A group of 1070 community-living persons aged 65 and over was assessed using the GMS-AGECAT package and other interviews at years 0 and 3. Year 3 interviewers were 'blind' to the findings at year 0, and the prevalence of organic disorders and depression was very similar in both years. According to the results at year 3, minimum and maximum prevalence figures for dementia at year 0 were 2.4% and 3.8% for moderate to severe and 0.4% and 2.4% for mild or early cases, with a best estimate of 3.5% and 0.8%, or 4.3% overall, divided into: senile, Alzheimer's type 3.3%; vascular 0.7%; and alcohol-related 0.3%. The overall incidence of dementia, clinically confirmed by six-year follow-up, was 9.2/1000 per year (Alzheimer type 6.3, vascular 1.9, alcohol related 1.0). Three years later, 72.0% of those with depressive psychosis and 62.3% of those with depressive neurosis were either dead or had some kind of psychiatric illness. Nearly 60% of milder depressive cases (7.2% of the total sample) had either died or developed a chronic mental illness. The outcome of depressive pseudodementias is equivocal so far. Findings at year 3 provide validation of AGECAT computer diagnosis against outcome; organic and depression diagnoses are seen to have important implications for prognosis.

In 1982-1983 a random sample of 1486 people aged 65 years and above was generated from general practitioner lists; 1070 were interviewed in the community using the Geriatric Mental State and a Social History questionnaire. The cohort was followed up by interview 3 years later. At year 3 the diagnostic computer program AGECAT diagnosed 44 incident cases of depression. Information from the depressed group's initial and further interviews was compared with a control group (which excluded cases of affective or organic mental illness). Univariate analysis yielded three factors that were significantly associated with the development of depression 3 years later: a lack of satisfaction with life; feelings of loneliness; and smoking. Multivariate analysis confirmed their independent effects and revealed 2 further factors attaining significance: female gender and a trigger factor, bereavement of a close figure within 6 months of the third-year diagnosis. Some other factors traditionally associated with depression, such as poor housing, marital status and living alone, failed to attain significance as risk factors.

A random community sample of subjects aged 65 and over was re-interviewed after three years by psychiatrists using the GMS and HAS. The relationship between drinking history and current psychiatric morbidity was examined. Men with a history of heavy drinking for five years or more at some time in their lives were found to have a greater than fivefold risk of suffering from a psychiatric disorder at the time of the interview. Among this group past alcohol consumption was significantly higher for those with a current psychiatric diagnosis compared with those who were well. This association between heavy alcohol consumption in earlier years and psychiatric morbidity in later life is not explained by current drinking habits.

SynopsisThree studies are reported using the Geriatric Mental State in one of its community forms on a total of 647 subjects aged over 65 and living in their own homes. The concordance between the computer diagnosis AGECAT and psychiatrists' diagnosis is at least as good in these replication samples as in the original studies (Copeland et al. 1986). AGECAT has been shown to be useful for epidemiological surveys and as a diagnostic guide for non-medical interviewers.