INTERNATIONAL JOURNAL OF GERIATRIC PSYCHIATRY, VOL.

12: 907±912 (1997)

PSYCHIATRIC DISORDERS IN ELDERLY FROM
THE INDIAN SUB-CONTINENT LIVING
IN BRADFORD
KUMUD BHATNAGAR1Ã AND JUDITH FRANK2

1

Consultant Psychiatrist for Elderly, Bradford Community Health NHS Trust, UK
2
Senior Experimental Ocer, University of Bradford and Bradford Community Health NHS Trust, UK

SUMMARY
In our community study of the prevalence of psychiatric disorders among elderly South Asian immigrants from the
Indian sub-continent (India, Pakistan and Bangladesh) living in Bradford (UK), we found depression in 20%, dementia in 4% and anxiety neurosis in 4%. Subjects were interviewed at their place of residence by a consultant psychiatrist familiar with their culture and language. The Hindi translation of the community version of the Geriatric
Mental State schedule (GMS-A) was also administered. Psychiatrist's ICD-9 diagnosis was compared with
GMS±AGECAT computerized diagnosis. We found low-level agreement in dementia cases (kappa 0.33) whereas the agreement in subjects with depression was high (kappa 0.81). In many subjects GMS-A made a diagnosis of dementia not diagnosed as cases by the psychiatrist, who had the bene®t of additional history information from carers in this population from a di€erent culture and educational background. These ®ndings are discussed along with suggestions and present limitations of GMS-A in the diagnosis of dementia in cross-cultural research. Larger studies are needed in this population (a) to ®nd out prevalence rates in countries of origin and (b) to investigate the author's (KB) observation of low rates of Alzheimer's type dementia in this population, which may have aetiological signi®cance.
# 1997 John Wiley & Sons, Ltd.
Int. J. Geriat. Psychiatry, 12, 907±912, 1997.
No. of Figures: 0. No. of Tables: 2. No. of References: 20.
KEY WORDS Ðepidemiology;

old age; mental disorder; dementia; depressive disorder; cross-cultural research

BACKGROUND TO THE STUDY
Bradford is an industrial city of 450 000 people in the north of England. According to the 1991
Census, there are 67 824 people aged 65 years and over, with 1270 South Asian immigrants from the
Indian sub-continent. These consist of 458 Indians,
770 Pakistanis and 42 from Bangladesh. In order to provide relevant psychiatric services to this group of ethnic minorities (whose numbers are set to rise in the future), we needed estimates of prevalence rates of psychiatric disorders in this population.
A literature search failed to ®nd any published study in this cultural group. While considerable research has been published from Europe (Kay et al., 1964; Copeland et al., 1987), the USA
Ã
Correspondence to: Dr K. Bhatnagar, Daisy Hill House,
Lyn®eld Mount Hospital, Heights Lane, Bradford, West
Yorkshire, UK BD9 6DP. Tel: 01274-363831. Fax: 01274363827.

CCC 0885±6230/97/090907±06$17.50
# 1997 John Wiley & Sons, Ltd.

(Copeland et al., 1987; Heyman et al., 1991),
South-East Asia (Kua, 1992) and Africa (Ben-Arie et al., 1983; Hendrie et al., 1995), the rates of psychiatric disorders have varied considerably.
Dementia rates of 2±8.6% and depression rates of 4.6±19.4% have been reported (Ben-Arie et al.,
1983; Copeland et al., 1987). We did not think we could apply these ®ndings to this speci®c cultural group and decided to do a community study to ®nd the prevalence rates. The consultant psychiatrist
(author KB) is familiar with the cultures, customs and main languages spoken by the immigrant
South Asians (Hindi, Urdu, Punjabi and Gujerati).
We decided to use a standardized psychiatric interview schedule for the study. The community version of the Geriatric Mental State (GMS-A)
(McWilliam et al., 1988) has been shown to be a reliable diagnostic instrument. Its Hindi translation was previously carried out by Sharma in 1989. In an earlier study (Copeland et al., 1987), good agreement was found (kappa 0.74) between raters'
Received 21 November 1996
Accepted 12 March 1997

908

K. BHATNAGAR AND J. FRANK

diagnosis and GMS±AGECAT computerized diagnosis. For depression and dementia the agreement was very good (kappa 0.80 and 0.88 respectively).
METHOD
Subjects for the study were selected by taking every
®fth Asian name from lists supplied by the Family
Health Service Authority of patients of general practitioners whose practices were located in inner city areas of Bradford (Little Horton, Manningham and Heaton), where the majority of the Asian population live. They were contacted by letter, which explained the study and encouraged participation, posted a week in advance of the interview.
The letter gave the date and time when the writer would visit the recipient's home accompanied by a doctor who would ask questions about their health.
They were informed that the doctor spoke Asian languages. A phone number with 24-hour answerphone cover was given in case the arranged time and date were inconvenient, or if there were any queries.
Subjects were visited at their place of residence.
On average each interview lasted about an hour.
During this time the psychiatrist administered the Hindi translation of GMS-A and recorded a clinical diagnosis of the person interviewed using
ICD-9 criteria (WHO, 1978). Living arrangements were recorded and additional background history obtained whenever relatives or carers were available.
GMS-A data from completed interviews were passed on for computerized diagnosis and case identi®cation by AGECAT programme (Dewey and Copeland, 1986). The agreement or concordance between psychiatrist's ICD-9 and GMS±
AGECAT diagnosis was statistically assessed using
Cohen's kappa (Cohen, 1960), which has a maximum value of 1 when agreement is perfect. Values of higher than 0.61 are considered to show good agreement. Prevalence rates and con®dence limits
(CL) were calculated.
RESULTS

was made with 22, while another 33 had moved house and a further 42 were out of the country on holiday to visit their relatives (mostly to Pakistan).
One person was hospitalized and one refused to participate. Thus it took us 213 contacts, made over several months, to successfully complete
100 interviews achieving a response rate of 50% of the e€ective sample.
Bradford's Asian elderly population consisted of 1270 immigrants from India, Pakistan and
Bangladesh, representing 36%, 61% and 3% respectively of the elderly population. The high mobility and travel to Pakistan resulted in a study population of 56% Indians, 36% Pakistanis and
8% Bangladeshis. Ocial as well as family interpreters were used, with eight Bengali-speaking elders. The di€erence in our study pro®le was not thought to be detrimental.
Thirty per cent of respondents were living with their spouse, 45% lived with sons or daughters.
Twenty-®ve per cent were living alone, including
15% in sheltered housing. Seventy-four per cent of respondents were male and 26% female, while in the community of elderly immigrant South Asians in Bradford there are 60% males and 40% females.
The age of subjects ranged from 65 to 89 years, the distribution being 51% between 65 and 69 years,
23% between 70 and 74 years, 13% between 75 and
79 years and 13% aged 80 and over.
Prevalence
The total prevalence of psychiatric disorders was
28% (CL 19.48%±37.07%) according to psychiatrist's ICD-9 diagnosis and 29% (CL 20.36%±
38.03% using GMS±AGECAT (Table 1). The rates for males and females were similar. The overall agreement was 84%, with a concordance of
84% and kappa 0.64 (Table 2).
Depression. Prevalence of depression diagnosed by the psychiatrist was 20% (CL 12.67%±28.75%)
Table 1. Prevalence of psychiatric disorders
Psychiatrist's diagnosis
All
Male
Female
(%)
(%)
(%)

The sample
Using a 1 X 5 selection method, 213 contacts were derived from 1065 elderly Asians who comprised
84% of the 1270 census population of Bradford.
One was under the age of 65 and 13 were deceased, giving an e€ective sample of 199. Of these, despite our persistent e€orts to trace subjects, no contact
INT. J. GERIAT. PSYCHIATRY, VOL.

12: 907±912 (1997)

Depression
Neurotic
Organic
Non-case
Total prevalence

20
4
4
72
28

20
4
4
72
28

19
4
4
72
27

AGECAT diagnosis (%)
20
2
7
71
29

# 1997 John Wiley & Sons, Ltd.

909

PSYCHIATRIC DISORDERS IN ELDERLY FROM INDIAN SUB-CONTINENT

and it was also 20% using GMS±AGECAT
(Table 2). There was full agreement in 17 of the
20 cases giving an overall agreement of 94%, a diagnostic concordance of 85% and kappa 0.81
(very good).
There was disagreement in three cases of depression diagnosed by the psychiatrist and di€erently by GMS±AGECAT. The latter had diagnosed two of these as subcases of depression and the third as anxiety neurosis. Examination of history showed features of depression as well as anxiety.
Also, there was disagreement in three cases of depression diagnosed by GMS±AGECAT but not by the psychiatrist. In one, the psychiatrist expressed uncertainty about caseness as the subject had realistic family worries, weepiness and anxiety symptoms. In the second, tearfulness and worries occurred when talking about all sons having gone abroad for several months but the symptoms were not persistent. The third revealed an extremely gruesome murder of a son and the subject felt life was not worth living; no biological symptoms were volunteered and the features suggested an understandable reaction to catastrophic stress.
Anxiety neurosis. Anxiety neurosis was diagnosed in 4% (CL 1.1%±9.9%) of subjects by the psychiatrist and in 2% (CL 0.24%±7.03%) by
GMs±AGECAT (Table 2). The diagnostic concordance reached was 20% and kappa 0.32. In one case there was full agreement and in the second the disagreement was at level of caseness. The remaining two showed histories of anxiety with some depressive symptoms and were classi®ed as depression subcases by GMS±AGECAT.
Dementia. Dementia was diagnosed in 4%
(CL 1.1%±9.9%) of cases by the psychiatrist and in 7% (CL 2.86%±13.83%) by GMS±AGECAT
(Table 2), reaching a diagnostic concordance of

29% and kappa 0.33. In two out of four cases there was a clear history of cerebrovascular accidents.
GMS±AGECAT agreed in only two of the four cases diagnosed as dementia by the psychiatrist. It queried a dementia subcase in the third, while the fourth was diagnosed by the psychiatrist as mild dementia with, depressive symptoms and as a depression subcase by GMS±AGECAT.
Additionally, GMS±AGECAT diagnosed ®ve subjects as dementia cases which were considered by the psychiatrist to be non-cases. All had complained of memory problems. Two were diagnosed by the psychiatrist to have benign forgetfulness of old age. A third was a non-case. The other two were illiterate, one gave `don't know' answers to memory questions and the other knew the correct Indian calendar month but not English dates. In these ®ve people competent social functioning was noted in the history.
DISCUSSION
Our study can be seen as a ®rst survey of the prevalence of mental disorders in elders from the Indian sub-continent, where some 20% of the world population lives. We also tried to ®nd the reliability of GMS-A in this cultural group.
Methodological issues
Like many of their counterparts in the developing world, our study sample was not used to participation in this form of survey. In spite of postal appointments being set up, it often took many visits and a high degree of persistence to secure successful participation of subjects. However, during the visits we found the cooperation of subjects excellent and carers hospitable. The importance of cultural competence and the linguistic

Table 2. Psychiatrist's diagnosis vs AGECAT diagnosis
AGECAT
diagnosis
Depression
Neurotic
Organic
Non-case
Psychiatrist total

Depression

Psychiatrist's diagnosis
Neurotic
Organic

17
1

1

2
20

3
4

Non-case

AGECAT total 3
2
2
4

5
64
72

20
2
7
71
100

Note: All conditions: overall agreement ˆ 84%; diagnostic concordance ˆ 84%; kappa ˆ 0X64 (good).
Depression: overall agreement ˆ 94%; diagnostic concordance ˆ 84%; kappa ˆ 0X81 (very good). Dementia: overall agreement ˆ 93%; diagnostic concordance ˆ 29%; kappa ˆ 0X33 (low).
# 1997 John Wiley & Sons, Ltd.

INT. J. GERIAT. PSYCHIATRY, VOL.

12: 907±912 (1997)

910

K. BHATNAGAR AND J. FRANK

ability of the interviewer cannot be overemphasized in research of this kind.
We are aware that one of the limitations of our study has been the use of a Hindi translation of
GMS-A in the entire elderly population from the
Indian sub-continent. Ideally, each of the subgroups studied, ie Punjabis, Gujeratis and Bengalis, should have been interviewed using validated instruments in each of their languages. It could be argued that such an approach would have tapped emotions more e€ectively and in¯uenced our
®ndings. However, it is worth noting that validated instruments for the elderly in the above languages simply do not exist. Also, Hindi is reasonably well understood by Punjabi and Gujerati elders, who form the majority (97%) of Bradford's Asian elders.
We thus decided to follow a pragmatic approach to evaluation of need as our prime purpose. Our task was ®rst and foremost an attempt to study this population, and we appreciate that further work may be needed to con®rm our ®ndings.
The problem of contact (as mentioned in the
Results section) was mostly due to the high mobility of our sample, particularly the elderly
Pakistanis. Many had gone to their home country for various periods, and also had moved house along with younger relatives. Despite repeat visits to addresses and attempts to gather information from neighbours, some people could not be contacted. To a lesser degree, there were problems due to the inaccuracy of the GP lists: 6% of the sample were dead and some names were incorrectly spelled. It is assumed this problem will lessen with the computerization of GP lists.
We had a lower response from Pakistani elderly and found in the main that it was due to their higher mobility, especially travelling to Pakistan on long holidays; they often have dual nationality, unlike elderly from India. Also, we are aware that to a small degree the Census may underenumerate.
However, this is likely to apply to all our study population and not just to the Pakistani subjects.
Depression
The higher rates of depression found in the
Bradford sample need to be con®rmed by larger studies. In an earlier community study (Copeland et al., 1987) the prevalence of depressive disorders in Liverpool was 11.3%, in New York 16.2% and in London 19.4%. The prevalence may have been due to additional social problems. High rates were also found in Cape Town, with South Africa
INT. J. GERIAT. PSYCHIATRY, VOL.

12: 907±912 (1997)

coloured elders using the Present State Examination schedule (Ben-Arie et al., 1983). This is in contrast to low rates found in Chinese elderly in
Singapore (Kua, 1992). The high rates observed could be a re¯ection of some constitutional vulnerability. However, our immigrant sample was living in the inner city with possible additional stresses contributing to depression. These include retirement from work, break-up of close family ties and social isolation, as well as low status of the elderly in the host community, which the older
South Asians in our survey may have found dicult to accept as it contrasts with their expectations. We noted that many Pakistani elders travel abroad but return to the UK when they are ill and this may partly account for the higher morbidity found in the e€ective sample. The expressed wish of many South Asian elders is to return to the country they originated from, but they stay in the United
Kingdom for family and health care reasons. The higher rates of depression could well have origins in higher reports of diabetes, heart disease and strokes in this group of immigrants. Our sample was predominantly male, re¯ecting the South Asian community census ®gures. It is dicult to say whether this a€ected our results signi®cantly as we failed to ®nd any sex di€erences in the prevalence rate. One observation when interviewing was the lack of guilt feelings and suicidal ideas in the depressed. In these cultures and religions, God is considered the sole arbiter of life and death. Loss of honour and social disgrace are often serious considerations for people.
Like other researchers, we found satisfactory rates of agreement in cases of depression between psychiatrist diagnosis and GMS±AGECAT. This reinforces our experience that across cultures the core symptoms of mood disorders are very similar, with culture adding colouring to some of them.
Dementia
Of considerable concern to us is the low level of agreement between psychiatrist and GMS±
AGECAT diagnosis in this cultural group. We found that many GMS-A questions designed to pick up organic disorder had serious shortcomings in a population from a di€erent culture, some of whom were only semi-literate. Some subjects had little concept of western calendar months and dates, their birth dates or name spelling. To them many questions seemed immaterial and of little practical value. Questions on subjective memory
# 1997 John Wiley & Sons, Ltd.

911

PSYCHIATRIC DISORDERS IN ELDERLY FROM INDIAN SUB-CONTINENT

problems seemed largely irrelevant. In the presence of dementia memory problems are often denied, while many cognitively intact persons openly admit to memory problems. Some subjects could not name United Kingdom prime ministers but were able to give the correct names of Asian leaders.
Islamic and Indian months were often remembered. A western rater would have had serious diculty in ®nding such subjects' competency.
While it could be argued that the diagnostic criteria used by psychiatrists may well be more restrictive, it is apparent that many questions about memory in GMS-A types of scales have a cultural and educational bias. Methodological diculties were also expressed in its use by others (Kua, 1992).
Diculties have been found too with the application of the Mini Mental State Examination
(MMSE) schedule, which tended to give low scores and high false positive dementia diagnoses in less educated American blacks and Hispanics (Anthony et al., 1982; Escobar et al., 1986; Bohnsted et al.,
1994). Illiterate Indian subjects scored less on the
Hindi version of the MMSE (Ganguli et al., 1995).
It is eminently clear that we need valid dementia scales that are culturally fair and free of educational bias. We found questions about current events in an individual's life, memory of meals taken, names of the visitors and religious festivals could serve the same purpose. Similarly, enquiries about the adequacy of social functioning were found to be of value. This needs con®rmation from informants and carers, which, while requiring extra e€ort, was found by us to be worthwhile rather than making a misdiagnosis of dementia. We feel strongly that it is not advisable in dementia research to adopt fully the item weights given to various memory questions in any scale developed from one particular educational and cultural group to another. It is with this in mind that a crossvalidation of a dementia screening test in a diverse ethnic population in Israel seems signi®cant
(Ritchie and Hallerman, 1989). They validated the IOWA screening test and found it to be very sensitive in diverse ethnic groups. An important feature is the fact that it does not assume literacy, and is relatively culturally fair. Instruments of this type are perhaps better tools for researchers in the cross-cultural ®eld.
It is the author's (KB) experience over the last decade working with elderly from the Indian subcontinent that Alzheimer's type of dementia seems very much less common than vascular dementia.
The small numbers in this particular study do not
# 1997 John Wiley & Sons, Ltd.

tell the whole story but only three out of 20 cases seen during the last 6 years seem to ®t Alzheimer's diagnosis, the majority of the others having cerebrovascular accidents in their history. Similar high rates of vascular dementia and low rates of
Alzheimer's dementia have been reported in black
Americans, Chinese Americans and Chinese in Singapore (Serby et al., 1987; Heyman et al.,
1991; Kua, 1994). Low risk for Alzheimer's type dementia was found in American Chinese and
Puerto Rican groups when compared to Jewish and
Italians in a study involving 6866 ®rst-degree relatives (Silverman et al., 1992). Further crosscultural research in these populations may reveal possible aetiological clues.
ACKNOWLEDGEMENT
We acknowledge the important contribution to our research made by Dr Vimal Sharma, Consultant Psychiatrist, Fazakerly Hospital, Liverpool. It was his translation of GMS-A into Hindi as well as his enthusiasm which made this study possible.
REFERENCES
Anthony, J. C., LeResche, L. U. and Folstein, M. F.
(1982) Limits of the `Mini-Mental State' as a screening test for dementia and delirium among hospital patients. Psychol. Med. 12, 397±408.
Ben-Arie, O., Swartz, L. and Teggin, A. F. (1983)
The coloured elderly in Cape TownÐa psychosocial, psychiatric and medical community survey. Part II.
Prevalence of psychiatric disorders. S. African Med. J.
64, 1056±1061.
Bohnstedt, M., Fox, P. J. and Kohastu, N. D. (1994)
Correlates of Mini-Mental State Examination scores among elderly demented patients: The in¯uence of race±ethnicity. J. Clin. Epidemiol. 47, 1381±1387.
Cohen, J. (1960) A coecient of agreement for nominal scales. Educ. Psychol. Measure. 20, 37±46.
Copeland, J. R. M., Dewey, M. E., Wood, N. et al.
(1987) Range of mental illness among elderly in the community. Prevalence in Liverpool using the GMS±
AGECAT package. Brit. J. Psychiat. 159, 815±823.
Copeland, J. R. M., McWilliam, C., Dewey, M. E. et al.
(1986) The early recognition of dementia in the elderly:
A preliminary communication about a longitudinal study using the GMS±AGECAT package (community version). Int. J. Geriatr. Psychiat. 1, 63±70.
Dewey, M. E. and Copeland, J. R. M. (1986) Computerised psychiatric diagnosis in the elderly: AGECAT.
J. Microcomput. Appl. 9, 135±140.
INT. J. GERIAT. PSYCHIATRY, VOL.

12: 907±912 (1997)

912

K. BHATNAGAR AND J. FRANK

Escobar, J., Burnam, A., Forsythe, A. et al. (1986)
Use of Mini-Mental State Examination (MMSE) in a community population of mixed ethnicity. Cultural and linguistic artefacts. J. Nerv. Ment. Dis. 170,
607±614.
Ganguli, M., Ratcli€, G., Chandra, V. et al. (1995) A
Hindi version of MMSE: The development of a cognitive screening instrument for a largely illiterate rural elderly population in India. Int. J. Geriatr.
Psychiat. 10, 367±377.
Hendrie, H. C., Osuntokun, B. O., Hall, K. S. et al. (1995) Prevalence of Alzheimer's disease and dementia in two communities: Nigerian Africans and African Americans. Am. J. Psychiat. 152, 1485±
1492.
Heyman, A., Fellenbaum, G., Prosnitz, B. et al. (1991)
Estimated prevalence of dementia among elderly black and white community residents. Arch. Neurol. 48,
594±598.
Kay, D. W. K., Beamish, P. and Roth, M. (1964) Old age mental disorders in Newcastle on Tyne. Part I. A study of prevalence. Part II. A study of possible social and medical causes. Brit. J. Psychiat. 110, 146±158,
668±682.

INT. J. GERIAT. PSYCHIATRY, VOL.

12: 907±912 (1997)

Kua, E. H. (1992) A community study of mental disorders in elderly Singaporean Chinese using the
GMS±AGECAT package. Austral. NZ J. Psychiat.
26, 502±506.
Kua, E. H. (1994) The ageing of elderly people.
Singapore Med. J. 35, 386±389.
McWilliam, C., Copeland, J. M. R., Dewey, M. E. et al.
(1988) The Geriatric Mental State Examination as a case-®nding instrument in the community. Brit. J.
Psychiat. 152, 205±208.
Ritchie, K. A. and Hallerman, E. F. (1989) Crossvalidation of a dementia screening test in a heterogenous population. Int. J. Epidem. 18, 717±719.
Serby, M., Chou, J. C. and Franssen, E. H. (1987)
Dementia in an American-Chinese nursing home population. Am. J. Psychiat. 144, 811±812.
Silverman, J. M., Li, G., Schear, S. et al. (1992) A crosscultural family history study of primary progressive dementia in relatives of non-demented elderly Chinese,
Italians, Jews and Puerto Ricans. Acta Psychiatr.
Scand. 85, 211±217.
World Health Organisation (1978) Mental Disorders
Glossary and Guide to their Classi®cation in Accordance with the Ninth Revision of the International
Classi®cation of Diseases (ICD-9). WHO, Geneva.

# 1997 John Wiley & Sons, Ltd.

M