A two-stage mental illness survey of a random sample of persons aged 17 years and over from a rural community in Cantabria, Spain, is described. In the first stage newly qualified doctors and final year medical students interviewed 1223 respondents (583 males and 640 females) at their homes, using the General Health Questionnaire (GHQ-60) and other psychopathological and social questionnaires. In the second stage senior psychiatrists carried out an at-home interview on a sample composed of all those who in the first stage scored above the cut-off point on the GHQ, and of a similar number of persons selected at random from two independent batches of below-threshold scorers on the GHQ. Because of this design the prevalence figures have to be weighted in order to represent the whole first stage sample.

Of the total population, 14·7% (8·1% of the men and 20·6% of the women) had psychiatric disorders as defined by the PSE-ID system. In males depression accounted for about twice as many cases as anxiety states, but in females there was a predominance of a combination of anxiety, phobic and obsessive conditions. Men presented a higher prevalence of disorders over the age of 35, with a peak around the age of forty, while in women the rise of prevalence was over the age of 45. There was, however, no significant association with marital status.

Unemployment was related to mental illness in males but not in females, while the reverse was true of the type of work. In both sexes the presence of children under fourteen in the household was not related to a rise in prevalence. Women exhibited a high rate of mental illness in the low educational level and in the low social and religious integration groups, but in men a rise in prevalence was found in the low social status, low educational level and low social integration groups. Lastly, in both sexes the presence of physical illness was related to mental disorders.

A two-stage screening strategy was applied in a single-handed general practice in northern Italy. The study was designed so that all the second-stage psychiatric interviews were conducted during the first phase of the study (group A patients), while self-report questionnaire (GHQ) and GP assessments only were collected during the second phase (group B patients).

The estimated true case rate of psychiatric disorder (36%) did not differ between the two phases of the study, but there were marked differences in the general practitioner's diagnostic behaviour. During the first phase (when psychiatric interviews were being conducted in the practice), her behaviour was heavily biased towards diagnosing psychiatric disorder; 61% of the patients were regarded by her as psychiatric cases and the hidden psychiatric morbidity rate was only 9%. During the second phase of the study (no psychiatric interviews in the practice), only 37% of the patients were regarded as psychiatric cases by the GP, and the hidden morbidity rate was 59%.

The results are interpreted as demonstrating that the presence of a psychiatrist conducting a two-stage screening survey in the practice had a profound effect on the diagnostic threshold. They also provide support for the use of the GHQ as a case-finding instrument in general practice.

Jockeys based in racing stables in the Newmarket and Epsom areas of England were screened using the 26-item Eating Attitudes Test. The mean score of the EAT was 14·9, significantly above that reported for males in other studies. The response rate was poor, but 10 jockeys agreed to a full psychiatric and eating interview. The majority of jockeys interviewed reported food avoidance, the use of saunas, and the abuse of laxatives. Diuretics and appetite suppressants were also used. Bingeing was common, but self-induced vomiting was unusual. Current weights were 13% below, and the lowest reported weights 21% below, matched population mean weights. The relationship between these forms of eating behaviour and clinical eating disorders is discussed.