The results of a 19-month follow-up study of persons identified as being at risk for bipolar affective disorder by the General Behavior Inventory (GBI) are reported Forty-five GBI cases and 64 noncase control subjects completed brief inventories reflecting subjects' psychological impairment over the follow-up period Overall, GBI cases continued to exhibit significant impairment relative to control subjects Two cases had been hospitalized for major depressive episodes following suicide attempts In addition, significantly higher proportions of GBI cases (a) were referred for treatment by family or friends, (b) reported frequent suicidal ideation, and (c) frequently considered seeking treatment Recently, we introduced a behavioral high-risk paradigm for the identification of persons at nsk for bipolar affective disorder (Depue et al., 1981) The paradigm focuses on individuals exhibiting subsyndromal bipolar disorder, or cyclothymia-a group which is at increased nsk for developing full syndromal bipolar illness (Akiskal, Djenderedjian, Rosenthal, & Kham, 1977). To identify cyclothymics in the general population, we developed the General Behavior Inventory (GBI), a first-stage case-identification inventory for cyclothymia In a previous article, we presented converging evidence for the validity of the GBI from diagnostic-interview, informant-report, family-history, chnical-characterlstics, and daily behavior-rating studies (Depue et al., 1981). In this article, we present evidence that individuals identified as cases by the inventory continue to exhibit significant impairment over the course of a 19-month follow up Method Subjects Subjects consisted of the 126 nonclmical (undergraduate) subjects who participated in the Depue et al. (1981) studies. A cutoff of 27 on the GBI, which maximized concordance between the mven-