Socio-demographic determinants of entry into and exit from long-term institutional care – a linked register based follow-up study of

older Finns

Pekka Martikainen

&

Elina Nihtilä

&

Heta Moustgaard

&

Others

Background on long-term institutional care

Long-term care costs are almost 50% of all health care cost

of the elderly, and institutional care is not the preferred living

arrangement of the elderly

With population ageing the demand and costs of long-term

care are expected to grow rapidly

At the proximate level entry is determined by health and

cognitive and physical functional status, but also more distal

socio-demographic factors have been shown to be important

Focus and aims

How institutional care is determined by living

arrangements (or marital status) and socioeconomic

position

We have three main focuses of interest: First entry into institutional care Duration of care Care over the life course

Types of data sources

Individual level data on the total Finnish population Based on

Censuses from 1970-2005 (every 5-years)Population registration 1987-2007 (annual)

Linked with data on e.g.Mortality (1971-2007)Hospital discharge and medicationTaxation

Linkage is based on social security numbers Allows for:

Longitudinal analysesMultilevel analyses: areas, couples/families

An example of data content for a study of An example of data content for a study of entry into institutional care (40% sample entry into institutional care (40% sample of 65+)of 65+)

Sociodemographic Sociodemographic factors:factors:SexSexAgeAgeMarital statusMarital statusLiving Living arrangements arrangements EducationEducationSocial classSocial classIncomeIncomeHousingHousingPartnerPartnerRegionRegion

Use of homecare Use of homecare servicesservices

Institutional care:Institutional care:

Care episodesCare episodes

Date of entryDate of entry

Date of exitDate of exit

Type of institutionType of institution

(in health & social (in health & social care sectors)care sectors)

Supply of care:Supply of care:

Regional coverage of Regional coverage of institutional careinstitutional care

STAKESStatisticsFinland

STAKES

STAKES

STAKES

Pension instituteHealth:Health:

MedicationMedication

Hospital dischargeHospital discharge

Definition of long-term institutional care

24-hour care in nursing homes, service homes, hospitals

and health centers lasting for over 90 days or confirmed

by a long-term care decision.

The over-90-days criterion was met if a patient had

stayed in the same institution or successively in different

institutions for the time required.

Approximately 75 per cent of first stays begun in hospitals

or health centres

Living arrangements(with spouse, alone, others)

Informal care

Adequate availability of informal care is an important

determinant of independent living in the community and

postpones entry into institutional care

The importance of informal care is amplified because of

deteriorating dependency ratio and potential shortage of

care staff in the formal care system

Distribution by gender and living arrangements. Finnish older adults aged 65 and over living in the community

72

36

22

50

614

0%

20%

40%

60%

80%

100%

Men (N=108 474) Women (N=172 248)

with others

alone

with spouse

Probability of survival without long-term institutionalisation by living arrangements among Finnish older adults living in the community at baseline

Nihtilä & Martikainen, Scandinavian Journal of Public Health 2008

Probability of survival without long-term institutionalisation by living arrangements among Finnish older adults living in the community at baseline

Nihtilä & Martikainen, Scandinavian Journal of Public Health 2008

Institutionalisation among Finnish men and women aged 65+; adjusted hazard ratios

Men WomenFisrt entry

Living arrangement (LA) with spouse 1.00 * 1.00 alone 1.72 * 1.34 * other 1.55 * 1.24 *Home ownership (HO) yes 1.00 1.00 no 1.70 * 1.35 *Household income (HHI) 1. Quartile (high) 1.00 1.00 2. Quartile 1.19 * 1.09 * 3. Quartile 1.38 * 1.22 * 4. Quartile (low) 1.39 * 1.28 *

First entry

Nihtilä & Martikainen, Scandinavian Journal of Public Health 2008

Life-events: widowhood

A major life-event among the elderly

(with spouse -> alone) May lead to:

decline in health (effects of grief) loss of emotional/social support loss of task support

These in turn may influence entry into care May interact strongly e.g. with poor health or low SES

Figure 1. Relative age-adjusted institutionalization rates in

relation to duration of bereavement, Finnish elderly 65+

(Reference = married)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

0-1 1-2 2-3 3-6 6-12 12-24 24-36 36-42 42+

Time since spouse's or partner's death (months)

Haz

ard

ratio

(ins

titut

iona

lizat

ion)

MEN

WOMEN

Nihtilä & Martikainen, American Journal of Public Health 2008

Duration of care & gender difference

Human experience; rather than first entry total

duration of care is what people are interested in Costs determined by duration of stay

Inequalities/differences may be compounded E.g. gender differences in long-term care use

Women have higher chances of entry (and may

enter in better health)Women stay for longer (lower chances of exit)

Hazard ratios (women vs. men) of institutionalisation

and mode of exit from institution

Total (N=280722) Those institionalised (N=35926)First entry Died in institution Returned home Total exit

Model Hazard ratio Hazard ratio Hazard ratio Hazard ratio

unadjusted 1.42 0.70 0.81 0.74age-adjusted 1.12 0.66 0.89 0.75age+living arrangement 0.97 0.68 0.92 0.77age+household income 1.10 0.67 0.90 0.75age+home ownership 1.09 0.66 0.90 0.75age+health status 1.08 0.67 0.88 0.75full model 0.94 0.68 0.91 0.77

Martikainen, Moustgaard, Murphy, Nihtilä, Koskinen, Martelin, Noro, The Gerontologist 2008

= >

Adjusted for age:

Women stay in care on average 1064 days

- if living with spouse at baseline 994 days

- if living alone at baseline 1105 days

Men stay in care on average 686 days

- if living with spouse at baseline 645 days

- if living alone at baseline 746 days

Income

Are the effects of income independent of other socio-demographic

factors?

The independent effects may relate to e.g. the ability to pay for private

home care

Household income / consumption unit Source: Tax Administration and the Social Insurance Institution Incomes of all household members, including wages, capital

income and taxable income transfers and accounts for taxes and

non-taxable income transfers. Adjusted for household composition

(OECD, 1982).

Table 3. Relative insitutitionalisation rates by income quintiles obtained from the different Cox regression models, Finnish women and men aged 65 years and over

BASIC EXPLANATORY MEDIATING ALLMODEL FACTORS FACTORS

Model (1) (2) (3) (4) (5) (6) (7)Control 1+living 1+education 1+car 1+house type 1+chronic All

variables (a) arrangements + social class + level of medical factors

+ home equipment in conditions

ownership dwelling

HR HR HR HR HR HR HRWomenHousehold income5. Quintile (highest) 1.00 1.00 1.00 1.00 1.00 1.00 1.004. Quintile 1.07 * 1.07 * 1.05 1.06 * 1.06 * 1.06 * 1.033. Quintile 1.23 * 1.22 * 1.17 * 1.21 * 1.21 * 1.21 * 1.14 *2. Quintile 1.30 * 1.26 * 1.20 * 1.27 * 1.27 * 1.26 * 1.15 *1. Quintile (lowest) 1.35 * 1.31 * 1.23 * 1.32 * 1.30 * 1.28 * 1.15 *

MenHousehold income5. Quintile (highest) 1.00 1.00 1.00 1.00 1.00 1.00 1.004. Quintile 1.16 * 1.17 * 1.05 1.10 * 1.16 * 1.15 * 1.07 *3. Quintile 1.40 * 1.37 * 1.19 * 1.26 * 1.38 * 1.41 * 1.20 *2. Quintile 1.52 * 1.42 * 1.26 * 1.31 * 1.49 * 1.51 * 1.18 *1. Quintile (lowest) 1.59 * 1.47 * 1.29 * 1.31 * 1.52 * 1.51 * 1.13 *

* Different from the reference group (1.00) at 5% significance level

(a) Control variables: age, first language, area characteristics (level of urbanisation, region of residence)

Nihtilä & Martikainen, Population Studies 2007

Use of care towards the end of life

Older age and people’s proximity to death are

the most significant determinants of health

status and health needs Part of older age groups’ higher cost will thus reflect

the greater number of people close to death Acute health care costs are strongly associated with

proximity to death, with more than a quarter of all

acute health care costs incurred in the last year of

life May imply that projections of costs for acute care

that do not account for proximity to death

overestimate future increases by up to 20-25%

Less is known about long-term care

Less is know about how social factors affect these

associations

Care use patterns before death have repercussion for

future care demand and costs, as well as individual

quality of life

Murphy and Martikainen

Murphy and Martikainen

Proximity to death is important for acute care, but age is

more important for long-term care

Marital status differentials are quite substantial and future

changes in marital status distributions may be important

for future projections of elderly use of health and social

services.

Summary

Overall, the results demonstrates the strength of register based data in the

analyses of long-term institutional care

The consequences of population ageing on long-term care are not simply

determined by number and age of people, and their individual functional

status

The examples presented here show that also social factors - living

arrangements, change in living arrangements and socioeconomic

characteristics - have important consequences for entry into long-term care

Summary continued …

Similar factors define duration of care with strong

evidence of compounding of differences

To the extent that LTC is determined by age rather than

proximity to death implies that cost saving similar to those

expected for acute care can not be obtained for LTC in

the coming decades