An informant interview for the diagnosis of dementia and depression in older adults (IDD-GMS)

AN INFORMANT INTERVIEW FOR THEDIAGNOSIS OF DEMENTIA AND DEPRESSION

IN OLDER ADULTS (IDD-GMS)SIMON LEWIS1*, KATY HINCHCLIFFE2, CORNELIUS KATONA3 AND GILL LIVINGSTON4

1Specialist Registrar in Psychiatry, Department of Psychiatry and Behavioural Sciences,University College London Medical School, London, UK

2Senior Registrar in Psychiatry, Department of Psychiatry and Behavioural Sciences,University College London Medical School, London, UK

3Professor of Psychiatry of the Elderly, Department of Psychiatry and Behavioural Sciences,University College London Medical School, London, UK

4Senior Lecturer in Psychiatry, Department of Psychiatry and Behavioural Sciences,University College London Medical School, London, UK

SUMMARY

Background. There has been no instrument developed for the di�erential diagnosis of psychiatric conditions usingan informant. The present study describes the development and validation of an informant interview for the diagnosisof dementia and depression in older adults (IDD-GMS). The IDD-GMS, as its name indicates, is based upon thewell-established Geriatric Mental State Schedule (GMS).

Method. Thirty older adults with psychiatric illnesses were identi®ed. An informant/carer was interviewed using theIDD-GMS. Questions from the GMS were altered to re¯ect the informant nature of the interview. Validity wascompared to ICD-10 diagnoses. Interrater reliability was determined.

Results. Using a hierarchical diagnostic system, receiver operating characteristics demonstrated one optimalcutpoint for sensitivity, 413 for dementia and 416 for depression, and one for speci®city, 413 for dementia and410 for depression.

Conclusion. The validity and reliability of the IDD-GMS falls within acceptable limits and indicates that theIDD-GMS can be used as a diagnostic instrument for dementia and depression. The IDD-GMS represents the ®rstinformant interview to achieve this. # 1998 John Wiley & Sons, Ltd.

Int. J. Geriat. Psychiatry, 13: 298±309, 1998.

KEY WORDSÐgeriatric psychiatry; informant interview; dementia; depression; GMS

Informants are routinely interviewed in clinicalpsychiatric practice to elicit whether a patient'spresenting symptoms represent a change and toclarify premorbid personality and functioning. Thisinformation is essential for diagnosis, managementand prognosis. Despite this, there are not manystandardized interviews for informants. A Medlinesearch (1966±1996, key word `informant'), exclud-ing learning disability and child and adolescentpsychiatry, found few reports of informant

interviews designed to be used by themselves forpsychiatric assessment, as opposed to givingadditional information after patient interview.Those found were concerned with either cognitivedecline and de®cit (Jorm and Jacomb, 1989; Kosset al., 1993; Sano et al., 1995) or personalityassessment (Pilgrim et al., 1993; Riso et al., 1994).

Comparison between direct patient assessmentand informant interview has concluded that bothcognitive testing and informant report are e�ectivescreening tools for dementia (Mulligan et al.,1996). In conjunction with interview information,they can accurately measure change as opposed tomaking a diagnosis in both dementia and depres-sion (Jorm et al., 1995). Informant interviews have

CCC 0885±6230/98/050298±12$17.50 Received 17 September 1997# 1998 John Wiley & Sons, Ltd. Accepted 15 December 1997

INTERNATIONAL JOURNAL OF GERIATRIC PSYCHIATRY, VOL. 13: 298±309 (1998)

*Correspondence to: Dr S. Lewis, A8, Mental Health Care ofOlder People, The Whittington Hospital, Highgate Hill,London N19 5NF, UK. Tel: 0171 530 2308. Fax: 0171 530 2304.

also been shown to be a satisfactory measure ofthe duration of symptoms in Alzheimer's disease(Sano et al., 1995). In addition, in comparison withother screening instruments for dementia, theyhave lower refusal rates and have been lessin¯uenced by the location of interview than patientquestionnaires (Ritchie and Fuhrer, 1992).

In a clinical setting an informant interview thatcould be used to diagnose common psychiatricdisorders may assist in the initial assessment ofreluctant or inaccessible patients via a relative orcarer. For example, a relative or carer may makeinitial contact to voice their concerns about aperson for whom they care who either refuses to seea mental health professional or is unable orunwilling to answer questions. An informantinterview could enable an initial working diagnosisto be made that would be useful for subsequentcommunity outreach work. In addition, an inform-ant interview could be used to inform the decisionabout whether a formal Mental Health Actassessment is necessary.

An informant instrument which generated aninitial diagnosis could help in several ways inachieving the Health of the Nation goals (Depart-ment of Health, 1992). The government hasprioritized several areas of health outcome. Themental health priorities are: to improve signi®-cantly the health and social functioning of mentallyill people; to improve the overall suicide rate by atleast 15% by the year 2000; and to reduce thelifetime suicide rate of severely mentally ill peopleby 33% by the year 2000. Older people have anincreased risk of suicide (McClure, 1984; Tobiaset al., 1992). This is associated with mentaldisorderÐin particular depressionÐin more than90% (Barraclough et al., 1974; Rich et al., 1986).However, most older people with depressionremain untreated (Livingston et al., 1990; Katonaet al., 1997). An instrument which allowed thedetection of untreated psychiatric disorder in olderpeople would therefore not only indirectly addressthe priorities in improving functioning but mightalso improve detection and treatment of a group athigh risk of suicide.

There has been no instrument developed forthe di�erential diagnosis of psychiatric conditionsusing an informant alone. The present studydescribes the development and validation of aninformant interview for the diagnosis of dementiaand depression in older adults based upon theGeriatric Mental State Schedule (Copeland et al.,1976, 1986) (IDD-GMS).

METHOD

Subjects

This study had approval from the local ethicalcommittee and took place in Islington, NorthLondon. Between March 1995 and September1996, new referrals from all parts of the service ofa consultant psychiatrist for older people's catch-ment area were considered for entry into the study(GL). Consenting relatives of new patients aged 65or more with a psychiatric diagnosis were inter-viewed.

Inclusion criteria were:

(i) At least one hour contact time with the patienttwice a week at the time of assessment and overthe course of the previous year

(ii) No previous knowledge or contact betweeninterviewers (SLandKH) andpatient or subject

Development of the interview

Design of the Geriatric Mental State ScheduleÐinformant version (GMS-I). The original ques-tionnaire was an informant version of the widelyused and validated psychiatric interview for olderadults, the computerized version of the GeriatricMental State Schedule (GMS-AGECAT)(Copeland et al., 1976, 1986). The adaptationinvolved altering the form of the GMS questionsso that they could be appropriately asked of aninformant as opposed to a patient, without alteringpsychopathological content. This version, namedthe Geriatric Mental State ScheduleÐinformantversion (GMS-I), was administered to 19 carers ofpatients and then compared to a gold standard.This was two clinicians' agreed diagnosis based ona direct full clinical assessment of the patient and areview of the patient's notes and mental state at thetime of the study (GL and CK). This informationwas used to classify patients according to theguidelines of the ICD-10 Classi®cation of Mentaland Behavioural DisordersÐClinical descriptionsand diagnostic guidelines (World HealthOrganization, 1992).

Using this method we were able to diagnosedementia with satisfactory validity (kappa 0.89).However, we were not able to diagnose depressionsatisfactorily (kappa 0.31). Other disorders (schizo-phrenia, dissociative disorders, etc) were foundto occur too infrequently to obtain statisticalmeasures of validity. In the light of this, questionspertaining to them were excluded in the subsequentdevelopment of the IDD-GMS.

# 1998 John Wiley & Sons, Ltd. INT. J. GERIAT. PSYCHIATRY, VOL. 13: 298±309 (1998)

INFORMANT INTERVIEW FOR DIAGNOSIS OF DEMENTIA AND DEPRESSION 299

Development of the IDD-GMS. Questions fromthe GMS-I with the highest correlation coe�cients(kappa5 0.14) to agreed clinical diagnoses wereretained. On this basis a modi®ed questionnaire,the IDD-GMS, was developed. The IDD-GMSretains 36 questions and takes approximately15 minutes to administer (see Appendix). A diag-nostic hierarchical model derived from Foulds(1976) was used, whereby a diagnosis of dementiaby the IDD-GMS excluded a diagnosis of depres-sion. The IDD-GMS questions are scored in asimilar manner to the GMS (Copeland et al., 1986):A `0' indicated no pathology above a clinical thres-hold and a `1' or `2' a pathology that was either`moderate' or `severe' in nature. The scores forquestions were summated into categories fordementia or depression.

Validity and reliability

The IDD-GMS was tested for validity bycomparison with the ICD-10 diagnoses agreed bytwo clinicians (CK and GL) as described above.Interrater reliability was tested by one rater tapingtheir interviews (SL or KH) and the other makingratings according to the tape. SL rated nine of theoriginal interviews and KH eight.

Data analysis

The data were entered for subjects who had beeninterviewed using the GMS-I, extracting theIDD-GMS items, and subjects interviewed byadministration of the IDD-GMS. Receiver operat-ing characteristics (ROC curves) were constructedin order to ascertain the e�ect of varying cutpointson sensitivity and speci®city (Murphy et al., 1987).Data comparison utilized SPSS for Windows(Norusis, 1991). Kappa was calculated by tele-scoping the ICD-10 diagnoses into the categoriesof dementia or depression or neither diagnosisand comparing them to IDD-GMS diagnoses.Spearman correlation coe�cients were calculatedin order to obtain values for statistical signi®cance.In addition, sensitivity and speci®city were calcu-lated.

RESULTS

Subjects

Thirty informants were recruited to the study.Nineteen were interviewed using the GMS-I and 11using the IDD-GMS. An additional subject refused

to participate. There were 17 female and 13 maleinformants representing 19 female and 11 malepatients. Table 1 represents the age and sexdistribution of patients and informants. We werenot able to obtain ages for eight of the informants.

IDD-GMS diagnoses

Clinical consensus diagnoses are summarized inTable 2, which shows the ICD-10 and IDD-GMSdiagnoses for the patients using a cuto� of 413for dementia and 416 for depression. Table 3demonstrates the concordance of the IDD-GMSwith the gold standard.

Cutpoints

Figs 1, 2 and 3 demonstrate the ROC curves fordiagnoses of dementia, depression and combinedhierarchical diagnoses respectively.

(a) Hierarchical cutpoints: 413 for dementia;416 for depression (Fig. 3). The sensitivity andspeci®city for these values were 84% and 80%respectively. The IDD-GMS correctly identi®eddementia in 16 and depression in ®ve subjects.There was one false positive and four falsenegatives. It correctly determined the absence ofdementia and depression in four subjects. Kappavalues for the IDD-GMS diagnoses compared toICD-10 were 0.86 for dementia (p5 0.000001),0.71 for depression (p5 0.00001) and 0.52 forneither diagnosis (p5 0.005).

Table 1. IDD-GMS age and sex distribution of patientsand informants

Age (yr) Patients Informants

Male

5 45 Ð 1

45±55 Ð 3

55±65 Ð 2

65±75 7 2

75±85 3 1

4 85 1 Ð

Female

5 45 Ð 4

45±55 Ð 1

55±65 Ð 4

65±75 8 2

75±85 8 1

4 85 3 1


300 S. LEWIS, K. HINCHCLIFFE, C. KATONA AND G. LIVINGSTON

(b) Hierarchical cutpoints: 413 for dementia;410 for depression (Fig. 3). The sensitivity andspeci®city for these values were 96% and 60%respectively. Kappa values for the IDD-GMSdiagnoses compared to ICD-10 were 0.86 fordementia (p5 0.000001), 0.69 for depression(p5 0.00001) and 0.71 for neither diagnosis(p5 0.00001).

Interrater reliability

Kappa values for interrater reliability were 1.00for dementia and 0.87 for depression.

DISCUSSION

The IDD-GMS is a new informant interview forthe diagnosis of dementia and depression. It is bothvalid and reliable and has been standardized in asample representing the target population. Itdistinguishes those respondents with dementiafrom those with depression and both of thesegroups from patients with neither diagnosis.

Development and ®ndings

The development of the IDD-GMS was notwithout di�culty. The ®rst instrument, the GMS-I,

Table 2. IDD-GMS diagnosis compared to clinicians' ICD-10 diagnosis

Patient IDD-GMS

depression

IDD-GMS

dementia

ICD-10 diagnosis (GL and CK) ICD-10 code

1 0 � Unspeci®ed dementia F03

2 0 � Dementia in Alzheimer's disease F00.1

3 � 0 Organic depressive disorder F06.32


5 � 0 Severe depressive episode F32.3

6 0 � Dementia in Parkinson's disease F02.3

7 � 0 Moderate depressive episode F32.1

8 0 � Alcohol dementia F10.73

9 0 0 Well at time of informant interview

10 0 � Dementia in Parkinson's disease F02.3

11 � 0 Moderate depressive episode F32.1

12 � 0 Severe depressive episode F32.3

13 0 � Vascular dementia F01.1


15 � 0 Persistent delusional disorder F22

16 0 0 Unspeci®ed dementia F00.1

17 0 0 Paranoid schizophrenia, episodic with de®cit F20.01

18 0 0 Organic depressive disorder F06.32

19 0 0 Moderate depressive episode F32.1




23 0 0 Bipolar a�ective disorder, current episode manic F31.2




27 0 0 Dementia in Alzheimer's disease F00.1

28 0 � Unspeci®ed mental disorder due to brain damage F06.9

29 0 � Dementia in Alzheimer's disease, atypical F00.2

30 0 0 Well at time of informant interview

Note: � indicates an IDD-GMS diagnosis, 0 indicates no IDD-GMS diagnosis.

Table 3. Concordance of the IDD-GMS with goldstandard diagnoses

IDD-GMS

diagnosis

Gold standard ICD-10 diagnosis

Dementia Depression Neither diagnosis

Dementia 16 Ð Ð

Depression 1 5 1

Neither diagnosis 1 2 4



Fig. 1. ROC curve: IDD-GMS dementia data illustrating cutpoints for dementia (O)

Fig. 2. ROC curve: IDD-GMS depression data illustrating cutpoints for depression (D)



could not be used to diagnose depression withvalidity. There are a number of possible reasons forthe problems encountered. Firstly, some of theGMS-I depression questions may have beenmisleading. The word `depression' is in commonuse by lay people and has a multitude of meaningswhich are not necessarily the same as thoseemployed by mental health professionals. Itappeared that the terms `depression' and `nervousbreakdown' were used synonymously by inform-ants such that, when referring to dementia, theyanswered questions about depressed moodpositively. Secondly, some of the symptoms ofdementia and depression are common to bothdiagnoses and, as a result of using questions aboutthese symptoms, speci®city and discriminantvalidity for the diagnosis of depression mightbe a�ected. Thirdly, dementia and depression cancoexist. Lastly, psychological di�culties in theinformant may have, for example, increasednegative cognitions and thus had an e�ect on theway questions were answered. There is an increasein psychiatric disorder linked with being a carer ofan older adult with a psychiatric illness. Depression

has been demonstrated in a higher percentage ofcarers of people with a psychiatric illness asopposed to carers of the physically ill (24% vs8%; Livingston et al., 1996).

While both the sensitivity and speci®city of theIDD-GMS were satisfactory at the ®rst cutpoints,413 for dementia and 416 for depression, theIDD-GMS was less sensitive than at cutpoints413for dementia and 410 for depression. At theformer cutpoints, the IDD-GMS did not establisha diagnosis in four patients who were su�eringfrom dementia or depression. Two patients notsu�ering from either dementia or depression werecorrectly classi®ed as not having an IDD-GMSdiagnosis and one patient with a persistent delu-sional disorder was classi®ed as depressed by theIDD-GMS. The IDD-GMS could be used as ascreening instrument for dementia and depressionand the latter cutpoints would be more appropriateas overall sensitivity is increased from 84% to 96%.

Two patients were classi®ed as `well at time ofinformant interview'. Although all the patientsrecruited to the study had a psychiatric diagnosis atentry point, in a number of cases there was a delay

Fig. 3. ROC curve: combined IDD-GMS data illustrating hierarchical cutpoints for dementia and depression



between assigning a patient to the study and theIDD-GMS being administered. In these two cases,patients had recovered at the point of informantinterview and the IDD-GMS correctly determinedthe absence of dementia and depression.

Limitations

This study has two main limitations. Thenumber of subjects led to small numbers in somesubgroups and the absence of the possibility ofmaking diagnoses other than dementia and depres-sion, and as such it can be seen as a pilot study.Although 30 informant interviews yielded validand reliable results, it is important to validate theIDD-GMS more widely. The instrument wasdeveloped to diagnose dementia and depressionas they are the commonest psychiatric diagnoses inolder adults. It would be desirable to extend theIDD-GMS to include other diagnoses. In addition,the IDD-GMS could be modi®ed and piloted forother age groups.

In our study, the IDD-GMS was administeredby psychiatrists. Its ability to be used by non-medical mental health professionals or non-healthprofessionals is unknown. Further study would benecessary to see if the IDD-GMS could be used bya wider range of personnel.

Use of the IDD-GMS

It is not justi®able for a de®nitive diagnosis to bemade without a direct assessment of a person withsymptoms. However, use of the IDD-GMS mightcontribute, via carers or other informants, to anincreased detection of mental illness, leading toincreased implementation of treatment pro-grammes, and improve treatment rates for undiag-nosed older adults with dementia or depression.This could result in earlier detection of psychiatricdisorder and in the long term a reduction in suiciderates. Postmortem diagnosis of dementia ordepression could be possible via an informant.Psychological autopsy methodology has previouslybeen proposed as a feasible model to study suicidein the elderly (Younger et al., 1990). This wouldallow the development of rational preventativestrategies and targeting of resources to those athighest risk. The IDD-GMS's use might lead toimprovements in the three desired outcomes of TheHealth of the Nation document described above.

Conclusion

The validity and reliability of the IDD-GMSfalls within acceptable limits and indicates that theIDD-GMS can be used to diagnose dementia anddepression. The IDD-GMS represents the ®rstinformant interview to achieve this.

ACKNOWLEDGEMENTS

We are indebted to Professor J. Copeland forallowing us to use the GMS as an original basis forthe development of the IDD-GMS.

We are grateful to all the patients and carers whoconsented to be a part of this study.

We would like to thank Mr R. Blizard forhis assistance with statistical analysis andDr M. Blanchard for helpful comments.

REFERENCES

Barraclough, B. M., Bunch, J., Nelson, B. et al.(1974) One hundred cases of suicideÐclinical aspects.Brit. J. Psychiat. 125, 355±373.

Copeland, J. R. M., Kelleher, M. J., Kellett, J. M. et al.(1976) A semi-structured clinical interview for theassessment of diagnosis and mental state in the elderly:The Geriatric Mental State Schedule 1. Psychol. Med.6, 439±449.

Copeland, J. R. M., Dewey, M. E. and Gri�ths-Jones,H. M. (1986) Psychiatric case nomenclature and acomputerised diagnostic system for elderly subjects:GMS and AGECAT. Psychol. Med. 16, 89±99.

Department of Health (1992) The Health of the NationÐA Strategy for Health in England. HMSO, London.

Foulds, G. A. (1976) The Hierarchical Nature of PersonalIllness. Academic Press, London.

Jorm, A. F. and Jacomb, P. A. (1989) The InformantQuestionnaire on Cognitive Decline in the Elderly(IQCODE): Socio-demographic correlates, reliability,validity and some norms. Psychol. Med. 19,1015±1022.

Jorm, A. F., Mackinnon, A. J., Henderson, A. S. et al.(1995) The Psychogeriatric Assessment Scales: Amulti-dimensional alternative to categorical diagnosisof dementia and depression in the elderly. Psychol.Med. 25, 447±460.

Katona, C., Manela, M. and Livingston, G. (1997)Co-morbidity with depression in older people. AgeingMent. Health 1, 57±62.

Koss, E., Patterson, M. B., Ownby, R. et al. (1993)Memory evaluation in Alzheimer's disease. Care-givers' appraisals and objective testing. Arch. Neurol.50, 92±97.



Livingston, G., Manela, M. and Katona, C. (1996).Depression and other psychiatric morbidity in carersof elderly people living at home. Brit. Med. J. 312,153±156.

Livingston, G., Thomas, A., Graham, N. et al. (1990)The Gospel Oak project: The use of services bydependent elderly people in the community. HealthTrends 22, 70±73.

McClure, G. M. G. (1984) Trends in suicide rate forEngland and Wales 1975±1980. Brit. J. Psychiat. 144,119±126.

Mulligan, R., Mackinnon, A., Jorm, A. F. et al. (1996)A comparison of alternative methods for screeningfor dementia in clinical settings. Arch. Neurol. 53,532±536.

Murphy, J. M., Haynes, R. D. and Guyatt, G. H. (1987)Performance of screening and diagnostic tests:Application of receiver operating characteristicanalysis. Arch. Gen. Psychiat. 44, 550±555.

Norusis, M. J. (1991) SPSS/PC4. SPSS Inc.Pilgrim, J. A., Mellers, J. D., Boothby, H. A. et al. (1993)

Inter-rater and temporal reliability of the StandardisedAssessment of Personality and the in¯uence ofinformant characteristics. Psychol. Med. 23, 779±786.

Rich, C. L., Young, D. and Fowler, R. C. (1986) SanDiego suicide study: 1. Young versus old subjects.Arch. Gen. Psychiat. 43, 577±582.

Riso, L. P., Klein, D. N., Anderson, R. L. et al. (1994)Concordance between patients and informants on thepersonality disorder examination. Am. J. Psychiat.151, 568±573.

Ritchie, K. and Fuhrer, R. (1992) A comparative studyof the performance of screening tests for seniledementia using receiver operating characteristicsanalysis. J. Clin. Epidemiol. 45, 627±637.

Sano, M., Devanand, D. P., Richards, M. et al. (1995)A standardised technique for establishing onset andduration of symptoms of Alzheimer's disease. Arch.Neurol. 52, 961±966.

Tobias, C. R., Pary, R. and Lippmann, S. (1992)Preventing suicide in older people. Am. Fam. Physician45, 1707±1713.

World Health Organization (1992) International Classi®-cation of Diseases, 10th edn (ICD-10). WHO, Geneva.

Younger, S. C., Clark, D. C., Oehmig-Lindroth, R. et al.(1990) Availability of knowledgeable informants for apsychological autopsy study of suicides committed byelderly people. J. Am. Geriatr. Soc. 38, 1169±1175.

APPENDIX

An informant interview for the diagnosis of dementia and depression in older adults (IDD-GMS)

Date:

Interviewee number:

Type of interview:Non-reliability interviewer � 0Reliability interviewer � 1Reliability observer � 2

Rater:

Sex of patient:Female� 1; male� 2

Sex of carer:Female� 1, male� 2

Age of patient:

Date of birth:

Orientation

For this section keep strictly to the wording of the questions and the instructions. If the subject fails togive an appropriate reply (correct or incorrect), the question may be repeated up to 3 times. If the subjectspontaneously changes his/her reply, then rate the changed reply. If the subject says `I do not know', theinterviewer should ask (once): `Could you try to remember?'.

1. If you asked them, can/could they spell their own name?Can/could not spell both names 0 1 8 9



2. Can/could they remember their date of birth?They do/did not know 0 1 8 9

3. Can/could they remember their age?They do/did not know 0 1 8 9

4. If you asked them, do/did they know the date?Error in day of week (2� error of more than 1 day) 0 1 2 8 9Error in month (2� error of more than 1 month) 0 1 2 8 9Error in year (2� error of more than 1 year) 0 1 2 8 9

5. Do/did they know their address?(Probe if necessary to ensure they know full address, including city, suburb or 0 1 8 9postal district but not post or zip code) Gives/gave incorrect orincomplete address

6. Do/did they sometimes mistake people for someone else and talk about them as if they weresomeone else?

They do/did mistake a person for someone else and talk about them as if 0 1 8 9they were someone else

7. Do/did they ever talk aimlessly.Talks/talked in an aimless fashion without getting to the point 0 1 8 9

Worry

8. What kind of things do/did they worry about? What about money or family problems, their ownhealth or someone else's health?

If worries are presentÐanything else?If no worries mentioned, skip to question 10

How much do/did they worry?Worries/worried a lot (ie about one or two speci®c things) 0 1 2 8 9

9. Does/did this worrying bother them a lot? Was it unpleasant? (Can/could they stop themselvesworrying?) Do/did the thoughts keep coming back?

Unpleasant worrying which keeps/kept coming back or can/could not 0 1 8 9be stopped

General anxiety

In the following items, rate free-¯oating subjective fear or anxiety. Do not confuse with worrying. If thecarer says the subject has/had general anxiety based upon delusions or false beliefs, eg being followed, itshould be rated here.

10. Do/did they get frightened? (Very anxious?) (What made them feel that way?)Fear or anxiety, out of proportion to the event, if any, that provoked the feeling 0 1 2 8 9

Depression

11. Have/had they been sad (depressed, miserable, in low spirits, blue) recently?Probe for depressed mood 0 1 2 8 9

12. Have/had they cried at all?Has/had been crying 0 1 2 8 9

13. Do you think that they have/had felt like crying (wanted to cry) without actually weeping?Has/had felt like crying 0 1 8 9



14. Do/did they seem depressed/cry/feel like crying most of the time? How long does/did it last? (Just afew hours at a time or longer than that?) How long have/had they been like this?

Depression, crying or feeling like crying lasts/lasted longer than just the 0 1 8 9occasional few hoursDepression, crying or feeling like crying is/was present most of the time 0 1 8 9Present for at least 2 continuous weeks in the last month (establish which 0 1 8 9part of the month for comparison with other symptoms) or in themonth before death

15. Have/had they felt that life was not worth living?Has/had felt life was not worth living 0 1 2 8 9

16. Do/did they talk about the future? Do/did they seem pessimistic? Despairing? If so, do you knowwhy?

Not pessimistic, but has/had empty expectations (lives from day to day) 0 1 8 9If pessimisticÐWhy is/was that? Have/had they felt really hopeless (despairing)? 0 1 8 9Is/was pessimistic or future seems/seemed bleak or can/could see no future at all,or future seems/seemed unbearableA general feeling of hopelessness/despair 0 1 8 9Pessimism obviously warranted by circumstances 0 1 8 9

17. Do/did they ever mention life being a burden? Do/did they mention ending their lives? (Harmingthemselves)

Has/had ever felt suicidal or wished to be dead 0 1 8 9If never, skip to item 20

18. When was that? In the last month? In the last year? (Or in last month or year before death)Never� 0 Sometimes� 1In last month 0 1 8 9In last year 0 1 8 9Has/had felt a wish to be dead for at least 2 weeks in the last month, or in 0 1 2 8 9month before death

19. Did they actually try anything? When was that? What did they do? (Or plan to do?) Why do you thinkthey felt that way?

Code as 1 if subject committed suicideHas/had done something or planned to do something about killing self 0 1 8 9Has/had rejected suicide but has/had wished to be dead because life 0 1 8 9is/was a burdenIf does not, skip to question 20

Looks/looked tense or worried 0 1 8 9Looks/looked sad, gloomy, mournful or depressed 0 1 8 9Looks/looked or sounds/sounded apprehensive or fearful 0 1 8 9Are/were tearful or crying 0 1 8 9

Memory

20. Are/were they aware of forgetting anything in particular? (Probe)Forgets/forgot names of family or friends, or misnames them 0 1 2 8 9(Do not include transient mistakes)Forgets/forgot where he/she has placed things 0 1 2 8 9

21. Do/did they have to make more e�ort to remember things than they used to? When did this begin?Has/had to make a greater e�ort to remember things than used to 0 1 8 9

22. Do/did they forget the names of famous people, eg the Prime Minister?Does/did not recall name of Prime Minister 0 1 8 9



Observation

23. Do/did you think they have/had di�culty with their memory?In interviewee's opinion, subject has/had di�culty with his/her memory 0 1 2 8 9

Somatic dysfunction

24. What is/was their appetite like? Do/did they enjoy their food? Are/were they eating more or less thanusual?

Diminution in the desire for food 0 1 2 8 9Increase in the desire for food 0 1 2 8 9

25. Have/had they lost any weight during the last 3 months or in the 3 months before death? About howmuch?

Lost 10 lb (4.5 kg) or more over the past 3 months or 3 months before death 0 1 8 9

26. Have/had they had trouble sleeping? Have/had they taken anything to help them sleep? How longhas/was it gone/going on for? What is/was happening?

Trouble with sleep or recent change in pattern 0 1 8 9

27. Have/had they had any di�culty falling asleep? Do/did they lie awake for long periods of time waitingto sleep? (If tablets taken, rate what interviewee feels would have happened without them)

Di�culty in falling asleep 0 1 8 928. Have/had they recently been waking up early in the morning and found it impossible to get back to

sleep? At what time? Is/was that their usual time? How often does/did it happen?Wakes/woke up 2 hours or more before normal time of awakening and 0 1 8 9can/could not get back to sleep, most nights for at least 2 weeks in thelast month or before death

Coping

29. Do/did they ®nd it di�cult to cope with the things they have/had to do every day?(NB: Do not rate physical disability)

Are/were not coping properly with everyday routine 0 1 2 8 9

30. Are/were they doing more, less, or about the same as usual?Doing less than usual. Rate only restrictions not imposed by environment 0 1 2 8 9

Guilt

31. Do/did they tend to blame themselves or feel guilty about anything? Or mention feelings ofworthlessness? (How long have/had they felt like this?) Is/was it reasonable?

Excessive guilt or self-blame over past and present peccadilloes 0 1 2 8 9(Do not include justi®able or minor self-blame)

Irritability

32. Do/did they seem more irritable (angry) lately/just before they diedIs/was more irritable (angry) 0 1 2 8 9Is/was more irritable on most days for at least 2 weeks in the last month or 0 1 2 8 92 weeks in the last month before death

33. Do/did they get angry with themselves?Gets/got angry with self 0 1 2 8 9



Interest

34. How is/was their interest in things? (Do/did they keep up their interests?)Has/had less interest in things in the last month than he/she used to have 0 1 2 8 9

35. What have/had they enjoyed doing? (Has/had there been any change?) (Did they used to enjoy doingthings?)

Almost nothing enjoyed 0 1 2 8 9If no decrease in interest or enjoyment, skip to end

36. When did you notice this loss of interest/enjoyment? When did it start? Has/had it been presentrecently? For how long? Is/was it there most days?

Falling o� of interest/enjoyment gradual over several years 0 1 2 8 9Falling o� of interest/enjoyment has occurred only within the last 0 1 2 8 93 months/3 months before death

The IDD-GMS scoring system

The IDD-GMS is scored in a similar way to the Geriatric Mental State Schedule (GMS).A `pathological' answer scores a 1 or 2.When not de®ned, and a choice between scoring 1 or 2 exists, a degree of judgement is necessary. For

example, in question 11: `Have/had they been sad recently?'. If the carer implies that the patient has been`quite' sad, score 1. If the carer says that the patient has been `very' sad, score 2.

The scores for each question are summated and cuto� criteria are used. The scheme for scoring isdetailed below. Some questions appear in both diagnostic categories and, where stated, only partquestions are scored.

Dementia

1, 2, 3, 4, 5, 6, 7, 20, 21, 22, 23, 24, 25, 26, 29, 30, 32, (1st part), 33, 34, 36 (1st part)Diagnostic cuto�: 413

Depression

8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36 (2nd part)Diagnostic cuto�: 416 or 410



Documents

An informant interview for the diagnosis of dementia and depression in older adults (IDD-GMS)