Signs and Symptoms (CHESS) 0.23812

3 months -0.18 0.16 -0.11 0.15 -0.06 (-0.49,0.37) 6 months -1.81 0.12 -1.71 0.11 -0.10 (-0.42,0.23) 12 months -0.81 0.21 -0.16 0.20 -0.65 (-1.24,-0.05) 18 months -0.35 0.27 -0.37 0.25 0.02 (-0.76,0.79)

Pain Scale 0.26513

3 months -0.19 0.16 0.07 0.15 -0.26 (-0.70,0.17) 6 months -0.16 0.18 -0.34 0.17 0.18 (-0.30,0.67) 12 months -0.54 0.24 -0.42 0.23 -0.12 (-0.78,0.54) 18 months 0.03 0.35 -0.72 0.31 0.74 (-0.19,1.67)

SECONDARY RESULTS FINDINGS

PAGE 92

Table 3-32: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for EuroQoL VAS, Cognitive Performance Scale (CPS), CHESS and Pain (Excluding residential home data) in Hutt Hutt

Usual Care New Service Treatment effect

(difference of differences) Treatment

* Visit Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Older person

EuroQoL Thermometer scale 0.94167 3 months 3.37 2.77 4.34 2.70 -0.98 (-8.63,6.68) 6 months 2.32 2.95 1.56 2.87 0.76 (-7.39,8.91) 12 months -2.50 3.64 -0.43 3.16 -2.06 (-11.60,7.47) 18 months -1.86 6.29 -2.05 5.79 0.18 (-16.70,17.07)

Cognitive Performance Scale (CPS) 0.05503

3 months 0.12 0.13 -0.04 0.13 0.16 (-0.20,0.52) 6 months 0.34 0.17 0.05 0.16 0.28 (-0.18,0.74) 12 months 0.05 0.25 0.29 0.22 -0.25 (-0.91,0.42) 18 months 0.34 0.35 -0.28 0.31 0.62 (-0.34,1.59)

Depression Rating Scale (DRS) 0.14205

3 months -0.14 0.33 -0.13 0.33 -0.01 (-0.93,0.91) 6 months -1.09 0.36 -1.36 0.35 0.26 (-0.72,1.25) 12 months -1.45 0.44 -0.50 0.38 -0.95 (-2.10,0.20) 18 months -1.10 0.76 0.62 0.70 -1.72 (-3.76,0.33)

Changes in Health, End-stage disease and Signs and Symptoms (CHESS) 0.01772

3 months -0.44 0.15 -0.55 0.14 0.11 (-0.30,0.51) 6 months -1.69 0.15 -2.00 0.15 0.30 (-0.12,0.73) 12 months -1.23 0.19 -0.81 0.17 -0.42 (-0.92,0.08) 18 months -1.02 0.33 -1.68 0.31 0.66 (-0.23,1.55)

Pain Scale 0.35945

3 months 0.04 0.17 -0.11 0.17 0.15 (-0.33,0.63) 6 months 0.04 0.19 -0.04 0.18 0.07 (-0.44,0.59) 12 months 0.30 0.23 -0.25 0.20 0.55 (-0.05,1.14) 18 months 0.06 0.39 0.39 0.36 -0.33 (-1.38,0.71)

SECONDARY RESULTS FINDINGS

PAGE 93

Table 3-33: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for EuroQoL VAS, Cognitive Performance Scale (CPS), CHESS and Pain (Including residential home data) in Hutt Hutt

Usual Care New Service Treatment effect

(difference of differences) Treatment

* Visit Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Older person

EuroQoL Thermometer scale 0.504513 months 3.76 2.64 6.13 2.58 -2.37 (-9.65,4.90) 6 months 2.84 2.75 0.06 2.71 2.79 (-4.83,10.40) 12 months 0.02 2.98 3.00 2.82 -2.98 (-11.07,5.11) 18 months 1.21 5.84 -0.10 4.99 1.31 (-13.82,16.44)

Cognitive Performance Scale (CPS) 0.20031

3 months 0.15 0.14 0.13 0.14 0.03 (-0.37,0.42) 6 months 0.35 0.17 0.23 0.17 0.12 (-0.36,0.59) 12 months 0.38 0.22 0.45 0.21 -0.06 (-0.67,0.55) 18 months -0.33 0.37 0.58 0.32 -0.91 (-1.92,0.10)

Depression Rating Scale (DRS) 0.32436

3 months -0.11 0.32 -0.20 0.31 0.09 (-0.78,0.97) 6 months -0.87 0.33 -1.00 0.33 0.13 (-0.79,1.05) 12 months -1.08 0.36 -0.25 0.34 -0.83 (-1.81,0.14) 18 months -0.12 0.69 0.37 0.60 -0.48 (-2.29,1.32)

Changes in Health, End-stage disease and Signs and Symptoms (CHESS) 0.49081

3 months -0.42 0.12 -0.51 0.12 0.09 (-0.25,0.44) 6 months -1.82 0.12 -1.95 0.12 0.14 (-0.21,0.49) 12 months -0.89 0.18 -0.89 0.17 0.00 (-0.48,0.48) 18 months -1.51 0.27 -1.16 0.24 -0.35 (-1.09,0.40)

Pain Scale 0.99780

3 months -0.02 0.16 -0.18 0.15 0.16 (-0.27,0.59) 6 months 0.02 0.16 -0.10 0.16 0.13 (-0.33,0.58) 12 months -0.02 0.18 -0.18 0.17 0.17 (-0.31,0.64) 18 months 0.13 0.34 -0.08 0.29 0.21 (-0.67,1.10)

SECONDARY RESULTS FINDINGS

PAGE 94

Table 3-34: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for EuroQoL VAS, Cognitive Performance Scale (CPS), CHESS and Pain (Excluding residential home data) in Christchurch Christchurch

Usual Care New Service

Treatment effect (difference of differences)

Treatment * Visit

Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Older person

EuroQoL Thermometer scale 0.33164 3 months 0.88 1.36 -0.09 1.41 0.97 (-2.88,4.82) 6 months 0.44 1.41 3.05 1.44 -2.61 (-6.55,1.33) 12 months 4.97 1.54 4.84 1.55 0.13 (-4.16,4.42) 18 months 6.47 2.25 4.62 1.98 1.85 (-4.02,7.73)

Cognitive Performance Scale (CPS) 0.39744

3 months -0.10 0.07 -0.05 0.07 -0.05 (-0.24,0.13) 6 months 0.00 0.08 -0.06 0.08 0.06 (-0.17,0.29) 12 months 0.20 0.09 0.33 0.09 -0.13 (-0.37,0.11) 18 months 0.31 0.13 0.42 0.11 -0.11 (-0.44,0.22)

Depression Rating Scale (DRS) 0.48563

3 months -0.14 0.13 -0.33 0.14 0.19 (-0.18,0.57) 6 months 0.00 0.14 -0.32 0.14 0.32 (-0.07,0.70) 12 months -0.14 0.15 -0.57 0.15 0.43 (0.01,0.85) 18 months -0.69 0.23 -0.65 0.20 -0.05 (-0.64,0.54)

Changes in Health, End-stage disease and Signs and Symptoms (CHESS) 0.55594

3 months -0.15 0.07 -0.09 0.08 -0.06 (-0.27,0.14) 6 months -0.77 0.07 -0.89 0.07 0.12 (-0.07,0.31) 12 months -0.03 0.10 -0.02 0.10 -0.01 (-0.28,0.27) 18 months -0.07 0.14 -0.08 0.13 0.01 (-0.37,0.38)

Pain Scale 0.86535

3 months -0.06 0.08 -0.03 0.08 -0.03 (-0.26,0.19) 6 months 0.07 0.08 0.19 0.08 -0.12 (-0.35,0.12) 12 months 0.01 0.09 0.06 0.09 -0.05 (-0.31,0.20) 18 months 0.06 0.14 0.02 0.12 0.04 (-0.33,0.41)

SECONDARY RESULTS FINDINGS

PAGE 95

Table 3-35: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for EuroQoL VAS, Cognitive Performance Scale (CPS), CHESS and Pain (Including residential home data) in Christchurch Christchurch

Usual Care New Service

Treatment effect (difference of differences)

Treatment * Visit

Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Older person

EuroQoL Thermometer scale 0.279693 months 1.13 1.33 0.45 1.39 0.69 (-3.08,4.45) 6 months 0.08 1.35 3.24 1.40 -3.16 (-6.96,0.64) 12 months 4.55 1.45 5.13 1.48 -0.58 (-4.63,3.48) 18 months 6.46 2.12 5.60 1.91 0.86 (-4.74,6.47)

Cognitive Performance Scale (CPS) 0.68297

3 months -0.09 0.07 -0.04 0.07 -0.05 (-0.23,0.13) 6 months 0.03 0.08 0.00 0.09 0.03 (-0.20,0.26) 12 months 0.28 0.09 0.38 0.09 -0.10 (-0.34,0.14) 18 months 0.42 0.13 0.48 0.12 -0.05 (-0.39,0.28)

Depression Rating Scale (DRS) 0.83149

3 months -0.18 0.13 -0.36 0.14 0.18 (-0.19,0.55) 6 months -0.11 0.13 -0.31 0.14 0.20 (-0.18,0.57) 12 months -0.28 0.14 -0.59 0.15 0.31 (-0.09,0.71) 18 months -0.71 0.22 -0.74 0.19 0.03 (-0.54,0.60)

Changes in Health, End-stage disease and Signs and Symptoms (CHESS) 0.77118

3 months -0.14 0.07 -0.08 0.07 -0.06 (-0.26,0.14) 6 months -0.81 0.06 -0.88 0.06 0.06 (-0.11,0.24) 12 months 0.02 0.09 -0.01 0.09 0.03 (-0.22,0.28) 18 months -0.05 0.14 -0.13 0.12 0.09 (-0.27,0.45)

Pain Scale 0.79873

3 months -0.04 0.08 -0.04 0.08 0.00 (-0.23,0.22) 6 months 0.07 0.08 0.16 0.08 -0.09 (-0.32,0.13) 12 months -0.03 0.09 0.01 0.09 -0.04 (-0.28,0.20) 18 months 0.04 0.13 -0.04 0.12 0.08 (-0.27,0.43)

SECONDARY RESULTS FINDINGS

PAGE 96

3.12.3 Presentation of GP visit findings

General practice (GP) visits were recorded throughout the trial. There is an assumption that if older people with more frail needs are living at home, then there will be increased use of GP services. This was not supported by the ASPIRE trial findings with little differences being identified between groups

Table 3-36: Number of GP visits at each follow-up Hamilton Lower Hutt Christchurch

Usual Care Community

FIRST Usual Care PIP Usual Care COSE N* % 2 N* % 2 N* % 2 N* % 2 N* % 2 N* % 2 Baseline 0 16 28.07 21 38.18 15 28.3 17 32.69 52 28.73 60 35.5 1 21 36.84 23 41.82 20 37.74 20 38.46 91 50.28 82 48.522 15 26.32 6 10.91 8 15.09 7 13.46 22 12.15 17 10.063 or more 5 8.77 4 7.27 10 18.87 6 11.54 16 8.84 10 5.92 Unknown - - 1 1.82 - - 2 3.85 - - - - 3 months

0 15 40.54 17 34.69 7 17.95 7 17.07 46 27.38 56 35.441 16 43.24 17 34.69 23 58.97 21 51.22 82 48.81 66 41.772 2 5.41 6 12.24 2 5.13 4 9.76 25 14.88 20 12.663 or more 1

2.70 5 10.20 3

7.69 5 12.207

4.17 9 5.70

Unknown 3 8.11 4 8.16 4 10.26 4 9.76 8 4.76 7 4.43 12 months

0 7 53.85 6 37.50 7 30.43 7 21.88 42 32.56 41 31.781 3 23.08 4 25.00 8 34.78 16 50.00 61 47.29 60 46.512 1 7.69 3 18.75 3 13.04 2 6.25 9 6.98 9 6.98 3 or more 1

7.69 2 12.50 1

4.35 1 3.13 5

3.88 6 4.65

Unknown 1 7.69 1 6.25 4 17.39 6 18.75 12 9.3 13 10.08 18 months

0 3 60.00 4 57.14 3 37.50 2 22.22 20 36.36 23 33.821 2 40.00 3 42.86 3 37.50 3 33.33 13 23.64 26 38.242 - - - - - - 2 22.22 6 10.91 8 11.763 or more - - - - -

- - - 7

12.73 5 7.35

Unknown - - - - 2 25 2 22.22 9 16.36 6 8.82

SECONDARY RESULTS FINDINGS

PAGE 97

3.12.4 Sub-group analysis

No statistically significant results were observed in any of the sub-group analyses in any region. Sub-group analysis was undertaken for: Needs Level (High or Very High); CHESS score; age group (65-75, 75-85, 85+); Gender; Ethnicity and; presence or absence of dementia15.

3.13 Informal caregiver

The role of the caregiver can not be understated and this is supported by the findings presented in the next section, whereby there is a statistically significant association between increasing caregiver burden as measured by the caregiver reaction assessment (CRA) and a rising risk of residential home admission. Caregivers in the ASPIRE population ranged from their adult children and their partners to spouses. No comprehensive analysis on the role of caregivers has been undertaken at this stage. This section does however present important data around caregiver stress levels.

The CRA measures caregiver stress or burden of care (Chapter II) and is a 24-item scale where the maximum score of 120 indicates maximum stress according to the scale. The CRA results are a vital aspect to assessing the relative success of the AIPI. There was no detectable difference in caregiver stress between the AIPI and usual care at any of the three sites. There is a possible non-significant trend in a reduction of caregiver stress in Hamilton and Hutt over time.

The SF36 has two components (physical and mental) and is discussed in Chapter II. The physical component includes physical, role, pain and general health, where as the mental component includes mental, role, social function and vitality. Both values are presented in the following tables and a reduction in scores indicates deterioration. The participants in Hamilton from baseline values are clearly the most disabled and frail out of the three sites and also the caregiver roles are more often being filled by their adult progeny and therefore it is likely that the impact of caring would be most apparent in this region. However, the

15 As previously discussed, this is a very difficult condition to accurately diagnose and define in the research

SECONDARY RESULTS FINDINGS

PAGE 98

adult children of the older person may be typically physically stronger to cope with the role of caregiving than spouses. Tables 3-37 and 3-38 illustrate that there is no overall treatment effect or per site on the SF36 physcial and mental domains. An interesting trend, whereby there is deterioration in the physical component of SF-36 for participants in the new service in Hamilton. It is difficult to draw firm conjectures from this finding, however, it may well be that caregivers of older people receiving usual care find the role physically challenging, though rewarding and is an area for further discussion. No trends are observed in either Hutt or Christchurch in terms of SF36.

The EuroQoL 5D has two components, the first includes five questions relating to self perceived health related quality of life and the second pertains to the 100 item temperature scale (Visual Analogue Scale). The caregiver was asked to complete the EuroQoL 5D on behalf of the older person, the results of the VAS are presented here, therefore an increase in values represents an improvement. The scale should be considered in light of the responses from the older person presented in Section 3.12.

It is difficult to draw any firm conclusions from the findings and no statistically significant results were observed.

SECONDARY RESULTS FINDINGS

PAGE 99

Table 3-37: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for Caregiver SF36, Caregiver Reaction Assessment and EuroQoL perception of older person (Excluding residential home data) in Hamilton Hamilton

Usual Care New Service Treatment effect

(difference of differences) Treatment

* Visit Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Caregiver

SF-36 Physical

Component Scale (PCS) 0.15687 3 months 0.08 1.55 1.44 1.37 -1.36 (-5.45,2.73) 6 months -1.89 1.76 0.23 1.58 -2.12 (-6.83,2.58) 12 months -1.14 2.46 -0.18 2.33 -0.96 (-7.66,5.74) 18 months 0.84 3.83 -9.69 3.81 10.53 (-0.14,21.21) Mental Component

Scale (MCS) 0.03854 3 months 4.87 1.85 1.82 1.63 3.06 (-1.86,7.97) 6 months 0.45 2.11 0.52 1.89 -0.07 (-5.71,5.58) 12 months 4.85 2.82 -1.27 2.65 6.11 (-1.58,13.80) 18 months -4.28 4.14 6.55 4.11 -10.82 (-22.40,0.76)

Caregiver Reaction

Assessment (CRA) 0.77973 3 months -2.73 1.33 -1.77 1.17 -0.96 (-4.48,2.56) 6 months -3.01 1.53 -3.36 1.38 0.35 (-3.72,4.43) 12 months -2.34 2.13 0.38 2.01 -2.73 (-8.53,3.08) 18 months -9.79 1.87 -8.34 1.78 -1.45 (-6.80,3.90)

EuroQoL Caregiver (How they feel about OP's health)

Thermometer scale 0.96455 3 months 0.51 3.73 5.13 3.29 -4.61 (-14.47,5.24) 6 months 2.87 4.36 5.24 3.92 -2.37 (-13.97,9.23) 12 months 9.26 6.19 9.30 5.88 -0.04 (-16.94,16.86) 18 months 2.91 9.63 5.13 9.56 -2.23 (-29.08,24.63)

SECONDARY RESULTS FINDINGS

PAGE 100

Table 3-38: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for Caregiver SF36, Caregiver Reaction Assessment and EuroQoL perception of older person (Including residential home data) in Hamilton Hamilton

Usual Care New Service Treatment effect

(difference of differences) Treatment

* Visit Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Caregiver

SF-36 Physical

Component Scale (PCS) 0.03885 3 months 0.17 1.30 1.87 1.27 -1.70 (-5.29,1.90) 6 months -0.68 1.44 1.08 1.42 -1.76 (-5.75,2.23) 12 months -1.28 1.93 -0.91 1.98 -0.36 (-5.83,5.10) 18 months 3.26 2.87 -7.23 3.12 10.49 (2.12,18.86) Mental Component

Scale (MCS) 0.01500 3 months 3.81 1.60 2.18 1.55 1.63 (-2.79,6.05) 6 months 1.04 1.77 0.80 1.74 0.24 (-4.70,5.18) 12 months 7.18 2.39 0.03 2.45 7.15 (0.38,13.93) 18 months -0.78 3.57 10.28 3.89 -11.06 (-21.50,-0.62)

Caregiver Reaction

Assessment (CRA) 0.52738 3 months -1.65 1.25 -3.15 1.21 1.50 (-1.94,4.95) 6 months -2.06 1.36 -4.15 1.35 2.09 (-1.70,5.88) 12 months -3.48 1.80 -2.32 1.84 -1.15 (-6.24,3.93) 18 months -7.49 1.71 -10.62 1.79 3.12 (-1.92,8.16)

EuroQoL Caregiver (How they feel about OP's health)

Thermometer scale 0.463503 months -0.10 3.11 4.86 3.01 -4.96 (-13.51,3.59) 6 months 1.63 3.47 4.63 3.42 -3.00 (-12.63,6.62) 12 months 9.37 4.90 3.73 5.06 5.64 (-8.27,19.55) 18 months -0.96 7.61 10.13 8.31 -11.09 (-33.32,11.15)

SECONDARY RESULTS FINDINGS

PAGE 101

Table 3-39: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for Caregiver SF36, Caregiver Reaction Assessment and EuroQoL perception of older person (Excluding residential home data) in Hutt Hutt

Usual Care New Service Treatment effect

(difference of differences) Treatment

* Visit Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Caregiver

SF-36 Physical

Component Scale (PCS) 0.00031 3 months 0.53 1.18 -0.69 1.23 1.22 (-2.23,4.66) 6 months 2.43 1.78 1.84 1.88 0.60 (-4.67,5.86) 12 months -2.24 1.79 -1.66 1.77 -0.58 (-5.72,4.56) 18 months -16.10 2.85 -1.60 2.62 -14.49 (-22.75,-6.24) Mental Component

Scale (MCS) 0.76997 3 months -1.09 1.45 0.40 1.53 -1.50 (-5.73,2.74) 6 months 0.97 2.06 1.23 2.19 -0.26 (-6.33,5.81) 12 months -2.29 3.20 2.37 3.16 -4.66 (-13.87,4.55) 18 months -11.12 4.46 -4.78 4.06 -6.35 (-18.88,6.19)

Caregiver Reaction

Assessment (CRA) 0.63402 3 months -1.21 1.14 -2.30 1.20 1.09 (-2.19,4.38) 6 months -0.14 1.23 -1.59 1.30 1.44 (-2.11,5.00) 12 months 0.92 1.48 -2.52 1.44 3.44 (-0.66,7.54) 18 months 5.50 3.03 0.66 2.37 4.83 (-2.79,12.46)

EuroQoL Caregiver (How they feel about OP's health)

Thermometer scale 0.99566 3 months -1.42 3.22 -1.09 3.40 -0.33 (-9.68,9.03) 6 months 2.82 3.44 2.00 3.65 0.82 (-9.16,10.80) 12 months -2.93 4.00 -2.81 3.97 -0.12 (-11.32,11.07) 18 months -5.09 7.68 -4.31 6.13 -0.79 (-20.28,18.70)

SECONDARY RESULTS FINDINGS

PAGE 102

Table 3-40: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for Caregiver SF36, Caregiver Reaction Assessment and EuroQoL perception of older person (Including residential home data) in Hutt Hutt

Usual Care New Service Treatment effect

(difference of differences) Treatment

* Visit Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Caregiver

SF-36 Physical

Component Scale (PCS) 0.00002 3 months 0.96 1.22 -0.42 1.29 1.38 (-2.22,4.97) 6 months 3.89 1.66 2.75 1.73 1.14 (-3.70,5.98) 12 months -2.04 1.85 0.10 1.88 -2.14 (-7.50,3.22) 18 months -16.82 2.99 -0.01 2.74 -16.81 (-25.21,-8.41) Mental Component

Scale (MCS) 0.23398 3 months -0.19 1.85 0.56 1.97 -0.75 (-6.12,4.62) 6 months 1.71 1.93 -0.16 2.04 1.88 (-3.68,7.44) 12 months -1.02 2.09 3.47 2.16 -4.49 (-10.46,1.47) 18 months -4.51 5.10 -1.49 3.46 -3.02 (-15.21,9.17)

Caregiver Reaction

Assessment (CRA) 0.49774 3 months -1.28 1.18 -2.33 1.26 1.05 (-2.40,4.50) 6 months -1.39 1.24 -1.41 1.30 0.02 (-3.55,3.60) 12 months -0.98 1.32 -2.76 1.37 1.77 (-2.02,5.56) 18 months 4.79 3.12 -0.90 2.13 5.69 (-1.81,13.19)

EuroQoL Caregiver (How they feel about OP's health)

Thermometer scale 0.210903 months -2.99 2.77 -1.73 2.97 -1.25 (-9.41,6.90) 6 months 1.06 3.56 0.76 3.77 0.30 (-10.11,10.71) 12 months -1.99 3.36 -9.46 3.39 7.47 (-2.17,17.11) 18 months -10.10 3.57 -10.12 3.11 0.02 (-11.81,11.84)

SECONDARY RESULTS FINDINGS

PAGE 103

Table 3-41: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for Caregiver SF36, Caregiver Reaction Assessment and EuroQoL perception of older person (Excluding residential home data) in Christchurch Christchurch

Usual Care New Service Treatment effect

(difference of differences) Treatment

* Visit Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Caregiver

SF-36 Physical

Component Scale (PCS) 0.07479 3 months 0.25 1.21 -0.20 1.29 0.44 (-3.02,3.91) 6 months -2.35 1.24 0.56 1.33 -2.91 (-6.49,0.66) 12 months 0.34 1.35 -1.62 1.53 1.96 (-2.03,5.95) 18 months -1.88 2.07 0.17 1.86 -2.04 (-7.51,3.42) Mental Component

Scale (MCS) 0.63887 3 months 0.55 1.36 -1.93 1.48 2.48 (-1.46,6.43) 6 months 0.71 1.41 -1.50 1.54 2.21 (-1.89,6.31) 12 months -2.02 1.57 -1.34 1.82 -0.69 (-5.41,4.04) 18 months -1.10 2.43 -3.27 2.19 2.17 (-4.25,8.58)

Caregiver Reaction

Assessment (CRA) 0.99629 3 months -0.85 0.89 -0.27 0.96 -0.58 (-3.16,2.00) 6 months 0.20 0.92 0.66 0.99 -0.45 (-3.11,2.21) 12 months 0.97 1.00 1.66 1.14 -0.69 (-3.67,2.28) 18 months 1.68 1.51 2.57 1.37 -0.89 (-4.89,3.11)

EuroQoL Caregiver (How they feel about OP's health)

Thermometer scale 0.61426 3 months 1.96 2.33 2.90 2.53 -0.94 (-7.71,5.82) 6 months 2.26 2.43 3.88 2.64 -1.61 (-8.67,5.44) 12 months -1.68 2.66 4.38 3.06 -6.06 (-14.02,1.90) 18 months -0.76 3.88 -0.60 3.53 -0.16 (-10.47,10.15)

SECONDARY RESULTS FINDINGS

PAGE 104

Table 3-42: Treatment effect for Scale Measurements Change from Baseline (Mixed model) at each visit for Caregiver SF36, Caregiver Reaction Assessment and EuroQoL perception of older person (Including residential home data) in Hamilton Christchurch

Usual Care New Service

Treatment effect (difference of differences)

Treatment * Visit

Interaction

change from

baseline SE

change from

baseline SE Mean 95%CI p-value Caregiver

SF-36 Physical

Component Scale (PCS) 0.06341 3 months 0.21 1.14 -0.19 1.25 0.40 (-2.91,3.71) 6 months -2.13 1.17 0.79 1.27 -2.93 (-6.31,0.45) 12 months -0.02 1.24 -1.46 1.36 1.44 (-2.18,5.05) 18 months -2.24 1.89 0.36 1.72 -2.61 (-7.62,2.40) Mental Component

Scale (MCS) 0.77833 3 months 0.24 1.44 -1.54 1.57 1.78 (-2.39,5.95) 6 months 0.27 1.48 -0.82 1.60 1.09 (-3.17,5.35) 12 months -1.25 1.58 -0.68 1.73 -0.57 (-5.17,4.03) 18 months -0.30 2.34 -2.71 2.14 2.41 (-3.81,8.63)

Caregiver Reaction

Assessment (CRA) 0.91180 3 months -0.81 0.89 -0.21 0.98 -0.60 (-3.19,1.99) 6 months 0.23 0.91 0.68 0.99 -0.45 (-3.09,2.20) 12 months 0.85 0.97 0.79 1.07 0.07 (-2.76,2.89) 18 months 1.21 1.44 2.48 1.33 -1.27 (-5.10,2.57)

EuroQoL Caregiver (How they feel about OP's health)

Thermometer scale 0.181293 months 2.28 2.43 3.17 2.65 -0.89 (-7.95,6.18) 6 months 2.40 2.49 4.46 2.70 -2.06 (-9.27,5.15) 12 months -2.72 2.66 5.74 2.93 -8.46 (-16.24,-0.69) 18 months -1.75 3.55 0.20 3.34 -1.95 (-11.53,7.63)

SECONDARY RESULTS FINDINGS

PAGE 105

3.14 Summary

The secondary results are somewhat complex to interpret though do provide some level of explanation for the primary results findings. The reduction in risk of residential home admission in participants enrolled in the Community FIRST programme may be explained by a statistically significant treatment effect in activities of daily living and a clear trend for improvement in activities of daily living over time, and may be a product of the restorative focus of the programme. There did not appear to be any trends for functional gains in either the Hutt PIP initiative or COSE for ADL function. However, COSE did result in a treatment effect for Instrumental Activities of Daily Living, whereby older people in COSE deteriorated at a lower rate. COSE is a NASC replacement care management programme and utilises standard home care services that do not have a promoting independence or restorative focus and therefore it is not surprising that the initiative did not demonstrate an improvement in ADL or IADL. It is surprising however that despite this the programme resulted in such a statistically significant reduction in residential home placement as presented in Section 1 of this chapter, and this may reflect the enhanced coordination with primary healthcare by liaising with general practitioners on their older patients, whereas NASC does not.

There appeared to be few trends in the quality of life of older people in any of the groups, though this may be an artefact of the publicised difficulties in assessing quality of life in older people.

Finally, there is a real risk that supporting an older person with high and complex needs at home increases caregiver burden. This was not found to be the case in any of the three initiatives evaluated in ASIRE as measured by the Caregiver Reaction Assessment. However, a deterioration was noted in the physical components of health related quality of life for usual care in Hamilton.

PREDICTIVE MODELLING FINDINGS

PAGE 106

Section 4: Analysis of risks of entry to residential care and hospitalisation (predictive modelling)

3.15 Introduction

Predictive modelling is the process by which a statistical model is created or chosen to try to best predict the probability of an outcome. In this instance, the MDS-HC Home Care Quality Indicators were utilised in order to explore their potential relationship with firstly hospitalisation and secondly permanent residential home admission. In addition, the EuroQoL Visual Analogue Scale (VAS) a scale from 0 to 100 which depicts how the individual feels at that point in time as well as the Caregiver Reaction Assessment.

Predictive modelling was undertaken in the ASPIRE study using the

30 MDS-HC Home Care Quality Indicators as well as the EuroQoL

Visual Analogue Scale and Caregiver Reaction Assessment. Hazard

Risk ratios were calculated for these variables using hospitalisation

and residential home admission as primary endpoints.

No medication review, negative mood and previous hospitalisation

were correlated with increasing the risk of hospitalisation. Where as,

inadequate meals, dehydration, ADL/rehab potential with no

therapies, failure to improve/incidence of decline in ADL, social

isolation, caregiver stress (CRA), negative mood and delirium were all

correlated with an increased risk of residential home placement.

Interestingly, when there is a failure to improve / incidence of decline

in ADL in an older person, they are over 11 times more likely to be

admitted to a residential home. Such a result is highly pertinent for

the increasing interest in restorative home support.

These findings are useful in the development of an evidence-based

service specification for ageing-in-place initiatives or indeed

restorative home support contracts.

PREDICTIVE MODELLING FINDINGS

PAGE 107

The findings16 are presented as Tables which provide the Hazard Ratio estimates adjusted for age, sex, ethnicity and treatment effect with 95% Confidence Intervals (CI) and p-values for both Hospitalisation and Residential Care Entry as the Primary end-points. Hazard Ratio (HR) estimates are interpreted as the magnitude of risk for having versus not having the risk factor except for CRA and EuroQoL. For example, the risk of entering Residential Care for people who have inadequate meals in Table 3-44 is 2.18 times higher than those who have adequate meals with 95% confidence interval falls between (CI: 1.15, 4.13). Although, HR estimates are provided for each variable, only those that are statistically significant at the 5% level are highlighted.

Predictive modelling is particularly useful in assisting in the development of service specifications for new home based services, where residential home admission and hospitalisation need to be avoided. For instance, the statistically significant predictors can be incorporated as key performance indicators and for this reason; these findings are particularly relevant for ongoing development of ageing in place initiatives.

Note. The predictive modelling exercise was undertaken on a sample set of 524 rather than the full 569 as a result of the need to run the enquiry prior to data lock down.

3.16 The risk of hospitalisation

Table 3-43 (over page) illustrates that no medication review (HR 1.9, or a 1.9 increased risk of hospitalisation if an older person does not have a medication review within the last 6 months), negative mood (HR 1.5) and previous hospitalisation within the last 90 days (HR 1.8) are all correlated with the risk of hospitalisation.

16 For the matters of expediency, the predictive modelling exercise was undertaken on an interim dataset

PREDICTIVE MODELLING FINDINGS

PAGE 108

Table 3-43: Hazard Ratio estimates with 95% CIs and p-values, Hospitalisation as the primary end-point

Possible Indicators Hazard Ratio

Lower CI

Upper CI

p-value

Prevalence of inadequate meals 1.06 0.58 1.94 0.8574 Prevalence of weight loss 1.17 0.57 2.42 0.6724 Prevalence of dehydration 1.01 0.67 1.51 0.9690 Prevalence of no medication review 1.934 1.141 3.278 0.0144 Failure to improve/incidence of bladder incontinence 0.792 0.311 2.019 0.6255

Failure to improve/incidence of skin ulcers 0.000 0.000 0.000 0.9935

Prevalence of no assistive devices among clients with mobility difficulties 2.267 0.631 8.145 0.2098

Prevalence of ADL/rehab potential with no therapies 0.647 0.390 1.074 0.0919

Failure to improve/incidence of decline in ADL 0.488 0.144 1.650 0.2484

Failure to improve/incidence of improve/incidence of impaired mobility at home 0.558 0.225 1.383 0.2077

Prevalence of falls 0.662 0.282 1.553 0.3432 Prevalence of social isolation 0.834 0.550 1.263 0.3909 Failure to improve/incidence of cognitive decline 2.447 0.794 7.540 0.1191

Prevalence of delirium 1.406 0.891 2.217 0.1429 Prevalence of negative mood 1.515 1.024 2.241 0.0376 Failure to improve/incidence of difficulty in communication 1.173 0.461 2.988 0.7376

Prevalence of disruptive or intense daily pain 1.108 0.738 1.663 0.6210

Prevalence of inadequate pain control 0.924 0.502 1.700 0.7995 Prevalence of neglect/abuse 1.956 0.718 5.328 0.1894 Prevalence of injuries 1.406 0.946 2.088 0.0915

Prevalence of not receiving flu vaccinations - - - -

Prevalence of hospitalisation 1.794 1.196 2.690 0.0047 Status of family care using CRA 1.025 0.994 1.057 0.1205

EuroQoL Self-assessment thermometer 0.991 0.981 1.000 0.0537

PREDICTIVE MODELLING FINDINGS

PAGE 109

3.17 The risk of residential home admission

The predictive modelling exercise around the residential home admission endpoint provides fascinating reading. Several findings can be highlighted. Firstly, the importance of nutrition and fluid intake is reinforced, whereby inadequate meals results in a 2.18 times increased risk of residential home admission and dehydration a 1.74 increase in risk. The findings around delirium are not surprising given the link with acute infection. However, it is disturbing to see the relationship nonetheless, particularly given the potentially reversible nature of delirium.

The extremely strong relationship between residential home admission and failure to prevent a decline in ADL function (i.e. an individual is 11 times more likely to be admitted to a residential home) provides considerable support for the restorative support philosophy particularly in light of the finding that therapy input has no effect. Although superficially, it may appear that a lack of therapy (specifically, occupational therapy and physiotherapy) input actually reduces the risk of residential home admission by 64%, a more likely explanation is that older people receive therapy input when there is an existing major disability and therefore need and therefore the risk of residential home admission is already high. However, what is clear is that physiotherapy and occupational therapy input does not appear to protect the older person from reaching this endpoint, which needs discussion.

Further, the findings around social isolation (HR 1.86) and negative mood (HR 2.17) provides further weight behind the need to develop services that focus on reintegrating older people back into their community, as is observed in the Community FIRST initiative in Hamilton and the COSE model in Christchurch.

For the Caregiver Reaction Assessment (CRA) scale, one unit increase will increase/decrease the risk by the magnitude of Hazard Ratio estimates. For example, one unit of CRA increase will increase the risk of entering Residential Care by 7 % with 95% confidence interval falls between (3%, 11 %). Such a finding illustrates the importance of the caregiver in maintaining the older person at home and ensuring that caregiver burden is minimised.

PREDICTIVE MODELLING FINDINGS

PAGE 110

Table 3-44: Hazard Ratio estimates with 95% CIs and p-values, Residential Care Entry as the primary end-point

Possible Indicators Hazard Ratio

Lower CI

Upper CI p-value

Prevalence of inadequate meals 2.18 1.15 4.13 0.0166 Prevalence of weight loss 1.63 0.69 3.85 0.2609 Prevalence of dehydration 1.74 1.04 2.92 0.0347 Prevalence of no medication review 1.23 0.52 2.89 0.6420 Failure to improve/incidence of bladder incontinence 3.39 0.81 14.20 0.0945

Failure to improve/incidence of skin ulcers 0.00 0.00 0.00 0.9975

Prevalence of no assistive devices among clients with mobility difficulties 0.00 0.00 0.00 0.9906

Prevalence of ADL/rehab potential with no therapies 0.38 0.20 0.71 0.0027

Failure to improve/incidence of decline in ADL 11.07 2.57 47.74 0.0013

Failure to improve/incidence of improve/incidence of impaired mobility at home

1.68 0.38 7.51 0.4961

Prevalence of falls 1.76 0.67 4.58 0.2508 Prevalence of social isolation 1.86 1.11 3.11 0.0190 Failure to improve/incidence of cognitive decline 1.04 0.26 4.27 0.9522

Prevalence of delirium 3.65 2.16 6.18 0.0000 Prevalence of negative mood 2.17 1.29 3.65 0.0034 Failure to improve/incidence of difficulty in communication 0.94 0.23 3.88 0.9307

Prevalence of disruptive or intense daily pain 1.21 0.71 2.08 0.4819

Prevalence of inadequate pain control 1.07 0.50 2.30 0.8670 Prevalence of neglect/abuse 2.41 0.75 7.77 0.1418 Prevalence of injuries 1.55 0.93 2.61 0.0949 Prevalence of not receiving flu vaccinations - - - -

Prevalence of hospitalization 1.11 0.66 1.85 0.7048 Status of family care using CRA 1.07 1.03 1.11 0.0006 EuroQoL Self-assessment thermometer 1.07 0.97 1.18 0.2048

OPERA FINDINGS

PAGE 111

Section 5: The impact of moving into a residential facility

3.18 Introduction

The OPERA (Older People Entering Residential Accommodation) utilised qualitative research methods to determine the influences which surround residential care entry and the older person’s satisfaction with the decisions regarding their permanent residence. In order to have a robust model, the findings have been sequentially structured through three phases.

17 OPERA is a separately funded piece of research work and sits outside the remit of ASPIRE

OPERA (Older People Entering Residential Accommodation)

provides the in-depth qualitative analysis around the ASPIRE study.

It is very clear that few quality of life indicators are appropriate for

older people in New Zealand and therefore qualitative findings and

the conclusions drawn from such are highly pertinent for this

population.

The findings of this ASPIRE sub-study17 indicates that there are a

number of factors that are highly important to the enrolled population

of older people such as: coping, support, decisions and place of

residence themes. The study also explored the process around

decision making in relation to placement and there appeared to be

some disagreement around who made the decision for residential

home placement, with the older person feeling that they were the

main decision maker, though both family and NASC felt that the

family was.

For the 131 older people interviewed, of those that had relocated to a

residential home 47% reported as being sad or very sad around the

decision to move, whereas 75% of people living in their own home

were happy or very happy with their decision to remain living at home.

OPERA FINDINGS

PAGE 112

Phase 1: The literature review, gave the grounding to the study by identifying the existing key decision makers and factors which influenced the older people’s residential care entry. The major influencing factors from the overseas literature reviews were: being over 80, living alone, being ‘white’, having difficulty with daily living functions and cognitive problems.

Phase 2: This was an explorative qualitative pilot study involving 13 older people, their caregivers, a Needs Assessment Co-ordination Service (NASC) and a Multi-disciplinary team (MDT). The analysis of the pilot identified three main themes: change, control and placement.

Phase 3: Phase 3 was developed from the pilot study and involved a group of 131 older people (who were enrolled in the larger ASPIRE study) with an initial and a follow up visit, and where possible their caregivers and a NASC. The analysis developed four themes; coping, support, decisions and residence, which formed a sequential progression from the initial event, to the older persons’ permanent residence.

3.19 Findings

A number of themes arose from the study and these can broadly be defined as the following: Change, control, placement, coping and support.

3.19.1 Change

The change theme described the triggers for the review of the existing support needs for the older person. The following spouses’ comments illustrated some of the categories within this theme, such as social implications.

“Primarily the problems were getting worse. We tried getting help in and it just wasn’t working…so they decided that I could get relief from [his wife] by her going into a rest home” (Spouse)

Also one husband who was managing risk and his wife’s dementia with support said:

She couldn’t live on her own unless you got a person to live in nearly 24 hours of the day. That is of course if I dropped down dead” (Husband).

OPERA FINDINGS

PAGE 113

3.19.2 Control

This theme referred to the person, (usually the older person, doctor, or family) who made the decision about the permanent residence of the older person. In most cases, when the older person was in hospital, they remembered a doctor talking about what they could not do, which appeared to be a one-way conversation. One caregiver illustrated this point by saying:

“The services were offered, not explained, ….He did not understand the language; they spoke in medical-speak ….Professional services ….give the person no choices….The process to know everything is very long and fraught with difficulty, it is like a snake pit.” (Niece).

Many older people had become resigned to the decision made for them, but there was a notable exception. One lady, post cerebro-vascular accident with dysphasia, asked the MDT on their weekly ward round if she could go home:

“I managed to say to the doctor ‘go home’...and he laughed at me and said ….you will go to a rest home.…The doctor then said if you can walk from your bed to the door then you can go home. My friend grabbed hold of my feet and slowly stood me up ...it took a whole day, but I got to the door.” (Older person)

This lady self-discharged later that day and is now living with support in her own home. “No alternatives really”, was often said, which points to communication and knowing what support was available within the community. One person felt that she had to go into a rest home because she didn’t want to encumber her family.

“Well there is no alternative really unless the family come into it…. you can’t go and live with them you know, they have their lives to live and that is what I mean about alternatives” (Older person)

3.19.3 Placement

Placement described some thoughts about the older person’s residence, such as adjusting to relocation, environmental hazards, the support that they received and financial issues. Cost was a perceived barrier to both residential care, and home.

OPERA FINDINGS

PAGE 114

“He was scared about [residential care] costs, so he just said he was fine,

he would stay at home, but he is not fine.” (Niece)

and,

“The cost of that [looking after the older person at home] is high also. They have to be driven to the doctor and to get medicines etc. If a person stays at home to look after their parent they do not get much financial support” (M.D.T.)

3.19.4 Coping

This was the way the older person managed, their situation including their disabilities, and outside forces which interacted with them. For clarification coping and its categories have been divided into two main subheadings:

Physical issues

A major part of coping was living and dealing with health concerns. This has been expressed by participants in a general and more specific manner such as:

“I feel my health is getting worse and soon I will have to go into a rest home” (Older person).

and more specifically:

“It is difficult for me now that my eyesight is so bad. I get letters from my daughter and I can’t read them. I open the letter and kiss it and then wait for my son to come and read it for me” (Older person).

The mobility category encompassed the ability to move from one place to another, and be able to get to basic areas such as the toilet or kitchen. While outside the house, it was the ability to clear the mail box, or go to the local shops which caused frustration. In some cases it was the family’s concern about the mobility and subsequent falls which predetermined residential care entry.

“The family wanted me to go in here [residential care]…..they thought that I would fall, they have always had that fear that I would fall and no one would hear and I would be left there” (Older person)

OPERA FINDINGS

PAGE 115

Two other important factors which older people had to cope with were ‘dementia,’ and ‘incontinence’.

“We do not want to move, my wife has dementia and I look after her, but we will stay here for as long as we can” (Older person)

and

“The nurse from the rest home came over to my unit [in a retirement village] and said that since I was incontinent that I should not be here, I should be in the rest home” (Older person).

Emotional issues

The over-riding feeling among the older people was a sense of loss, often cause by a death, relocation to a smaller unit, or residential care. Some missed their youth and the way they ‘used to be.’ One rest home resident commented:

“I would have liked to continue in my own home. It still makes me sad when I think about it. One of the saddest things was that I had to leave my cats. I had two lovely cats….. They took my cats away from me, I don’t know what has happened to them, and I don’t like to ask. They were really my family. … I do so much miss them. I am still sad about all this and in here I get no sun in my room and just sit here all day…. I guess you could say I am home sick” (Older person).

Loneliness was another of the emotional issues which was found both at home, and within residential care. At home it was the lack of family and friends, or the opportunities to get out and mix with others. At times family would move the older person to be near them, and in doing so immediately dislocate them from their previous peer group. Families were invariably busy with jobs, children and social activities. While they offered support to the older person, there were plenty of hours in the day and night where the older person had the potential to be lonely. In the rest home it was often a feeling of isolation, within a sea of people. People would be isolated in their rooms, either voluntarily or involuntarily, or isolated within a group due to an inability to communicate. The reasons for this were either extrinsic or intrinsic, and could be the inability of one or more people to communicate due to deafness, speech deficits, mental confusion, or all three. One person living in a rest home commented:

OPERA FINDINGS

PAGE 116

“I wish that I had died when I had my last stroke. I don’t have any friends here in this place. I sit at the table with five others and none of them speak. Two are deaf, one just answers in monosyllables and the other sleeps all the time. I just stay in my room and watch television because there is little else that I can do when I have no friends” (Older person).

A feeling of impending death, led to a lack of desire to do anything, make any changes, and to just accepting the inevitable closure to their lives. There were comments like “I won’t be here much longer,” “now I see no future for myself and would like to just fade out” and “there is nothing to live for now, so I have just given up.” For some the ‘not wanting to be a burden’ was their reason for residential care entry.

“I found that I was relying too much on my family …so I thought that it was time for me to make a move. So that is why I came in” (Older person).

Coping had a lot to do with adequate support, and a good feeling of self efficacy in a supportive environment.

3.18.5 Support

The support theme talked about the assistance given to the older person, which enabled them to live within a safe and fulfilling environment. This theme has been divided into two categories, ‘family and friends,’ and ‘community support services.’

Family and friends

This category described how the older person felt about the support from their family. Many of the people who had spouses had been living with them for upwards of 40 years and a separation, either by death, or residential care was clearly difficult. Children being busy with their own families or work, and not having much time were constant comments. Because children were busy, the older people hesitated to ‘bother’ them. Some older people did not feel close to their children, and therefore did not get much assistance from them. One older person commented that her daughter said:

OPERA FINDINGS

PAGE 117

“I didn’t leave nursing to turn around and look after elderly people again, not even my own mother.” (Older person)

Although in many cases the caregivers were managing with extra support when needed, there was caregiver stress evident in some situations. Some spouses, or children, were themselves not well which increased the stress levels. The N.A.S.C. described one older person who entered residential care because of caregiver stress, as follows:

“She was categorised at hospital level care. She had home help and daily assistance. Her husband called us to say that he was unable to manage her any more, so he suggested rest home” (N.A.S.C.).

Community support services

Most of the older people in the study who were living at home were accessing some form of community support, mostly home help or showering. Some people were really happy with the home service provided, while others were frustrated, and said it was really hard to get to know what was available, and then really hard to get it, get it organized, and finally get it running smoothly. A common complaint was there was no service during the holidays, weekends, or when their usual support services person was sick. In general the information about services was limited, and the older people thought that they were just given what the agency thought that they needed.

3.19.6 Decisions theme

This theme describes who made the decisions about the permanent residence of the older person after an assessment for support needs. It encompassed the opinions of the older person, caregivers and the NASC Table 3-45 demonstrates that the majority (76%) of the older people felt that they had the power to decide to either remain at home, or enter residential care. A lesser number of older people thought that their family was involved in that decision (32%), however the reverse was seen in the caregivers report (Table 3-46) and NASC report (Table 3). Interestingly both the NASC and caregivers gave lower numbers for the older person’s influence. Overall the majority of the decision makers were

OPERA FINDINGS

PAGE 118

the older person and their family; however the hospital doctor and the general practitioner were the next most commonly noted decision makers.

Table 3-45: Older people’s report on decision makers

Older Family Hospital General Hospital Social Friends District Practice Physio NASC Occupation-Strength of person doctor practitioner nurse worker nurse nurse therapist al therapist

Influence n=131 Number of decision makers Very strong 100 42 12 7 4 2 3 1 2 1 1 1strong 13 20 7 4 4 1 1 1 1 0 1 1some 8 8 7 5 4 3 0 2 0 1 0 0minimal 5 2 0 0 0 0 2 0 0 1 0 0none 5 59 105 115 119 125 125 127 128 128 129 129

Table 3-46: Caregivers’ report on decision makers Older Family Hospital General Hospital Social Friends District Practice Physio NASC Occupation-

Strength of person doctor practitioner nurse worker nurse nurse therapist al therapistInfluence n=24 Number of decision makers

Very strong 8 18 2 0 0 0 0 0 0 0 0 0strong 4 4 1 0 0 0 0 0 0 0 0 0some 8 0 2 4 1 0 0 1 0 0 0 0minimal 2 1 0 0 0 0 0 0 0 0 0 0none 2 1 19 20 23 24 24 23 24 24 24 24

Table 3-47: NASC report on major decision makers Older Family Hospital General Hospital Social Friends District Practice Physio NASC Occupation-

Strength of person doctor practitioner nurse worker nurse nurse therapist al therapistInfluence n=12 Number of decision makers

Very strong 2 10 3 3 1 2 0 0 0 0 0 0strong 2 0 2 0 1 2 0 0 0 0 0 0some 2 0 1 2 0 1 0 0 0 0 0 0minimal 5 0 0 0 0 0 0 0 0 0 0 0none 1 2 6 7 10 9 12 12 12 12 12 12

The older people often thought that the hospital doctors were prescriptive, giving them no choice as mentioned by one person:

“I had a spell in hospital and after that they said that I couldn’t go home to my wife, I was too much trouble and too heavy, so I had to come here [rest home]” (Older person).

In other cases it was the family who wouldn’t give the older person the choice:

“I would like to go into a rest home, but my husband won’t let me. I would get more attention there. He goes out too much and I never see him” (Older person).

OPERA FINDINGS

PAGE 119

and a comment from the N.A.S.C.

“The family decided that it was too dangerous for her to stay in her own house and they couldn’t have her in theirs, so she was put in a rest home. It was the family’s decision. Mrs K wanted to go back home” (NASC).

Other people went into residential care for respite care, and just continued on there after the respite had ended. In this case it was usually the older person who found that they felt more secure, so made the decision to stay in residential care. One man described his respite care experience as “being treated like royalty”. The majority of older people (90%) who were living at home felt they had a major influence on the decision about where they were going to live, as seen in Table 3-48. However, for those older people who entered residential care, their range of influence was much lower (53%) as seen in Table 5. The percentage of family having a major influence was similar in both situations (46%). Few older people (12%) thought that the doctors influenced the decision to returning home, while 40% of the older people felt that the hospital doctors had a major influence on the decision to enter residential care. The percentage of influence which the general practitioner had for those people living at home was doubled when the older people entered residential care (13%).

Table 3-48: Percentage of decision makers for the older person to return home

Older people Family General Hospital Doctorsn=123 Practitioner doctor combined

Major influence (5,4) 90.24 46.34 6.50 12.20 9.35Some influence (3,2) 7.32 7.32 3.25 5.69 4.47No influence (1) 2.44 2.44 90.24 82.11 86.18

Table 3-49: Percentage of decision makers for the older person to enter residential care

Older people Family General Hospital Doctorsn=15 Practitioner doctor combined

Major influence (5,4) 53.33 46.67 13.33 40 26.67Some influence (3,2) 13.33 20.00 13.33 0 6.67No influence (1) 33.33 33.33 73.33 60 66.67

OPERA FINDINGS

PAGE 120

3.19.7 Residence

This described where the older person was permanently living. The term ‘home’ within this study includes retirement villages and living with relatives, but excluding rest homes, and continuing care hospitals (residential care). The term home conjured up many different pictures to people, for example, a garden, friends, sitting in peace having meals, independence, family and family pets. At least 60% of the people had lived in their homes for more than ten years, with some remaining in the one home for over 50 years. There was a common feeling of wanting to remain there.

“We have been here 50 years and I don’t want to move. My wife finds it a bit hard, but I want to stay here” (Older person).

Most people related the time they had lived in their home as the reason they wanted to stay there. People were living with concerns such as poor accessibility for ambulances, steps which could not be negotiated, and living on the hill. Transport and accessibility to shops proved a problem with poor taxi services, and no close bus service.

“I lost my license and that has really put a dampener on my activities. The taxi service is really bad. It doesn’t run on Sundays at all up here, we have to get a taxi from Lower Hutt all the way up here and they take an hour or so to get here….Sometimes they don’t show up at all. I get frustrated” (Older person).

Relocation was something that many older people did when they felt that their family home or the garden had become too big, or they were too far from family. Retirement villages were a good option for some who felt that they were closer to services and “like minded people.” However one person said that the reason he wouldn’t go in to a retirement village was because of the bad contracts that they offered.

“I moved to this house 20 years ago. My previous house was too big, well it wasn’t the house really… it was the garden” (Older person).

Many older people who were admitted to residential care had previously been in an acute hospital, and may also have been in respite care after discharge from hospital. The following reports from the N.A.S.C. were not uncommon:

OPERA FINDINGS

PAGE 121

“The old lady had a fall and went into hospital. At discharge she went into respite care and from there just stayed in the rest home” (NASC).

Also people thought that to go into residential care was inevitable if you lived long enough:

“Oh yes that [going into the rest home] was the wisest thing to do really wasn’t it. It is not my cup of tea this sort of life, but I mean it comes, well if you live long enough it comes to us all possibly” (Older person)

Happiness about the older people’s residence decision was diverse, as seen in Table 3-50 with 25% older people being happy in residential care and 75% being happy at home.

Table 3-50: Older people’s feelings about their residential placement decision

Feeling about the decision People living in residential care (n =19)

People living at home (n =71)

Very happy 4 21 Happy 1 32

No strong feeling 5 11 Sad 7 4

Very sad 2 3

The figure below interprets the data from the table above, through collapsing the categories (Very happy and Happy to one and Sad and Very Sad to one).

OPERA FINDINGS

PAGE 122

26 26

47

75

1510

0

10

20

30

40

50

60

70

80

Very happy or happy No strong feeling Sad or very sad

Satisfaction with place of residence

Perc

enta

ge o

f tot

al

People living in residential care People living at home

Figure 3-10: Older people's feelings about their decision around their residential placement decision

One older person said that she really enjoyed living at the rest home with the friendly atmosphere, and the fact that there were lots of people to talk to. Others were sad, particularly if they had left a spouse at home. The change of environment and habits was hard for some as indicated below:

“Yes I miss that, [home] yes you do really, you can more or less have what you want at home, it is not the same at all really. But they are very good here [rest home] and you certainly don’t go hungry, but it is not quite the same. It is a bit of a loss. I do go to bed quite early here because there is nothing else to do here. I get up at 7am to have a shower, it is very early. I never did that at home” (Older person)

3.20 Summary

The qualitative findings commenced with a literature review (Phase 1) which identified areas of risk of entering residential care, and the people who made the residence decisions. Several major risk factors for residential care entry were highlighted such as, being over 80 years of age, living alone, carer not coping, being ‘white’, difficulty with daily functions and cognitive dysfunction. A sequential model was developed from the four

OPERA FINDINGS

PAGE 123

themes for Phase 3 which were: coping, support, decisions and residence themes. These themes described the stages and processes which the older person, and where possible their carer went through, as well as thoughts from the NASC. The residence theme described the older person’s satisfaction with the residence decision. The vast majority of older people felt that they had a major influence in their residence decision, with the family being the next most common decision maker. If the older person lived at home 75% said they were happy, while if they lived in residential care 25% said they were happy. Overall the older people were happier at home, and felt that they had considerable input into the decision of whether they lived at home or residential care.

KEY FINDINGS

PAGE 124

KEY FINDINGS

PAGE 125

Chapter IV: key findings

4.1 Introduction

ASPIRE has aimed to assess the effectiveness of the Promoting Independence Programme (PIP) in Lower Hutt, Community FIRST (Flexible Integrated Restorative Support Team) in Hamilton and COSE (Coordination of Services for Elderly) in Christchurch. All are examples of ageing-in-place initiatives (AIPI), which specifically target older people with high and very high needs and support them to remain living in the community. The ASPIRE trial has sought to assess the impact of the AIPI on older people in preventing or delaying entry into permanent residential care as well as their role in reducing mortality. In addition, this project has aimed to evaluate the cost-effectiveness of the three initiatives and the results are available in ASPIRE (Report II). This chapter will present the key findings with some level of explanation surrounding each. The key findings are not presented in any order of priority.

The current Support Needs Level Assessment and categorisation system used by the NASC services to determine allocation of funding appeared to be highly variable across the three District Health Boards under investigation. It appeared that older people in Christchurch were assessed as being able to enter Residential care with a lower level of disability than those living in Hamilton and Lower Hutt. The interRAI MDS-HC assessment tool used by the research team provided a more rigorous and standardised method of assessment. This variation is probably a factor in variations between the key outcomes achieved in different services, such as reducing mortality or admission to residential care.

All three services appeared to reduce the risk of mortality compared with usual services. This varied from 28% in Community FIRST, 14% in PIP and 10% in COSE in comparison to older people in usual care. Although these figures are not statistically significant they reflect a clear trend in each service and are consistent with other results in the trial.

All three AIPI reduced the risk of entry to residential facilities when compared to usual care. Specifically, COSE reduced the risk of entry of older people to residential care in

KEY FINDINGS

PAGE 126

comparison with the usual care NASC services by 43% (reduction). Community FIRST appeared to result in a reduction of risk of entry to residential care by 33%, in comparison to usual care, though given the lower sample size this was not statistically significant. The Masonic Promoting Independence Programme appeared to reduce risk of entry to residential care by 16% in comparison to usual care and as with Community FIRST; given the lower sample size this was not statistically significant.

Caregiver stress levels did not appear to increase in the intervention groups in comparison to usual care, despite the higher number of older people with high and complex needs remaining living at home.

An improvement in the independence levels of older people (Activities of Daily Living) within Community FIRST was noted, in comparison to usual care. No improvement in function was noted in the COSE or PIP initiatives in comparison to usual care.

Predictive modelling of the likelihood of older people being hospitalised or entering residential care was carried out using all the older people in the sample. This produced interesting results consistent with much overseas research.

If a functional decline occurs in older people and the deterioration is not stopped, the older person is 11 times more likely to enter residential care.

An older person is almost twice as likely to enter residential care if they are socially isolated.

If an older person reports as having a negative mood, they are over twice as likely to be admitted to residential care.

For every one unit increase on the Caregiver Reaction Assessment (which measures caregiver stress), there is a 7% increased risk of residential care entry.

When an older person experiences inadequate meals and dehydration, they are over twice and 1.7 times more likely to be admitted to residential care, respectively.

Delirium is highly correlated with risk of admission to residential care; those older people with delirium are 3.6 times more likely to be institutionalised.

KEY FINDINGS

PAGE 127

A lack of medication review (almost twice as likely), negative mood (1.5 times more likely) and previous hospitalisation (1.8 times more likely) are all correlated with increased risk of hospitalisation.

The findings show there are a number of factors highly important to the enrolled population of older people such as coping, support, decision making and place of residence.

The study also explored the process by which older people entered residential care. Whilst the majority of older people often felt they had made the decision (to enter residential care) in most cases both the family and the NACS services thought that the family had been the main decision-makers. Further, nearly half of those who had entered residential care were sad or very sad about the decision. By contrast three quarters of those living in their own homes were happy or very happy with their decision to remain living at home.

4.2 Conclusions

Although the findings presented here can be viewed as is, it is important to note that interpretation must be undertaken in light of the cost-effectiveness of the relative ageing-in-place initiatives, which is available in ASPIRE (Report II). Also of note are the very different approaches each initiative took to facilitate ageing-in-place. Where as there are clear benefits in exploring multiple means to support older people to age-in-place, there is a tendency to compare the AIPI evaluated here and the relative success each achieved. In actuality, the strength of ASPIRE is to isolate those factors that are effective in facilitating ageing-in-place to allow new and existing services to evolve and develop.

Perhaps the most important finding of ASPIRE is that older people with high and complex needs who may enter permanent residential care are choosing on the whole to remain living at home with no apparent increase in risks to themselves.

STUDY LIMITATIONS

PAGE 128

Chapter V: Study Limitations

5.1 Introduction

Health service research is invariably challenging as it is not possible or indeed useful to create an artificial experimental environment. In order to generalise from the study, interventions need to be based in the real world and face every day common issues, such as recruitment and retention issues as well as referrals. The fact that ASPIRE was evaluating three services that were not developed or controlled by the research team provides considerable strengths but at the same time many limitations. This final section presents these limitations which must be considered when interpreting the results of this study.

5.2 Recruitment

The main study limitation is without doubt the low sample size in Hutt and in particular Hamilton. At the study onset, it was intended that recruitment would last 12 months with 12 months of follow-up. There were few issues in referrals, refusals or recruitment in Christchurch, but by 12 months, only approximately 100 were recruited to Hutt and less than 55 in Hamilton. At 12 months, a decision was made by the research team in collaboration with Waikato DHB and the MoH to continue recruitment in Hamilton for a further six months in an effort to increase the sample size. This naturally presents further problems as although the sample size increased past 100, almost 50% of the sample had follow-up data less than one year.

These difficulties in recruiting participants to ASPIRE in Hamilton presented particular challenges to the research team. Although the rate at which older people refused participation was not any higher than the other sites, the referrals from NASC were particularly low and required considerable intervention from both the DHB and the research team; as reflected by the considerably lower than anticipated sample size. On one level, this is perfectly understandable; Community FIRST had only been operating for a few months prior to commencement of the trial and therefore it would be anticipated that NASC

STUDY LIMITATIONS

PAGE 129

may have had concerns around the skills and abilities of staff within the new Community FIRST initiative. An important lesson from this trial for other DHBs considering adopting a similar approach would be to follow a mutually agreed implementation pathway.

The Masonic Promoting Independence Programme (PIP) had similar though not quite so dramatic issues in recruiting participants. As with Community FIRST, although the Masonic PIP had been running for some time in Levin, it was relatively new in the Hutt Valley and the service specifications required re-working prior and during implementation.

5.3 Design issues

When interpreting the relative effectiveness of all three services, it is important to consider the limitations of ASPIRE itself. ASPIRE provides a snapshot of how service are performing during a set time period and therefore if services are operating sub-optimally then the evaluation reflects that. All three services were undergoing a period of transition. COSE was beginning a roll-out from a very small pilot to almost half of Christchurch, Community FIRST had only been running for a few months prior to ASPIRE and were having issues in staff recruitment. The Masonic PIP experienced a change in service specification from slow stream rehab to PIP and further during that period experienced considerable staff turnover. Further, there was a loss of short term stay rehabilitation beds at the Masonic village with the preference to use other facility beds when required.

5.4 Selection bias

Selection bias occurs when an investigator attributes study results to the effect of the independent variable when in reality the results could be explained by differences in the subjects before the experimental intervention was implemented. Several issues arose over the course of the investigation that may have confounded data in this manner. The research team had little control over the referrals to the programme and therefore the sample may not be representative as health professionals may have decided that a number of potential participants should be relocated directly to residential care. Although there were many efforts to minimise this phenomena, there is a possibility that older people were pre-screened by the Multi-disciplinary team.

STUDY LIMITATIONS

PAGE 130

Randomisation facilitates an equal distribution of known and unknown confounders across groups and although statistical testing revealed that the AIPI and usual care groups were similar, there were some differences observed at baseline in the two groups such as memory loss which may have influenced the results over time.

5.5 Maturation

Changes that take place in subjects because of development or the passage of time rather than as a response to an intervention, can also pose a threat to internal validity. For example, analysis around independence levels (assessed by ADL) in Hamilton reveal a mean decline in the usual care group that may have accentuated the associated improvement in the LAY group. An inherent difficulty in researching older people is that there is an expectation that a number of participants will decline in function and health over the period of a lengthy study. Thus, the declines observed in the usual care group may be part of a normal age-related deterioration.

5.6 Testing

Assessments were undertaken multiple times on participants involved in the study. Participants may have remembered their responses to questions at previous assessment points are altered their answers as a consequence of being exposed to them several times.

A further difficulty arose in that, although all attempts were made to ensure that follow-up assessments were undertaken on the same time and day, on occasions the subject cancelled or re-scheduled appointments. The extent to which this occurred was minor, however, in a few cases, it was not possible to assess the subject at the same time and day and as a result, there is considerable potential for recording errors. To illustrate, ‘morning stiffness’ is a measurable phenomenon that seems to be associated with strength and other functional abilities, which appear to improve over the course of the day. Thus, although the number of occasions where this issue arose was few, there is always the possibility that data could be confounded in some manner.

Although the MDS-HC proved to be an invaluable means to draw comparisons regarding the independence and relative frailty of older people enrolled in each region, it also

STUDY LIMITATIONS

PAGE 131

appeared to have several key limitations. One of which was the perceived inability of the tool to detect change over time, particularly in relation to the ADL and IADL scale.

It is interesting that none of the services appeared to demonstrate a statistically significant change in IADL function, despite a statistically significant change in ADL function within Community FIRST. Therefore, there is a distinct possibility that this lack of change could be an artefact of the relatively poor scope of the IADL scale. When comparing the MDS-HC IADL scale with other scales assessing extended function, it omits important areas such as walking outside and crossing roads as well as a host of other items. Clinically, one would anticipate demonstrating a shift in IADL items before ADL items and therefore the lack of change observed here may be a product of the assessment tool, which appears to have been addressed in Version 3.0 release of MDS-HC.

Measuring quality of life is difficult and the EuroQoL 5D was utilised in this study more for from a cost utility perspective. Where as the full EuroQoL results are reported in ASPIRE (Report II), the visual analogue scale was presented in this report. There were no real clinically relevant trends observed across any of the sites in ASPIRE from either the older person’s perspective on their quality of life or the caregivers. This is surprising, as one would anticipate observing improvements, if independence levels increased as observed in Hamilton. However, this could well be as a result of the considerably difficulty in assessing quality of life.

5.7 Hawthorne effect

The Hawthorne effect occurs when participants respond in a specific manner because they are aware that they are participants in a research project. This may have influenced the findings in some way, though it is impossible to determine to what extent.

5.8 Experimenter effects

Experimenter effects stem from characteristics of the researcher. Attributes such as age, gender and specifically profession and facial expressions may influence subject’s behaviour and responses. To illustrate, when the subjects were being assessed they may have behaved differently because of their familiarity with the researcher. Indeed, several

STUDY LIMITATIONS

PAGE 132

researchers have pointed to subjects becoming so well acquainted with them that they exerted themselves in the assessments more than they would normally do so, in order to achieve better results. This issue is of particular relevance to the present study as many of the subjects had few weekly visits and would therefore look forward to the researcher attending for the assessment. Although these threats are minimised by the usual care group, it remains an inherent weakness.

APPENDICES

PAGE 133

APPENDICES

PAGE 134

DEFF = 1 + (m-1)* ICC

where m is the cluster size and ICC is Intra-cluster-correlation

Appendices

Appendix 1: Statistical analysis Additional sample size and cluster size calculations

The formula for Design Effect is:

As we can see from the formula above if the cluster size is large, even a small value of ICC can have a large effect on sample size. The sample size required for a cluster-randomised study is the sample size for individual randomised study times DEFF. Below are various scenarios of the DEFF in the Christchurch trial assuming 35% event rate in the control group.

Design Effect Sample Size required in Christchurch Total sample size required

1 277 832

1.5 416 971

2 555 1109

2.5 693 1248

3 832 1387

For the purpose of the setting the total sample size to allow 90% Power when all 3 centres are combined, we have assumed that the DEFF is equal to 2 (cluster size=11, ICC=0.1). This gives us the anticipated total sample size required, from the scenarios presented below, of 1109 patients.

APPENDICES

PAGE 135

Handling of Christchurch data

Since one of the study centres was a cluster-randomised trial, it needed to be handled differently from the other two individually randomised trials. Since the data within a cluster is correlated there was a need to consider clustering effect when conducting the analysis of the Christchurch data. Therefore, the standard analysis methods were not suitable for these data. This section discusses the analysis methods that account for the clustering effect.

Proc Phreg in SAS was used to handle dependent data by addressing the COVB (AGGREGATE) option in the procedure (Glidden, 2004; Aalen, 1995). Hence, the Hazard ratio adjusted with cluster effect will be presented. The next step was to utilise meta-analysis to pool the entire three hazard ratios together. Standard Meta-analysis does not allow for clustering effect. There are several statistical methods for the meta-analysis for cluster randomized trials, such as adjusted Mantel-Haenszel test, Ratio estimator procedure, woolf procedure, Generalized estimating equation approach, etc. However, in this instance, only the adjusted Mantel-Haenszel test was considered (Donner, 2001; Donner and Klar, 2002). Adjusted Mantel-Haenszel test is represented by the formula:

2

1221

121

21221112

1221

2

)1/()(

])/[)]()([(

jjjj

s

jjjjjjj

jjjjjjj

s

jjjj

CMH

MCMCMAMAMM

CMCMAMAAMA

−+−

+−−−=

∑

∑

=

=χ

where ∑=

=ijn

lijlij mM

1 is the number of subjects in intervention group i in trial j amd ijA is defined as the number of successes in intervention group i in trial j.

ijijl

n

lijlij MmmC

ij

/]ˆ)1(1[1

ρ−+= ∑= , which are the clustering correction factors (estimated

design effect), where ρ̂ is the estimated intra-cluster correlation coefficient. Then the ‘clustered Mantel-Haenszel estimator’ (Donner et al, 2001; Donner and Klar, 2002) is given by:

∑∑==

Ψ=Ψs

jjCj

s

jjCCMH WW

11

ˆˆ, where

jjj

j

j

jjC PP

MC

MC

W 211

2

2

1

1 ˆ)ˆ1(][ −+= −

APPENDICES

PAGE 136

[To conduct analysis for secondary endpoints, a decision was made to ignore the clustering effect, when ICC is less than 0.05].

APPENDICES

PAGE 137

Appendix 2: The interRAI Minimum Data Set – Home Care (MDS-HC) The MDS-HC has several inherent scales. These are:

Activities of Daily Living Scales

The MDS items have been combined to form two types of summary measures (Morris et al. 1999):

1. a single MDS ADL Self-Performance Hierarchy Scale and;

2. two versions of additive ADL scales based on the same item pool (MDS ADL Short Form and the MDS ADL-Long Form Scales).

ADL Self-Performance Hierarchy Scale

The ADL Hierarchy Scale is a measure of ADL performance and categorises ADLs according to stages at which they can no longer be performed. The aim of this scale is to reflect the disablement process rather than to simply sum reduction in function. The scale is based upon the four ADL items used in the ADL short form (i.e., personal hygiene, toilet use, locomotion, eating). Early loss ADLs (e.g., personal hygiene) are given lower scores than those lost at a later stage (e.g., eating). For each of these four items, potential difficulty is scored from 0 (independence) to 4 (total dependence). A 6-point hierarchical scale is created as follows:

APPENDICES

PAGE 138

Table 5-51: ADL self-performance hierarchy scale

Score Description Use of four ADL items 0 Independent All four score 0 1 Supervision required All four score 1 or less AND at least one scores

1 2 Limited impairment All four score 2 or less AND at least one scores

2 3 Extensive assistance

required (I) Eating and locomotion both score less than 3 AND personal hygiene and toilet use both score 3 or greater

4 Extensive assistance required (II)

Eating OR Locomotion score 3

5 Dependent Eating OR Locomotion score 4 6 Total dependence All four score 4

These seven categories, which describe the older person’s ADL self-performance, can be grouped into four impairment levels: relatively independent (score = 0 or 1), limited impairment (score = 2), extensive help (score = 3 or 4), and more severely or totally dependent (score 5 or 6). Morris et al. (1999) reported that the ADL hierarchy scale provides evaluations of ADLs similar to other established ADL measures and is also able to reliably assess change in ADL impairment levels over time.

APPENDICES

PAGE 139

ADL Short Form Scale

This scale provides a summary measure of the client’s ability to perform ADLs and is based on four items: Personal hygiene, toilet use, locomotion and eating. The scale has a range of 0 to 16, with higher values indicating greater difficulty in performing activities.

ADL Long Form Scale

The ADL Long form is a summary scale capturing all 7 of the ADL items. Each item is scored 0 to 4, creating a scale with a range from 0 to 28.

Figure 11: ADL Hierarchy Scale (Morris et al, 1999)

APPENDICES

PAGE 140

The Cognitive Performance Scale (CPS) The cognitive performance scale (Morris, Fries et al. 1994) is a hierarchical index used to rate the cognitive status of older people. It has been validated against the Mini Mental State Examination, the Test for Severe Impairment (Morris, Fries et al. 1994; Hartmaier, Sloane et al. 1995) and includes four items: Short term memory, cognitive skills for daily decision making, expressive communication and eating. Based on the individual’s impairment level on these four items, a CPS score scale ranging from 0-6 is derived (equivalent scores on the Mini Mental State Examination (MMSE) are shown below).

Table 5-52: CPS rating scale

CPS score Description Equivalent Average

MMSE 0 Intact 25 1 Borderline intact 22 2 Mild impairment 19 3 Moderate impairment 15 4 Moderate/severe impairment 7 5 Severe impairment 5 6 Very severe impairment 1

Figure 5-13 highlights the scoring process in more detail.

APPENDICES

PAGE 141

Depression Rating Scale (DRS) (Burrows et al, 2000)

The Depression Scale can be used as a clinical screen for depression. It is based on 7 items embedded within the MDS-HC: Negative statements; persistent anger; expressions of unrealistic fears; repetitive health complaints; repetitive anxious complaints; sad, pained worried facial expressions and; tearfulness. Each of these 7 items is coded according to symptom frequency, resulting in a possible DRS score of 0 to 14. Burrows et al. (2000) demonstrated that scale scores of three or greater indicated major and minor depressive disorders in a nursing home population. They also established the criterion validity of this

Figure 5-12: The Cognitive Performance Scale

APPENDICES

PAGE 142

scale through comparisons with both the Hamilton Depression Rating Scale and the Cornell Scale for Depression.

Table 5-53: The Depression Rating Scale (DRS)

Score Indicators of depression

0, 1 or 2 Made negative statements

0, 1 or 2 Persistent anger with self or others 0, 1 or 2 Expressions (including non-verbal) of what appear to be unrealistic fears

0, 1 or 2 Repetitive health complaints

0, 1 or 2 Repetitive anxious complaints/concerns (non-health related)

0, 1 or 2 Sad, pained, worried facial expressions 0, 1 or 2 Crying, tearfulness

0 = Indicator not exhibited in last 30 days 1 = Indicator of this type exhibited up to five days a week 2 = Indicator of this type exhibited daily or almost daily (6, 7 days a week)

Instrumental activities of daily living (IADL) scales IADL difficulty scale: The IADL Difficulty Scale is a hierarchical index that measures difficulty with three IADLs: ordinary housework, preparing meals and using the telephone. Client scores are combined to create a scale ranging from 0 to 6, such that higher scores reflect greater difficulty on these IADLs as follows:

Table 5-54: IADL difficulty scale

IADL Difficulty Score Description 0 No difficulty on any of three IADLs 1-3 Some difficulty in 1 or more areas 4-6 Great difficulty in 1 or more areas

APPENDICES

PAGE 143

IADL involvement Scale

This scale is based upon a sum of 3 items in section H1 of the MDS-HC: ordinary housework, meal preparation and phone use. Individual items are summed to produce a scale that ranges from 0 to 9 (higher scores indicate greater difficulty in performing instrumental activities).

Table 5-55: IADL Involvement Scale

Score Instrumental Activities of Daily Living

0, 1, 2 or 3 Meal preparation

0, 1, 2 or 3 Ordinary housework

0, 1, 2 or 3 Managing finance

0, 1, 2 or 3 Managing medications 0, 1, 2 or 3 Phone use

0, 1, 2 or 3 Shopping

0, 1, 2 or 3 Transportation

Scoring in self-performance:

0 = INDEPENDENT – did on own 1 = SOME HELP – help some of the time 2 = FULL HELP – performed with help all of the time 3 = BY OTHERS – performed by others

Changes in Health, End-stage disease and Signs and Symptoms (CHESS)

The CHESS scale was developed to detect frailty and instability in health. The CHESS attempts to identify individuals at risk of serious decline and can serve as an outcome where the objective is to minimize problems related to frailty (e.g., declines in function) in the elderly population. The CHESS scale was originally developed for use with the MDS 2.0 and has since been adapted for use with the MDS-HC. The CHESS first creates a subscale by counting across the following health symptoms:

APPENDICES

PAGE 144

Table 5-56: The CHESS scale

Assessment item MDS-HC item code Vomiting K2e Dehydration L2c Leaving food uneaten L2b Weight loss L1a Shortness of breath K3e Oedema K3d

This score takes on values of either 0 (no symptoms), 1 (at least one symptom present) or 2 (2 or more symptoms present). This subscale is then added together with a score of 1 for each of the items on end stage disease (K8e), decline in cognition (B2b) and decline in ADL (H3) to result in a 6 point scale with scores ranging between 0 (meaning no instability) to 5 (for the highest level of instability). In the long-term care population, there is a clear differentiation of all six levels of CHESS scores, and higher levels are associated with a reduction in survival over time (Hirdes et al. 2003).

Pain Scale for the Minimum Data Set

The Pain Scale for the MDS was initially developed for use in nursing homes and later translated for use with the MDS-HC. The scale uses two items on the MDS (K4a and K4b) to create a score that ranges from 0 to 3 whereby 0 = no pain, 1 = less than daily pain, 2 = daily pain but not severe and 3 = severe daily pain. The Pain Scale has been shown to be highly predictive of pain on the Visual Analogue Scale in nursing home residents in the US (Fries, Simon et al. 2001).

APPENDICES

PAGE 145

Figure 13: The Pain Scale

APPENDICES

PAGE 146

Appendix 3: Survival plots The following figures are survival plots for firstly residential home placement and secondly mortality. The graphs provide a useful method of identifying the events (i.e. either permanent residential home placement or death) in the enrolled sample, as each drop in the plot indicates an event. However, events are relative to the enrolled population at that point and therefore the actual plots can be potentially misleading in small sample sizes as seen in Hamilton. For instance, if one older person is admitted to a residential facility at 18 months in Hamilton, it appears on the plot as being more of a decline than when one older person is admitted to a residential facility in Christchurch, as the enrolled sample in Christchurch at 18 months is larger and therefore one event may only be represented by a 5% drop (in a sample of 20), whereas in Hamilton it may be represented by a 20% drop (in a sample of 5). Notwithstanding this, the plots are useful up until 6 – 12 months into the study.

APPENDICES

PAGE 147

0. 00

0. 25

0. 50

0. 75

1. 00

survdeat h

0 100 200 300 400 500 600 700 800

STRATA: TREATMENT_GROUP_I D=17 Censored TREATMENT_GROUP_I D=17TREATMENT_GROUP_I D=18 Censored TREATMENT_GROUP_I D=18

Figure 5-14: Kaplan-Meier curves between treatment groups in Hamilton region (using Death as Primary endpoint)

0. 00

0. 25

0. 50

0. 75

1. 00

survdeat h

0 100 200 300 400 500 600 700 800

STRATA: TREATMENT_GROUP_I D=17 Censored TREATMENT_GROUP_I D=17TREATMENT_GROUP_I D=18 Censored TREATMENT_GROUP_I D=18

Figure 5-15: Kaplan-Meier curves between treatment groups in Lower Hutt region (using Death as Primary endpoint)

________ = AIPI ________ = Usual care

________ = AIPI ________ = Usual care

APPENDICES

PAGE 148

0. 00

0. 25

0. 50

0. 75

1. 00

survdeat h

0 100 200 300 400 500 600 700 800

STRATA: TREATMENT_GROUP_I D=17 Censored TREATMENT_GROUP_I D=17TREATMENT_GROUP_I D=18 Censored TREATMENT_GROUP_I D=18

Figure 5-16: Kaplan-Meier curves between treatment groups in Christchurch region (using Death as Primary endpoint)

0. 00

0. 25

0. 50

0. 75

1. 00

survresi d

0 100 200 300 400 500 600 700 800

STRATA: TREATMENT_GROUP_I D=17 Censored TREATMENT_GROUP_I D=17TREATMENT_GROUP_I D=18 Censored TREATMENT_GROUP_I D=18

Figure 5-17: Kaplan-Meier curves between treatment groups in Hamilton region (using Residential Care entry as Primary endpoint)

________ = AIPI ________ = Usual care

________ = AIPI ________ = Usual care

APPENDICES

PAGE 149

0. 00

0. 25

0. 50

0. 75

1. 00

survresi d

0 100 200 300 400 500 600 700 800

STRATA: TREATMENT_GROUP_I D=17 Censored TREATMENT_GROUP_I D=17TREATMENT_GROUP_I D=18 Censored TREATMENT_GROUP_I D=18

Figure 5-18: Kaplan-Meier curves between treatment groups in Lower Hutt region (using Residential Care entry as Primary endpoint)

0. 00

0. 25

0. 50

0. 75

1. 00

survresi d

0 100 200 300 400 500 600 700 800

STRATA: TREATMENT_GROUP_I D=17 Censored TREATMENT_GROUP_I D=17TREATMENT_GROUP_I D=18 Censored TREATMENT_GROUP_I D=18

Figure 5-19: Kaplan-Meier curves between treatment groups in Christchurch region (using Residential Care entry as Primary endpoint)

________ = AIPI ________ = Usual care

________ = AIPI ________ = Usual care

APPENDICES

PAGE 150

0. 00

0. 25

0. 50

0. 75

1. 00

survpr i m

0 100 200 300 400 500 600 700 800

STRATA: TREATMENT_GROUP_I D=17 Censored TREATMENT_GROUP_I D=17TREATMENT_GROUP_I D=18 Censored TREATMENT_GROUP_I D=18

Figure 5-20: Kaplan-Meier curves between treatment groups in Hamilton region (using combined primary outcome as Primary endpoint)

0. 00

0. 25

0. 50

0. 75

1. 00

survpr i m

0 100 200 300 400 500 600 700 800

STRATA: TREATMENT_GROUP_I D=17 Censored TREATMENT_GROUP_I D=17TREATMENT_GROUP_I D=18 Censored TREATMENT_GROUP_I D=18

Figure 5-21: Kaplan-Meier curves between treatment groups in Lower Hutt region (using combined primary outcome as Primary endpoint)

________ = AIPI ________ = Usual care

________ = AIPI ________ = Usual care

APPENDICES

PAGE 151

0. 00

0. 25

0. 50

0. 75

1. 00

survpr i m

0 100 200 300 400 500 600 700 800

STRATA: TREATMENT_GROUP_I D=17 Censored TREATMENT_GROUP_I D=17TREATMENT_GROUP_I D=18 Censored TREATMENT_GROUP_I D=18

Figure 5-22: Kaplan-Meier curves between treatment groups in Christchurch region (using combined primary outcome as Primary endpoint)

________ = AIPI ________ = Usual care

APPENDICES

PAGE 152

Appendix 4: Secondary outcome data (Excluding data arising from participants once entered residential home) The following 45 plots illustrate graphically changes in secondary outcome data over time. The variables measured are:

1. ADL Short Form scale

2. ADL Self-performance scale

3. ADL Long Form scale

4. IADL Difficulty Scale

5. IADL Involvement Scale

6. IADL Summary Scale

7. Cognitive Performance Scale (CPS)

8. Depression Rating Scale (DRS)

9. CHESS Scale

10. Pain Scale

11. Short Form – 36 (Physical)

12. Short Form – 36 (Mental)

13. Caregiver Reaction Assessment (CRA)

14. EuroQoL Visual Analogue Scale

15. EuroQoL Visual Analogue Scale (completed by caregiver for older person)

NOTE: These plots exclude data collected after an older person has been admitted to residential care.

APPENDICES

PAGE 153

Figure 5-23: Repeated measures of ADL short form scale in Hamilton

APPENDICES

PAGE 154

Figure 5-24: Repeated measures of ADL short form scale in Lower Hutt

APPENDICES

PAGE 155

Figure 5-25: Repeated measures of ADL short form scale in Christchurch

APPENDICES

PAGE 156

Figure 5-26: Repeated measures of ADL self-performance scale in Hamilton

APPENDICES

PAGE 157

Figure 5-27: Repeated measures of ADL self-performance scale in Lower Hutt

APPENDICES

PAGE 158

Figure 5-28: Repeated measures of ADL self-performance scale in Christchurch

APPENDICES

PAGE 159

Figure 5-29: Repeated measures of ADL Long form scale in Hamilton

APPENDICES

PAGE 160

Figure 5-30: Repeated measures of ADL Long form scale in Lower Hutt

APPENDICES

PAGE 161

Figure 5-31: Repeated measures of ADL Long form scale in Christchurch

APPENDICES

PAGE 162

Figure 5-32: Repeated measures of IADL difficulty scale in Hamilton

APPENDICES

PAGE 163

Figure 5-33: Repeated measures of IADL difficulty scale in Lower Hutt

APPENDICES

PAGE 164

Figure 5-34: Repeated measures of IADL difficulty scale in Christchurch

APPENDICES

PAGE 165

Figure 5-35: Repeated measures of IADL Involvement scale in Hamilton

APPENDICES

PAGE 166

Figure 5-36: Repeated measures of IADL Involvement scale in Lower Hutt

APPENDICES

PAGE 167

Figure 5-37: Repeated measures of IADL Involvement scale in Christchurch

APPENDICES

PAGE 168

Figure 5-38: Repeated measure of IADL Summary scale in Hamilton

APPENDICES

PAGE 169

Figure 5-39: Repeated measures of IADL Summary scale in Lower Hutt

APPENDICES

PAGE 170

Figure 5-40: Repeated measures of IADL Summary scale in Christchurch

APPENDICES

PAGE 171

Figure 5-41: Repeated measures of CPS scale in Hamilton

APPENDICES

PAGE 172

Figure 5-42: Repeated measures of CPS scale in Lower Hutt

APPENDICES

PAGE 173

Figure 5-43: Repeated measures of CPS scale in Christchurch

APPENDICES

PAGE 174

Figure 5-44: Repeated measures of DRS scale in Hamilton

APPENDICES

PAGE 175

Figure 5-45: Repeated measures of DRS scale in Lower Hutt

APPENDICES

PAGE 176

Figure 5-46: Repeated measures of DRS scale in Christchurch

APPENDICES

PAGE 177

Figure 5-47: Repeated measures of CHESS scale in Hamilton

APPENDICES

PAGE 178

Figure 5-48: Repeated measures of CHESS scale in Lower Hutt

APPENDICES

PAGE 179

Figure 5-49: Repeated measures of CHESS scale in Christchurch

APPENDICES

PAGE 180

Figure 5-50: Repeated measures of Pain scale in Hamilton

APPENDICES

PAGE 181

Figure 5-51: Repeated measures of Pain scale in Lower Hutt

APPENDICES

PAGE 182

Figure 5-52: Repeated measures of Pain scale in Christchurch

APPENDICES

PAGE 183

Figure 5-53: Repeated measures of SF36 PCS in Hamilton

APPENDICES

PAGE 184

Figure 5-54: Repeated measures of SF36 PCS in Lower Hutt

APPENDICES

PAGE 185

Figure 5-55: Repeated measures of SF36 PCS in Christchurch

APPENDICES

PAGE 186

Figure 5-56: Repeated measures of SF36 MCS in Hamilton

APPENDICES

PAGE 187

Figure 5-57: Repeated measures of SF36 MCS in Lower Hutt

APPENDICES

PAGE 188

Figure 5-58: Repeated measures of SF36 MCS in Christchurch

APPENDICES

PAGE 189

Figure 5-59: Repeated measures of CRA in Hamilton

APPENDICES

PAGE 190

Figure 5-60: Repeated measures of CRA in Lower Hutt

APPENDICES

PAGE 191

Figure 5-61: Repeated measures of CRA in Christchurch

APPENDICES

PAGE 192

Figure 5-62: Repeated measures of EQ VAS of OP in Hamilton

APPENDICES

PAGE 193

Figure 5-63: Repeated measures of EQ VAS of OP in Lower Hutt

APPENDICES

PAGE 194

Figure 5-64: Repeated measures of EQ VAS of OP in Christchurch

APPENDICES

PAGE 195

Figure 5-65: Repeated measures of EQ VAS of OP from Caregiver in Hamilton

APPENDICES

PAGE 196

Figure 5-66: Repeated measures of EQ VAS of OP from Caregiver in Lower Hutt

APPENDICES

PAGE 197

Figure 5-67: Repeated measures of EQ VAS of OP from Caregiver in Christchurch

APPENDICES

PAGE 198

Appendix 5: Secondary outcome data (Including data arising from participants once entered residential home) The following 45 plots illustrate graphically changes in secondary outcome data over time. The variables measured are:

16. ADL Short Form scale

17. ADL Self-performance scale

18. ADL Long Form scale

19. IADL Difficulty Scale

20. IADL Involvement Scale

21. IADL Summary Scale

22. Cognitive Performance Scale (CPS)

23. Depression Rating Scale (DRS)

24. CHESS Scale

25. Pain Scale

26. Short Form – 36 (Physical)

27. Short Form – 36 (Mental)

28. Caregiver Reaction Assessment (CRA)

29. EuroQoL Visual Analogue Scale

30. EuroQoL Visual Analogue Scale (completed by caregiver for older person)

NOTE: These plots include data arising from all scales from all assessment points

APPENDICES

PAGE 199

Figure 5-68: Repeated measures of ADL short form scale in Hamilton

APPENDICES

PAGE 200

Figure 5-69: Repeated measures of ADL short form scale in Lower Hutt

APPENDICES

PAGE 201

Figure 5-70: Repeated measures of ADL short form scale in Christchurch

APPENDICES

PAGE 202

Figure 5-71: Repeated measures of ADL self-performance scale in Hamilton

APPENDICES

PAGE 203

Figure 5-72: Repeated measures of ADL self-performance scale in Lower Hutt

APPENDICES

PAGE 204

Figure 5-73: Repeated measures of ADL self-performance scale in Christchurch

APPENDICES

PAGE 205

Figure 5-74: Repeated measures of ADL Long form scale in Hamilton

APPENDICES

PAGE 206

Figure 5-75: Repeated measures of ADL Long form scale in Lower Hutt

APPENDICES

PAGE 207

Figure 5-76: Repeated measures of ADL Long form scale in Christchurch

APPENDICES

PAGE 208

Figure 5-77: Repeated measures of IADL difficulty scale in Hamilton

APPENDICES

PAGE 209

Figure 5-78: Repeated measures of IADL difficulty scale in Lower Hutt

APPENDICES

PAGE 210

Figure 5-79: Repeated measures of IADL difficulty scale in Christchurch

APPENDICES

PAGE 211

Figure 5-80: Repeated measures of IADL Involvement scale in Hamilton

APPENDICES

PAGE 212

Figure 5-81: Repeated measures of IADL Involvement scale in Lower Hutt

APPENDICES

PAGE 213

Figure 5-82: Repeated measures of IADL Involvement scale in Christchurch

APPENDICES

PAGE 214

Figure 5-83: Repeated measure of IADL Summary scale in Hamilton

APPENDICES

PAGE 215

Figure 5-84: Repeated measures of IADL Summary scale in Lower Hutt

APPENDICES

PAGE 216

Figure 5-85: Repeated measures of IADL Summary scale in Christchurch

APPENDICES

PAGE 217

Figure 5-86: Repeated measures of CPS scale in Hamilton

APPENDICES

PAGE 218

Figure 5-87: Repeated measures of CPS scale in Lower Hutt

APPENDICES

PAGE 219

Figure 5-88: Repeated measures of CPS scale in Christchurch

APPENDICES

PAGE 220

Figure 5-89: Repeated measures of DRS scale in Hamilton

APPENDICES

PAGE 221

Figure 5-90: Repeated measures of DRS scale in Lower Hutt

APPENDICES

PAGE 222

Figure 5-91: Repeated measures of DRS scale in Christchurch

APPENDICES

PAGE 223

Figure 5-92: Repeated measures of CHESS scale in Hamilton

APPENDICES

PAGE 224

Figure 5-93: Repeated measures of CHESS scale in Lower Hutt

APPENDICES

PAGE 225

Figure 5-94: Repeated measures of CHESS scale in Christchurch

APPENDICES

PAGE 226

Figure 5-95: Repeated measures of Pain scale in Hamilton

APPENDICES

PAGE 227

Figure 5-96: Repeated measures of Pain scale in Lower Hutt

APPENDICES

PAGE 228

Figure 5-97: Repeated measures of Pain scale in Christchurch

APPENDICES

PAGE 229

Figure 5-98: Repeated measures of SF36 PCS in Hamilton

APPENDICES

PAGE 230

Figure 5-99: Repeated measures of SF36 PCS in Lower Hutt

APPENDICES

PAGE 231

Figure 5-100: Repeated measures of SF36 PCS in Christchurch

APPENDICES

PAGE 232

Figure 5-101: Repeated measures of SF36 MCS in Hamilton

APPENDICES

PAGE 233

Figure 5-102: Repeated measures of SF36 MCS in Lower Hutt

APPENDICES

PAGE 234

Figure 5-103: Repeated measures of SF36 MCS in Christchurch

APPENDICES

PAGE 235

Figure 5-104: Repeated measures of Caregiver Reaction Assessment in Hamilton

APPENDICES

PAGE 236

Figure 5-105: Repeated measures of Caregiver Reaction Assessment in Lower Hutt

APPENDICES

PAGE 237

Figure 5-106: Repeated measures of Caregiver Reaction Assessment in Christchurch

APPENDICES

PAGE 238

Figure 5-107: Repeated measures of EuroQoL Visual Analogue Scale of Older Person in Hamilton

APPENDICES

PAGE 239

Figure 5-108: Repeated measures of EuroQoL Visual Analogue Scale of Older Person in Lower Hutt

APPENDICES

PAGE 240

Figure 5-109: Repeated measures of EuroQoL Visual Analogue Scale of Older Person in Christchurch

APPENDICES

PAGE 241

Figure 5-110: Repeated measures of EuroQoL Visual Analogue Scale of Older Person from Caregiver in Hamilton

APPENDICES

PAGE 242

Figure 5-111: Repeated measures of EuroQoL Visual Analogue Scale of Older Person from Caregiver in Lower Hutt

APPENDICES

PAGE 243

Figure 5-112: Repeated measures of EuroQoL Visual Analogue Scale of Older Person from Caregiver in Christchurch

APPENDICES

PAGE 244

Appendix 6: Adverse events Table 5-57: Adverse events count (including RH data) for Hamilton

Hamilton

Usual Care New Service

Description n of

events %1 n of

people %2 n of

events %1 n of

people %2

x2 test p-

value

Top 10 AEs

U.T.I. - - - - 2 100.00 2 100 0.2366

C.V.A 1 33.33 1 33.33 2 66.67 2 66.67 0.6147

COPD - - - - 1 100.00 1 100 0.4868

Heart Attack 2 100 2 100 - - - - 0.4995

Congestive Heart Attack 4 100 3 100 - - - - 0.2492

Heart Failure 6 60.00 6 60.00 4 40.00 4 40.00 0.7526

Pneumonia 1 25.00 1 25.00 3 75.00 3 75.00 0.3610

Delirium 13 86.67 8 80.00 2 13.33 2 20.00 0.1075

Parkinson's Disease 2 50.00 2 66.67 2 50.00 1 33.33 1.0000

Other 50 42.74 47 43.93 67 57.26 60 56.07

Total number of AEs 87 50.88 84 49.12

Falls (OP 9.23) 43 40.57 23 48.94 63 59.43 24 51.06 0.8503

Hospitalisation (any) 31 50.00 19 48.72 31 50.00 20 51.28 0.8645

Serious Falls 11 55.00 7 53.85 9 45.00 6 46.15 0.9234

Serious AEs - if admit to hospital or death 46 48.42 28 50.91 49 51.58 27 49.09 0.9378

APPENDICES

PAGE 245

Table 5-58: Adverse events count (including RH data) for Hutt

Lower Hutt

Usual Care New Service

Description n of

events %1 n of

people %2 n of

events %1 n of

people %2 x2 test

p-value

Top 10 AEs

U.T.I. 1 33.33 1 33.33 2 66.67 2 66.67 0.6147

C.V.A 4 40.00 4 40.00 6 60.00 6 60.00 0.5358

COPD - - - - - - - - -

Heart Attack - - - - 1 100 1 100 0.4868

Congestive Heart Attack 2 100 2 100 - - - - 0.4995

Heart Failure 1 50.00 1 50.00 1 50.00 1 50.00 1.0000

Pneumonia 4 57.14 2 50.00 3 42.86 2 50.00 1.0000

Delirium - - - - - - - - -

Parkinson's Disease 3 75.00 2 66.67 1 25.00 1 33.33 1.0000

Other 42 52.50 39 52.70 38 47.50 35 47.30

Total number of AEs 60 51.72 56 48.28

Falls (OP 9.23) 54 45.76 21 46.67 64 54.24 24 53.33 0.6205

Hospitalisation (any) 24 53.33 18 51.43 21 46.67 17 48.57 0.8721

Serious Falls 11 84.62 8 80.00 2 15.38 2 20.00 0.1075

Serious AEs - if admit to hospital or death 32 52.46 25 52.08 29 47.54 23 47.92 0.9681

APPENDICES

PAGE 246

Table 5-59: Adverse events count (including RH data) for Christchurch

Christchurch

Usual Care New Service

Description n of

events %1 n of

people %2 n of

events %1 n of

people %2

x2 test p-

value

Top 10 AEs

U.T.I. 13 72.22 8 61.54 5 27.78 5 38.46 0.6418

C.V.A 3 50.00 3 50.00 3 50.00 3 50.00 1.0000

COPD 14 77.78 8 80.00 4 22.22 2 20.00 0.1075

Heart Attack 4 26.67 4 28.57 11 73.33 10 71.43 0.1062

Congestive Heart Attack 9 81.82 7 77.78 2 18.18 2 22.22 0.1776

Heart Failure 2 40.00 2 40.00 3 60.00 3 60.00 0.6785

Pneumonia 4 66.67 4 66.67 2 33.33 2 33.33 0.6865

Delirium 1 100 1 100 - - - - 1.0000

Parkinson's Disease 6 85.71 2 66.67 1 14.29 1 33.33 1.0000

Other 217 48.87 201 51.15 227 51.13 192 48.85

Total number of AEs 275 51.31 261 48.69

Falls (OP 9.23) 175 53.52 77 52.74 152 46.48 69 47.26 0.7622

Hospitalisation (any) 138 49.46 83 47.98 141 50.54 90 52.02 0.3357

Serious Falls 34 45.95 20 45.45 40 54.05 24 54.55 0.5137

Serious AEs - if admit to hospital or death 147 48.68 88 47.31 155 51.32 98 52.69 0.2139

APPENDICES

PAGE 247

Appendix 7: Questionnaires

This section includes the ASPIRE Older Person questionnaire, the caregiver questionnaire and the OPERA sub-study questionnaire.

Note. Data were collected electronically using software developed by The University of Auckland and therefore these questionnaires are included for completeness only.

APPENDICES

PAGE 248

OPERA Questionnaire

QUESTIONNAIRE FOR OLDER PEOPLE OPERA Hamilton Lower Hutt Christchurch

Date of interview…………….. Participant number ……………….

SECTION ONE

Q1 Tell me about what you know of community help.

Q2 Tell me about your family involvement.

5 4 3 2 1Q3 The following questions are about how much Strongly Neither agree Strongly

control you feel that you have over your life agree or disagree disagreea I can influence many of the things

that happen to meb I am confident I can solve most

of the problems I havec I can do just about anything if I

am determined enough to do itd What happens to me in the future

mostly depends on mee Sometimes I feel that I am being

pushed around5 4 3 2 1

Q4 Do you worry about your family doing too Hardly Verymuch? ever often

5 4 3 2 1Q5 Do you feel that you are causing them Hardly Very

too much concern? ever often

5 4 3 2 1Q6 Were you given any other alternatives of Residential With my In a retiement

where you could live? care family village

5 4 3 2 1Q7 Do you worry about money? Hardly Very

ever often

5 4 3 2 1Q8 What influence did money have on where Very strong Strong Small Minimal No

you would live influence influence influence influence influence

5 4 3 2 1Q9 Did any of the following people have an influence Very strong Strong Some Minimal No

on where you are living now? influence influence influence influence influencea Geriatricianb Hospital Nursec District Nursed Social Workere Occupational therapistf Physiotherapist g General Practitionerh Practice NurseI NASCj Friendsk Familyl Myself

Agree Disagree

Never Sometimes Often

No

Never Sometimes Often

In own home

Never Sometimes Often

APPENDICES

PAGE 249

Q10 Tell me about that decision

5 4 3 2 1Q11 Do you feel happy or sad about the decision Very No strong Very

of where you are living? happy feeling sad

Q12 If there was one thing that you could change what would it be?

SECTION TWO Only for those participants living in the community4 3 2 1

Q13 Are there concerns at home which would stop Strongly Moderately Minimallyyou staying there? concerned concerned concerned

a Fear of fallingb Fear of burglaryc Housingd Environmental aspectse People at homef My inability to copeg Lonelinessh Personal hygieneI Personal healthj Shoppingk No support from the communityL No close familym Other

SECTION THREE Only for those participants living in residential care4 3 2 1

Q14 What were the concerns at home which stopped Strongly Moderately Minimallyyou staying there? concerned concerned concerned

a Fear of fallingb Fear of burglaryc Housingd Environmental aspectse People at homef My inability to copeg Lonelinessh Personal hygieneI Personal healthj Shoppingk No support from the communityL No close familym Other

5 4 3 2 1Q15 If all the support you need was provided where Residential Living with In my own

ever you wanted, where would you rather be now? care my family home Hospital Other

Happy Sad

Unconcerned

Unconcerned

APPENDICES

PAGE 250

OPERA QUESTIONNAIRE FOR THE PRIMARY CAREGIVER Hamilton Lower Hutt Christchurch

Date of interview…………….. Participant number ……………….SECTION ONE

Q1 Tell me about what you know of community help.

Q2 Tell me about what happened after the older person was assessed as needing care

5 4 3 2 1Q3 Did any of the following people influence the Very strong Strong Some Minimal

decision of where the older person should live? influence influence influence influencea Geriatricianb Hospital Nursec District Nursed Social Workere Occupational therapistf Physiotherapistg General Practitionerh Practice NurseI NASCj Friendsk FamilyL The older person

Q4 Tell me about that decision

Q5 Do you feel there were any other alternativesif so what were they?

5 4 3 2 1Q6 How do you feel about the decision? Very Very

happy sad

4 3 2 1Q7 Money influenced the decision of where Strong Moderate Minimal no

the older person would live. influence influence influence influence

For primary care givers of a community resident please go to Section TwoFor primary care givers of a residential care resident please go to Section Three

No influence

Happy SadRelieved

APPENDICES

PAGE 251

SECTION TWO Only for caregivers of community residents

4 3 2 1Q8 Are there concerns at home which would stop Strongly Moderately Minimally

the older person staying there? concerned concerned concerneda Fear of fallingb Fear of burglaryc Housing d Environmental aspectse People at homef Their inability to copeg Their lonelinessh Their personal hygieneI Their healthj Shoppingk No support from the communityL No close familym Other

SECTION THREE Only for caregivers of residential care residents

5 4 3 2Q9 What were there concerns at home which Strongly Moderately Minimally

stopped the older person staying there? concerned concerned concerneda Fear of fallingb Fear of burglaryc Housing d Environmental aspectse People at homef Their inability to copeg Their lonelinessh Their personal hygieneI Their healthj Shoppingk No support from the communityL No close familym Other

Unconcerned

Unconcerned

APPENDICES

PAGE 252

SURVEY FOR THE GENERAL PRACTITIONER Hamilton Lower Hutt Christchurch

Date of interview…………….. Participant number ……………….

5 4 3 2 1Q1 If safety was an issue would you make placement Hardly Very

suggestions? ever often

5 4 3 2 1Q2 How experienced do you consider yourself to be Very Somewhat Very

in determining placement issues for the older person? experienced experienced inexperienced

5 4 3 2 1Q3 Do you discuss support needs with Hardly Very

the older person? ever often

5 4 3 2 1Q4 Do you inform the older person when Hardly Very

you feel that staying at home is not an option? ever often

5 4 3 2 1Q5 Would you consider any of the following to be Hardly Very

mandatory for placement within residential care ever oftena No familyb Environmental hazardsc Patient depressiond Dementiae Co-morbid medical conditionf Their fear of fallingg Their personal safetyh Decreased level of ADLsI Positive attitude to residential carej Other

Q6 Can you tell me what you do when you find that a person needs more care than is presently provided in the home?

The following questions are specifically about #……….. Patient's name.

5 4 3 2 1Q7 In your opinion what was the main influence on the Environment Dementia Medical Fear of

placement decision? hazards condition falling

10 9 8 7 6Personal Decreased Positive attitude Home care Very good

safety level of ADLs to resthome available home situation

5 4 3 2 1Q8 Did you discuss the future with the older person's Hardly Very

family? ever often

5 4 3 2 1Q9 Do you feel that the older person was happy with Very Very

your suggestions? happy sad

5 4 3 2 1Q10 Do you feel that the closest family member was Very Very

happy with your suggestions happy sad

5 4 3 2 1Q11 Did any of the following people influence the Very strong Strong Some Minimal

decision of where the older person should live? influence influence influence influencea Geriatricianb Hospital Nursec District Nursed Social Workere Occupational therapistf Physiotherapistg General Practitionerh Practice NurseI NASCj Friendsk FamilyL The older person

No influence

Never Sometimes Often

Experienced Inexperienced

Often

Relieved

Never Sometimes Often

Never Sometimes

Never Sometimes Often

No family

Happy Sad

Never Sometimes Often

Happy Sad

Relieved

APPENDICES

PAGE 253

SURVEY FOR THE MULTI-DISCIPLINARY TEAM Hamilton Lower Hutt Christchurch

Date of interview…………….. Participant number ……………….

5 4 3 2 1Q1 Which people are involved in the multi-disciplinary Occupational Physio Social

team meetings? therapist therapist worker

10 9 8 7 6Domiciliary The older

staff person

5 4 3 2 1Q2 How experienced do you feel the team is Very Somewhat Very

in making discharge and support needs decisions? experienced experienced Inexperienced

5 4 3 2 1Q3 Does the team discuss the support needs with Hardly Very

the older person? ever often

5 4 3 2 1Q4 Does your team make suggestions for placement? Hardly Very

ever often

5 4 3 2 1Q5 Does the team inform the older person when Hardly Very

going home should not be an option? ever often

5 4 3 2 1Q6 If so who informs the older person? Occupational Physio Social

therapist therapist worker

10 9 8 7 6Domiciliary

staff

5 4 3 2 1Q7 Would you consider any of the following to be Hardly Very

mandatory for placement within residential care ever oftena No familyb Environmental hazardsc Patient depressiond Dementiae Co-morbid medical conditionf Their fear of fallingg Their personal safetyh Decreased level of ADLsI Positive attitude to residential carej Other

Often

Never Often

Never

Never Sometimes

Doctor

Never

NASC Family

Doctor Nurse

NASC Family Other

Experienced Inexperienced

Other

Sometimes Often

Nurse

Sometimes Often

GP

Sometimes

APPENDICES

PAGE 254

OPERA SURVEY FOR THE NASC Hamilton Lower Hutt Christchurch

Date of interview……………………………………. Participant number ……………….

5 4 3 2 1Q1 How experienced do you feel the team is Very Somewhat Very

in making placement decisions? experienced experienced inexperienced

5 4 3 2 1Q2 Does the team discuss the placement decisions Hardly Very

with the older person? ever often

5 4 3 2 1Q3 Does the team inform the older person when Hardly Very

going home should not be an option? ever often

5 4 3 2 1Q4 Would you consider any of the following to be Hardly Very

mandatory for placement within residential care ever oftena No familyb Environmental hazardsc Patient depressiond Severe Dementiae Co-morbid medical conditionf Their fear of fallingg Their personal safetyh Decreased level of ADLsI Positive attitude to residential carej Other

ns are specifically about #……….. Patient's name.

5 4 3 2 1Q5 What was the major influence on the Environment Dementia Medical Fear of

placement decision? hazards condition falling

10 9 8 7 6Personal Decreased Positive attitude Home care Very good

safety level of ADLs to resthome available home situation

5 4 3 2 1Q6 Do you feel that the older person was happy with Very Very

the outcome of the NASC meeting? happy sad

5 4 3 2 1Q7 Do you feel that the closest family member was Very Very

happy with the outcome of the NASC meeting? happy sad

5 4 3 2 1Q8 Did any of the following people influence the Very strong Strong Some Minimal

decision of where the older person should live? influence influence influence influencea Geriatricianb Hospital Nursec District Nursed Social Workere Occupational therapistf Physiotherapistg General Practitionerh Practice NurseI NASCj Friendsk FamilyL The older person

Q9 Tell me about anything else which was a factorin the placement decision?

No influence

Experienced Inexperienced

Never Sometimes Often

SometimesNever Often

Never Sometimes Often

Happy Sad

Sad

No family

Relieved

Relieved

Happy

APPENDICES

PAGE 255

The following questions are specifically about #……….. Patient's name.

5 4 3 2 1Q8 What was the major influence on the Environment Dementia Medical Fear of

placement decision? hazards condition falling

10 9 8 7 6Personal Decreased Positive attitude Home care Very good

safety level of ADLs to resthome available home situation

5 4 3 2 1Q9 Who discussed the future with the older person? Occupational Physio Social

therapist therapist worker

10 9 8 7 6Domiciliary

staff

5 4 3 2 1Q10 Who discussed the future with the older person's Occupational Physio Social

family? therapist therapist worker

10 9 8 7 6Domiciliary The older

staff person

5 4 3 2 1Q11 Do you feel that the older person was happy with Very Very

the outcome of the teams meeting? happy sad

5 4 3 2 1Q12 Do you feel that the closest family member was Very Very

happy with the outcome of the teams meeting? happy sad

5 4 3 2 1Q13 Did any of the following people influence the Very strong Strong Some Minimal

decision of where the older person should live? influence influence influence influencea Geriatricianb Hospital Nursec District Nursed Social Workere Occupational therapistf Physiotherapistg General Practitionerh Practice NurseI NASCj Friendsk FamilyL The older person

No influence

Other

Doctor

No family

Doctor Nurse

Other

NASC Family

NASC Family GP

Relieved

Relieved

Nurse

Happy Sad

Happy Sad

Note. Older person's questionnaire, questions 4 and 8 and 13 and 14 are reverse scored

APPENDICES

PAGE 256

ASPIRE Older Person Questionnaire

APPENDICES

PAGE 257

APPENDICES

PAGE 258

APPENDICES

PAGE 259

APPENDICES

PAGE 260

APPENDICES

PAGE 261

APPENDICES

PAGE 262

APPENDICES

PAGE 263

APPENDICES

PAGE 264

APPENDICES

PAGE 265

APPENDICES

PAGE 266

APPENDICES

PAGE 267

APPENDICES

PAGE 268

APPENDICES

PAGE 269

APPENDICES

PAGE 270

APPENDICES

PAGE 271

APPENDICES

PAGE 272

APPENDICES

PAGE 273

APPENDICES

PAGE 274

APPENDICES

PAGE 275

APPENDICES

PAGE 276

APPENDICES

PAGE 277

APPENDICES

PAGE 278

APPENDICES

PAGE 279

APPENDICES

PAGE 280

APPENDICES

PAGE 281

ASPIRE: Older Person Questionnaire

REFERENCES

PAGE 282

References:

AAlen OO. Analysis of Dependent Survival data applied to lifetimes of Analgam Fillings. Statistical in Medicine 1995;14:1819-1829. Aiken, L. H., S. P. Clarke, et al. (2001). "Cause for concern: nurses' reports of hospital care in five countries." LDI Issue Brief 6(8): 1-4. Bear, M. (1990). "Social networks and health: impact on returning home after entry into residential care homes." Gerontologist 30(1): 30-4. Bergen, A. (1992). "Case management in community care: concepts, practices and implications for nursing." Journal of Advanced Nursing 17(9): 1106-13. Bergen, A. (1994). "Case management in the community: identifying a role for nursing." Journal of Clinical Nursing 3(4): 251-7. Bergen, A. (2003). "Care management revisited: a follow-up study." British Journal of Community Nursing 8(1): 16-23. Bernabei, R., F. Landi, et al. (1998). "Randomised trial of impact of model of integrated care and case management for older people living in the community." BMJ 316(7141): 1348-51. Blackhall,L.J., Frank,G., Murphy,S.T., Michel,V., Palmer,J.M., & Azen,S.P. (1999). Ethnicity and attitudes towards life sustaining technology. Social Science & Medicine 48(12), 1779-1789. Bookwala,J., Coppola,K.M., Fagerlin,A., Ditto,P.H., Danks,J.H., & Smucker,W.D. (2001). Gender differences in older adults' preferences for life-sustaining medical treatments and end-of-life values. Death Studies . 25(2). Braun,K.L., Onaka,A.T., & Horiuchi,B.Y. (2001). Advance directive completion rates and end-of-life preferences in Hawaii. Journal of the American Geriatrics Society 49(12), 1708-1713. Bryan, B., D. Dickerson, et al. (1994). "Nursing case management: affecting quality and cost." Hospital Materiel Management Quarterly 16(2): 63-75. Budge, C., J. Carryer, et al. (2003). "Health correlates of autonomy, control and professional relationships in the nursing environment." J Adv Nurs 42(3): 260-268. Burrows AB, Morris JN, Simon SE, Hirdes JP, Phillips C. Development of a minimum data set-based depression rating scale for use in nursing homes. Age & Ageing 2000;29(2):165-72. Burrows, A. B., J. N. Morris, et al. (2000). "Development of a minimum data set-based depression rating scale for use in nursing homes." Age & Ageing 29(2): 165-72. Carmel,S. (1999). Life-sustaining treatments: what doctors do, what they want for themselves and what elderly persons want. Social Science and Medicine 49(10), 1401-1408.

REFERENCES

PAGE 283

Carmel,S., & Mutran,E.J. (1999). Stability of elderly persons' expressed preferences regarding the use of life-sustaining treatments. Social Science & Medicine 49(3), 303-311. Carpenter, I., Francis, S., Roberts, S., & Wayman, C. (2002, May). InterRAI UK MDS Home Care Assessment Instrument For Community Care User's Manual [Internet Website]. InterRAI UK. Retrieved 9th March, 2003, from the World Wide Web: http://www.interrai.org Cicirelli,V.G., MacLean,A.P., & Cox,L.S. (2000). Hastening death: A comparison of two end-of-life decisions. Death Studies 24(5), 401-419. Coleman, B. J. (1995). "European models of long-term care in the home and community." International Journal of Health Services 25(3): 455-74. Colvez A, Joel ME, Ponton-Sanchez A, Royer AC. Health status and work burdon of Alzheimer patients' informal caregivers. Health Policy. 2002;60:219–33. Coppola,K.M., Bookwala,J., Ditto,P.H., Lockhart,L.K., Danks,J.H., & Smucker,W.D. (1999). Elderly adults' preferences for life-sustaining treatments: The role of impairment, prognosis, and pain. Death Studies 23 617-634. Crawley, W. D. (1994). "Case management: managing the nurse case manager." Health Care Supervisor 12(4): 84-9. Curtis,J.R., Patrick,D.L., Engelberg,R.A., Norris,K., Asp,C., & Byock,I. (2002). A measure of the quality of dying and death. Initial validation using after-death interviews with family members. Journal of Pain & Symptom Management 24(1), 17-31. Dales FW. The Respite Care Scheme: A partnership in care. In: National Gerontology Conference; 1998. Dalziel, L. (2001). The New Zealand Positive Ageing Strategy. Lower Hutt, Ministry of Social Policy. Department of Health. National Service Framework for Older People. 2001. The Stationery Office. Ditto,P.H., Smucker,W.D., Danks,J.H., Jacobson,J.A., Houts,R.M., Fagerlin,A., Coppola,K.M., & Gready,R.M. (2003). Stability of older adults' preferences for life-sustaining medical treatment. Health Psychology 22(6), 605-615. Donner A, Klar N. Issues in the meta-analysis of cluster randomized trials. Stat Med 2002;21(19):2971-80. Donner A, Klar N. Issues in the meta-analysis of cluster randomized trials. Statistical in Medicine 2002;21:2971-2980. Donner A, Piaggio G, Villar J. Statistical methods for the Meta-analysis of cluster randomization trials. Statistical Methods in Medical Research 2001;10:325-338. Donner A. Design and analysis of cluster randomization trials in health research: London : Arnold, 2000., 2000. Dunn, R. B., P. M. Guy, et al. (1995). "Can a house call by a public health nurse improve the quality of the discharge process for geriatric patients?" Clinical Performance & Quality Health Care 3(3): 151-5.

REFERENCES

PAGE 284

Dyson, R. (2002). Health of Older People Strategy: health sector action to 2010 to support positive ageing. Lower Hutt, Ministry of Health. EFCCP Audit working party 2001. www.efgcp.prg/workingparties/audit/protcom1.htm

EFCCP Audit working party 2001. www.efgcp.prg/workingparties/audit/protcom1.htm

Eggert, G. M., J. G. Zimmer, et al. (1991). "Case management: a randomized controlled study comparing a neighborhood team and a centralized individual model." Health Services Research 26(4): 471-507. Emanuel,E.J., & Emanuel,L.L. (1998). The promise of a good death. The Lancet 351(1002), SII21-SII29. Fairrow,A.M., McCallum,T.J., & Messinger-Rapport,B.J. (2004). Preferences of older African-Americans for long-term tube feeding at the end of life. Aging & Mental Health 8(6), 530-534. Fitzgerald, J. F., D. M. Smith, et al. (1994). "A case manager intervention to reduce readmissions.[comment]." Archives of Internal Medicine 154(15): 1721-9. Fried,T.R., & Bradley,E.H. (2003). What matters to seriously ill older persons making end-of-life treatment decisions?: A qualitative study. Journal of Palliative Medicine 6(2), 237-244. Fried,T.R., Bradley,E.H., & Towle,V.R. (2002). Assessment of Patient Preferences: Integrating Treatments and Outcomes. Journals of Gerontology Series B: Psychological Sciences and Social Sciences 57(6), S348-S354. Fried,T.R., Bradley,E.H., Towle,V.R., & Allore,H. (2002). Understanding the Treatment Preferences of Seriously Ill Patients. The New England Journal of Medicine 346(14), 1061-1066. Fried,T.R., van,D.C., O'Leary,J.R., Tinetti,M.E., & Drickamer,M.A. (1999). Older persons' preferences for site of terminal care. Annals of Internal Medicine 131(2), 109-112. Fries BE, Simon SE, Morris JN, Flodstrom C, Bookstein FL. Pain in U.S. nursing homes: validating a pain scale for the minimum data set.[comment]. Gerontologist 2001;41(2):173-9. Fries, B. E., S. E. Simon, et al. (2001). "Pain in U.S. nursing homes: validating a pain scale for the minimum data set.[comment]." Gerontologist 41(2): 173-9. Gagnon, A. J., C. Schein, et al. (1999). "Randomized controlled trial of nurse case management of frail older people.[comment]." Journal of the American Geriatrics Society 47(9): 1118-24. Gallo,J.J., Straton,J.B., Klag,M.J., Meoni,L.A., Sulmasy,D.P., Wang,N.Y., & Ford,D.E. (2003). Life-sustaining treatments: what do physicians want and do they express their wishes to others? Journal of the American Geriatrics Society 51(7), 961-969. Gauthier,D.M., & Froman,R.D. (2001). Preferences for care near the end of life: scale development and validation. Research in Nursing & Health 24(4), 298-306. Gibson, S. J., S. M. Martin, et al. (1994). "CNS-directed case management. Cost and quality in harmony." Journal of Nursing Administration 24(6): 45-51.

REFERENCES

PAGE 285

Giuliano, K. K. and C. E. Poirier (1991). "Nursing case management: critical pathways to desirable outcomes." Nursing Management 22(3): 52-5. Given CW, Given B, Stommel M, Collins C, King S, Franklin S. The caregiver reaction assessment (CRA) for caregivers to persons with chronic physical and mental impairments. Research in Nursing & Health 1992;15(4):271-83. Given, C. W., B. Given, et al. (1992). "The caregiver reaction assessment (CRA) for caregivers to persons with chronic physical and mental impairments." Research in Nursing & Health 15(4): 271-83. Glidden DV. Modelling clustered survival data from multicentre clinical trials. Statistical in Medicine 2004;23:369-388. Gott,M., Seymour,J., Bellamy G, Clark,D., & Ahmedzai,S. (2004). Older people's views about home as a place of care at the end of life. Palliative Medicine 18 460-467. Gramelspacher,G.P., Zhou,X.H., Hanna,M.P., & Tierney,W.M. (1997). Preferences of physicians and their patients for end-of-life care. Journal of General Internal Medicine 12(6), 346-351. Hardy R, Thompson S. Detecting and Describing Heterogeneity in Meta-Analysis. Stat Med 1998;17:841-856. Hartmaier SL, Sloane PD, Guess HA, Koch GG, Mitchell CM, Phillips CD. Validation of the Minimum Data Set Cognitive Performance Scale: agreement with the Mini-Mental State Examination. Journals of Gerontology Series A-Biological Sciences & Medical Sciences 1995;50(2):M128-33. Hartmaier, S. L., P. D. Sloane, et al. (1995). "Validation of the Minimum Data Set Cognitive Performance Scale: agreement with the Mini-Mental State Examination." Journals of Gerontology Series A-Biological Sciences & Medical Sciences 50(2): M128-33. Harvard Medical School. What family members can do. 2002; Available online. URL: http://www.intelihealth.com/IH/ihtPrint/WSIHW000/32070/32074/350668.html?d=dmtContent&hide=t&k=basePrint Accessed 31 August, 2003.

Hawes, C., J. N. Morris, et al. (1995). "Reliability estimates for the Minimum Data Set for nursing home resident assessment and care screening (MDS)." Gerontologist 35(2): 172-8. Hawkins,N.A., Ditto,P.H., Danks,J.H., & Smucker,W.D. (2005). Micromanaging death: Process preferences, values, and goals in end-of-life medical decision making. Gerontologist 45(1), 107-117. Hayes JM. Respite for caregivers: A community-based model in a rural setting. J Gerontol Nurs. 1999;25:22–7. Hayes, R. J. and S. Bennett (1999). "Simple sample size calculation for cluster-randomized trials." International Journal of Epidemiology 28(2): 319-26. Hays,J.C., Galanos,A.N., Palmer,T.A., McQuoid,D.R., & Flint,E.P. (2001). Preference for place of death in a continuing care retirement community. Gerontologist 41(1), 123-128. Heeren,O., Menon,A.S., Raskin,A., & Ruskin,P. (2001). Religion and end of life treatment preferences among geriatric patients. International Journal of Geriatric Psychiatry 16(2), 203-208.

REFERENCES

PAGE 286

Help the Aged. End-of-Life: making decisions around the end-of-life. Help the Aged . 2002. 10-10-2005. Hendriksen, C., E. Lund, et al. (1984). "Consequences of assessment and intervention among elderly people: a three year randomised controlled trial." British Medical Journal Clinical Research Ed. 289(6457): 1522-4. Henwood,M., & Debate of the Age Health and Care Study Group. The future of health and care of older people: the best is yet to come. 1999. London, Age Concern. The Millennium Papers. Higginson,I.J. Priorities and Preferences for end of life care in England, Wales and Scotland. 2003. London, National Council for Hospice and Specialist Palliative Care Services. Hing, E. and B. Bloom (1991). "Long-term care for the functionally dependent elderly." American Journal of Public Health 81(2): 223-5. Hirdes J, Frijters D, Teare G. The MDS-CHESS Scale: A new measure to predict mortality in the institutionalised elderly. J Am Ger Soc 2003;51(1):96-100. Hirdes, J., D. Frijters, et al. (2003). "The MDS-CHESS Scale: A new measure to predict mortality in the institutionalised elderly." J Am Ger Soc 51(1): 96-100. Hughes, S. L., A. Giobbie-Hurder, et al. (1999). "Relationship between caregiver burden and health-related quality of life." Gerontologist 39(5): 534-45. Hurley,A.C., Volicer,L., & Blasi,Z.V. (2000). End-of-Life Care for Patients With Advanced Dementia. JAMA: The Journal of the American Medical Association 284(19), 2449-244a. Kahana,B., Dan,A., Kahana,E., & Kercher,K. (2004). The personal and social context of planning for end-of-life care. Journal of the American Geriatrics Society 52(7), 1163-1167. Kane, R. L. and R. A. Kane (1995). "Long-term care." JAMA 273(21): 1690-1. Kelner,M. (1995). Activists and delegators: elderly patients' preferences about control at the end of life. Social Science & Medicine 41(4), 537-545. Kinney, J. M. and M. A. Stephens (1989). "Hassles and uplifts of giving care to a family member with dementia." Psychology & Aging 4(4): 402-8. Klinkenberg,M., Willems,D.L., Onwuteaka-Philipsen,B.D., Deeg,D.J.H., & van der Wal,G. (2004). Preferences in end-of-life care of older persons: after-death interviews with proxy respondents. Social Science & Medicine 59(12), 2467-2477 Koffman,J., & Higginson,I.J. (2004). Dying to be home? Preferred location of death of first-generation black Caribbean and native-born white patients in the United Kingdom. Journal of Palliative Medicine 7(5), 628-636. Kramer, B. J. (1997). "Differential predictors of strain and gain among husbands caring for wives with dementia." Gerontologist 37(2): 239-49. Landi, F., E. Tua, et al. (2000). "Minimum data set for home care: a valid instrument to assess frail older people living in the community." Med Care 38(12): 1184-90. Landi, F., G. Gambassi, et al. (1999). "Impact of integrated home care services on hospital use." J Am Geriatr Soc 47(12): 1430-4.

REFERENCES

PAGE 287

Leichtentritt,R.D., & Rettig,K.D. (1999). Meanings and attitudes toward end-of-life preferences in Israel. Death Studies 23(4), 323-358. Leichtentritt,R.D., & Rettig,K.D. (2000). The good death: Reaching an inductive understanding. Omega-Journal of Death and Dying 41(3), 221-248. Litwin H. Social network type and morale in old age. Gerontologist 2001;41(4):516-24. Litwin, H. (1998). "Social network type and health status in a national sample of elderly Israelis." Social Science & Medicine 46(4-5): 599-609. Litwin, H. (2001). "Social network type and morale in old age." Gerontologist 41(4): 516-24. Lloyd, L (2000), Dying in Old Age: promoting wellbeing at the end of life, Mortality, 5(2), 171.88 Lynn,J., Teno,J.M., Phillips,R.S., Wu,A.W., Desbiens,N., Harrold,J., Claessens,M.T., Wenger,N., Kreling,B., & Connors,A.F., Jr. (1997). Perceptions by family members of the dying experience of older and seriously ill patients. SUPPORT Investigators. Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments. Annals of Internal Medicine 126(2), 97-106. Marie Curie Cancer Care. Views about dying at home: Survey of the UK general public. 2004. Ref Type: Report Matsumura,S., Bito,S., Liu,H., Kahn,K., Fukuhara,S., Kagawa-Singer,M., & Wenger,N. (2002). Acculturation of attitudes toward end-of-life care: a cross-cultural survey of Japanese Americans and Japanese. Journal of General Internal Medicine 17(7), 531-539. McHorney CA, Ware JE, Jr., Lu JF, Sherbourne CD. The MOS 36-item Short-Form Health Survey (SF-36): III. Tests of data quality, scaling assumptions, and reliability across diverse patient groups. Medical Care 1994;32(1):40-66. McHorney CA, Ware JE, Jr., Raczek AE. The MOS 36-Item Short-Form Health Survey (SF-36): II. Psychometric and clinical tests of validity in measuring physical and mental health constructs. Medical Care 1993;31(3):247-63. McHorney CA, Ware JE, Jr., Rogers W, Raczek AE, Lu JF. The validity and relative precision of MOS short- and long-form health status scales and Dartmouth COOP charts. Results from the Medical Outcomes Study. Medical Care 1992;30(5 Suppl):MS253-65. McHorney, C. A., J. E. Ware, Jr., et al. (1992). "The validity and relative precision of MOS short- and long-form health status scales and Dartmouth COOP charts. Results from the Medical Outcomes Study." Medical Care 30(5 Suppl): MS253-65. McHorney, C. A., J. E. Ware, Jr., et al. (1993). "The MOS 36-Item Short-Form Health Survey (SF-36): II. Psychometric and clinical tests of validity in measuring physical and mental health constructs." Medical Care 31(3): 247-63. McHorney, C. A., J. E. Ware, Jr., et al. (1994). "The MOS 36-item Short-Form Health Survey (SF-36): III. Tests of data quality, scaling assumptions, and reliability across diverse patient groups." Medical Care 32(1): 40-66. McKenzie, C. B., N. G. Torkelson, et al. (1989). "Care and cost: nursing case management improves both." Nursing Management 20(10): 30-4. Authors

REFERENCES

PAGE 288

McMurdo, M.E. (1999). Exercise in old age: time to unwrap the cotton wool. British Journal of Sports Medicine. 33 (5):295-6 Ministry of Health (2002). Support needs assessment and service co-ordination: Policy, procedure and information reporting guidelines. Lower Hutt. Moos, R. (1995). Work Environment Scale Manual 3rd Edn. Palo Alto, CA, Consulting Psychologists Press. Mor, V., O. Intrator, et al. (1997). "Changes in hospitalization associated with introducing the Resident Assessment Instrument.[comment]." Journal of the American Geriatrics Society 45(8): 1002-10. Morris JN, Fries BE, Mehr DR, Hawes C, Phillips C, Mor V, et al. MDS Cognitive Performance Scale. Journal of Gerontology 1994;49(4):M174-82. Morris JN, Fries BE, Morris SA. Scaling ADLs within the MDS. Journals of Gerontology Series A-Biological Sciences & Medical Sciences 1999;54(11):M546-53. Morris, J. F., BE; Bernabei, R; Ikegami, N; Gilgen, R; Steel, K; Carpenter I. (1996). RAI Home care assessment manual. Washington, DC, InterRAI Corporation. Morris, J. N., B. E. Fries, et al. (1994). "MDS Cognitive Performance Scale." Journal of Gerontology 49(4): M174-82. Morris, J. N., B. E. Fries, et al. (1997). "Comprehensive clinical assessment in community setting: applicability of the MDS-HC.[comment]." Journal of the American Geriatrics Society 45(8): 1017-24. Morris, J. N., B. E. Fries, et al. (1999). "Scaling ADLs within the MDS." Journals of Gerontology Series A-Biological Sciences & Medical Sciences 54(11): M546-53. Motenko, A. K. (1989). "The frustrations, gratifications, and well-being of dementia caregivers." Gerontologist 29(2): 166-72. National Statistics. Focus on Older People. National Statistics Online . 2005. 10-10-2005. Ref Type: Electronic Citation Navari,R.M., Stocking,C.B., & Siegler,M. (2000). Preferences of Patients With Advanced Cancer for Hospice Care. JAMA: The Journal of the American Medical Association 284(19), 2449. Nijboer C, Triemstra M, Tempelaar R, Sanderman R, van den Bos GA. Measuring both negative and positive reactions to giving care to cancer patients: psychometric qualities of the Caregiver Reaction Assessment (CRA). Social Science & Medicine 1999;48(9):1259-69. Nijboer, C., M. Triemstra, et al. (1999). "Measuring both negative and positive reactions to giving care to cancer patients: psychometric qualities of the Caregiver Reaction Assessment (CRA)." Social Science & Medicine 48(9): 1259-69. Pacala, J. T., C. Boult, et al. (1995). "Case management of older adults in health maintenance organizations." Journal of the American Geriatrics Society 43(5): 538-42. Parsons, M. et al (2003). Disability Workforce Survey I, The Ministry of Health, Wellington Parsons, M. et al (2004). Disability Workforce Survey II, The Ministry of Health, Wellington

REFERENCES

PAGE 289

Parsons, M. et al (2204) Disability Workforce Survey III, The Ministry of Health, Wellington Parson, M. et al (2004). Service User Survey, The Ministry of Health, Wellington Pathy, M. S., A. Bayer, et al. (1992). "Randomised trial of case finding and surveillance of elderly people at home.[comment]." Lancet 340(8824): 890-3. Patrick,D.L., Curtis,J.R., Engelberg,R.A., Nielsen,E., & McCown,E. (2003). Measuring and Improving the Quality of Dying and Death. Annals of Internal Medicine 139(5_Part_2), 410-415. Patrick,D.L., Engelberg,R.A., & Curtis,J.R. (2001). Evaluating the Quality of Dying and Death. Journal of Pain and Symptom Management 22(3), 717-726. Patrick,D.L., Pearlman,R.A., Starks,H.E., Cain,K.C., Cole,W.G., & Uhlmann,R.F. (1997). Validation of preferences for life-sustaining treatment: implications for advance care planning. Annals of Internal Medicine 127(7), 509-517. Pearlin, L. I., J. T. Mullan, et al. (1990). "Caregiving and the stress process: an overview of concepts and their measures.[comment]." Gerontologist 30(5): 583-94. Phipps,E., True,G., Harris,D., Chong,U., Tester,W., Chavin,S.I., & Braitman,L.E. (2003). Approaching the end of life: attitudes, preferences, and behaviors of African-American and white patients and their family caregivers. Journal of Clinical Oncology 21(3), 549-554. Rantz, M. J., D. R. Mehr, et al. (2000). "Initial field testing of an instrument to measure: observable indicators of nursing home care quality." J Nurs Care Qual 14(3): 1-12. Ratner,E., Norlander,L., & McSteen,K. (2001). Death at home following a targeted advance-care planning process at home: the kitchen table discussion. Journal of the American Geriatrics Society 49(6), 778-781. Rogers, M., J. Riordan, et al. (1991). "Community-based nursing case management pays off." Nursing Management 22(3): 30-4. Rosenfeld,K.E., Wenger,N.S., & Kagawa-Singer,M. (2000). End-of-life decision making - A qualitative study of elderly individuals. Journal of General Internal Medicine 15(9), 620-625. Runciman, P., C. T. Currie, et al. (1996). "Discharge of elderly people from an accident and emergency department: evaluation of health visitor follow-up." Journal of Advanced Nursing 24(4): 711-8. Salvage, A. V., D. A. Jones, et al. (1989). "Opinions of people aged over 75 years on private and local authority residential care." Age & Ageing 18(6): 380-6. Schiff,R., Rajkumar,C., & Bulpitt,C. (2000). Views of elderly people on living wills: interview study. BMJ 320(7250), 1640-1641. Seale,C. (2000). Demographic change and the experience of dying. In D.Dickenson, Johnson M, & J.Samson Katz (Eds.), Death, Dying and Bereavement (pp. 35-43). London: Sage. Seeman, T. E. and L. F. Berkman (1988). "Structural characteristics of social networks and their relationship with social support in the elderly: who provides support." Social Science & Medicine 26(7): 737-49. Seymour,J. (2003). Technology and "natural death": A study of older people. Z Gerontol Geriat 36(5), 339-346

REFERENCES

PAGE 290

Seymour,J., Gott,M., Bellamy,G., Ahmedzai,S.H., & Clark,D. (2004). Planning for the end of life: the views of older people about advance care statements. Social Science & Medicine 59(1), 57-68. Seymour,J.E., Bellamy,G., Gott,M., Ahmedzai,S.H., & Clark,D. (2002). Good deaths, bad deaths: older people's assessments of the risks and benefits of morphine and terminal sedation in end-of- life care. Health Risk & Society 4(3), 287-303. Shrout PE, Fleiss JL. Intraclass Correlations: Uses in Assessing Rater Reliability. Psychological Bulletin 1979;86(2):420-428. Singer,P.A., Martin,D.K., & Kelner,M. (1999). Quality End-of-Life Care: Patients' Perspectives. JAMA: The Journal of the American Medical Association 281(2), 163-168. Smith,R. A good death. British Medical Journal 320(129), 130. 2000. Ref Type: Journal (Full) StatisticsNewZealand (1996). Household Disability Survey. Lower Hutt. StatisticsNewZealand (2002). 2001 Census Snapshot 9 (Older People). Steel, K. (1991). "Home care for the elderly. The new institution." Archives of Internal Medicine 151(3): 439-42. Steinhauser,K.E., Christakis,N.A., Clipp,E.C., McNeilly,M., Grambow,S., Parker,J., & Tulsky,J.A. (2001). Preparing for the End of Life: Preferences of Patients, Families, Physicians, and Other Care Providers. Journal of Pain and Symptom Management 22(3), 727-737. Steinhauser,K.E., Christakis,N.A., Clipp,E.C., McNeilly,M., McIntyre,L., & Tulsky,J.A. (2000a). Factors considered important at the end of life by patients, family, physicians, and other care providers. JAMA 284(19), 2476-2482. Steinhauser,K.E., Clipp,E.C., McNeilly,M., Christakis,N.A., McIntyre,L.M., & Tulsky,J.A. (2000b). In search of a good death: Observations of patients, families, and providers. Annals of Internal Medicine 132(10), 825-832. Stewart,A.L., Teno,J., Patrick,D.L., & Lynn,J. (1999). The Concept of Quality of Life of Dying Persons in the Context of Health Care. Journal of Pain and Symptom Management 17(2), 93-108. Straton,J.B., Wang,N.Y., Meoni,L.A., Ford,D.E., Klag,M.J., Casarett,D., & Gallo,J.J. (2004). Physical functioning, depression, and preferences for treatment at the end of life: the Johns Hopkins Precursors Study. Journal of the American Geriatrics Society 52(4), 577-582. Strauss,A. (1971). Anguish: the case history of a dying trajectory. San Francisco: Sociology Press. Stuck, A. E., H. U. Aronow, et al. (1995). "A trial of annual in-home comprehensive geriatric assessments for elderly people living in the community.[comment]." New England Journal of Medicine 333(18): 1184-9. Swindle, D. N., J. L. Weyant, et al. (1994). "Nurse case management: collaborative beyond the hospital walls." Journal of Case Management 3(2): 51-5.

REFERENCES

PAGE 291

Teno,J.M., Casey,V.A., Welch,L.C., & Edgman-Levitan,S. (2001). Patient-Focused, Family-Centered End-of-Life Medical Care: Views of the Guidelines and Bereaved Family Members. Journal of Pain and Symptom Management 22(3), 738-751. Thomas,C., Morris,S.M., & Clark,D. (2004). Place of death: preferences among cancer patients and their carers. Social Science & Medicine 58(12), 2431-2444. Vetter, N. J., D. A. Jones, et al. (1984). "Effect of health visitors working with elderly patients in general practice: a randomised controlled trial." British Medical Journal Clinical Research Ed. 288(6414): 369-72. Vig,E.K., & Pearlman,R.A. (2003). Quality of Life while dying: a qualitative study of terminally ill older men. Journal of the American Geriatrics Society 51 1595-1601. Vig,E.K., Davenport,N.A., & Pearlman,R.A. (2002). Good deaths, bad deaths, and preferences for the end of life: A qualitative study of geriatric outpatients. Journal of the American Geriatrics Society 50(9), 1541-1548. Ware JE, Jr., Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Medical Care 1992;30(6):473-83. Ware, J. E., Jr. and C. D. Sherbourne (1992). "The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection." Medical Care 30(6): 473-83. Weinberger, M., B. Nagle, et al. (1994). "Assessing health-related quality of life in elderly outpatients: telephone versus face-to-face administration." J Am Geriatr Soc 42(12): 1295-9. Wenger GC. A network typography, from theory to practice. Journal of Aging Studies 1991;5(1):147-162. Wenger GC. Support Networks for Older People: A Guide for Practitioners.: Centre for Social Policy Research and Development, University of Wales, Bangor., 1994. Wenger, G. C. (1991). "A network typography, from theory to practice." Journal of Aging Studies 5(1): 147-162. Wenger, G. C. (1994). Support Networks for Older People: A Guide for Practitioners., Centre for Social Policy Research and Development, University of Wales, Bangor. Wenger, G. C. (1996). Social network research in gerontology. The sociology of Ageing: International Perspectives. V. Minichiello, N. Chappell, A. Walker and H. Kendig. Melbourne, Australia, International Sociological Association: 60-81. Wenger, G. C. (1997). "Social networks and the prediction of elderly people at risk." Aging & Mental Health 1: 311-320. Williams,R. (1990). A Protestant Legacy: attitudes to death and illness among older Aberdonians. Oxford: Clarendon Press. Wilson,D.M. (2000). End-of-life care preferences of Canadian senior citizens with caregiving experience. Journal of Advanced Nursing 31(6), 1416-1421. Wolf GT & Makuch RW (1980); “Editorial: A classification system for protocol deviations in Clinical trials” Cancer Clin Trials 3: 101-3. Young,M., & Cullen,L. (1996). A Good Death: conversations with East Londoners. London: Routledge.

REFERENCES

PAGE 292

Yurk,R., Morgan,D., Franey,S., Stebner,J.B., & Lansky,D. (2002). Understanding the continuum of palliative care for patients and their caregivers. Journal of Pain & Symptom Management . 24(5), 01.