STUDY 3.7.1 Summary of methods
A study missed by the search strategy and pointed out by one of the reviewers is described below.87
Subjects were randomly drawn from four Manchester general-practice lists. They were initially contacted by letter, phone, or home visits. All available and consenting subjects were administered the GHQ-12. All those scoring 3 or more on the GHQ-12 and one in four of those scoring less than 3 were selected for a second-stage interview with Schedules for Clinical Assessment in Neuropsychiatry (SCAN)88 by a trained psychiatrist. They were all provided with an ICD-10 diagnosis. Those taking part in the second-stage interview were also given a short explanatory model questionnaire. Ethnicity was de®ned using the 1991 census categories. Although the paper does not make this explicit, it implies that these were self-assigned.