Development of informant-based screening tests for dementia is an emerging field. The reliability and validity of the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE), an instrument that screens for dementia in general, have been established. We conducted a study to validate a French version of the IQCODE as a screening test for Alzheimer's dementia in the elderly living in the community.

In the Canadian Study of Health and Aging, subjects were screened in their own homes using the Modified Mini-Mental State Exam (3-MS). Those who screened positive, along with a sample of subjects who screened negative, were referred for a complete clinical examination. In Quebec, an informant was asked to complete the French version, IQCODE (F), at that time. Based on the final clinical diagnoses, performances of the IQCODE (F) and Mini-Mental State Examination (MMSE, converted from 3-MS) in screening for Alzheimer's disease were evaluated.

Of the 237 subjects, the mean IQCODE (F) score was 3.4 (s.d. = 0.6), on a 5-point scale (1 = improvement in condition over the past 10 years, 5 = marked deterioration, 3 = no change). The mean MMSE score was 23.1 (s.d. = 4.5). The scores on the two scales were correlated (r = –0.44, P < 0.001). The IQCODE (F) scores were unrelated to education (r = –0.07, P > 0.3) in contrast to the MMSE scores (r = 0.28, P < 0.001). With respect to a diagnosis of probable Alzheimer's disease, the IQCODE (F) (cut-off point 3.6) had a sensitivity of 75% and a specificity of 95.6%. The sensitivity and specificity of the MMSE (cut-off point 23) were 70% and 82.3% respectively.

The findings of the IQCODE (F) are consistent with those of the English version in correlation with the MMSE and apparent freedom from educational bias. The IQCODE is superior to the MMSE as a screening test for probable Alzheimer's disease in the elderly living in the community. It may be a useful addition to the screening tests already available, especially for the less well educated.

The paper by McGuigan in this issue of the BJP examines the frequency of errors in statistical methods presented in papers published in the BJP in 1993. In that year, 248 papers appeared, 164 contained numerical results, and 65 of these contained statistical errors. So, one might infer, some 26% of the papers in the BJP draw potentially doubtful conclusions. Moreover, the level of misunderstanding implied by some of the examples McGuigan cites is frightening.

Determining the prevalence of psychiatric disorders among remand populations has been made a priority in England and Wales. Differences in legal process and psychiatric services in Scotland make similar research there important.

Demographic data were collected on 389 prisoners, the Clinical Interview Schedule was completed and cognitive function assessed.

The prevalence of major psychiatric disorders was low. Less severe symptoms were more common. The sample was of average IQ, but low educational attainment. Reported drug abuse was high.

Few of those interviewed required hospital care, but other symptoms and drug-related problems may place heavy demands on prison medical and psychiatric services.