Analysis of non-response in a prospective study of depression in Los Angeles County

The occurrence of depression in a probability sample of 1003 Los Angeles County adults was measured in a 1979-80 prospective cohort study. The primary measure of depression was the 20-item Center for Epidemiologic Studies-Depression (CES-D) scale. Of the original 1003 adults, 893 were successfully re-interviewed at a second time period (reasons for not being interviewed originally and at the second time period are given). Demographic data on those not re-interviewed are contrasted with data from those who did respond. The mean depression level and percentage classified as depressed at the first interview were higher in non-respondents than in respondents to the second interview. Techniques are given for assessing whether discontinued participation is due to exposure or disease factors. For this study, it was shown that dropping out of the study was mainly due to exposure (or demographic) factors and not to depression as such. Two methods for estimating the level of disease at the second time period for respondents and dropouts are illustrated.