The 1990s have seen major economic growth and declines in unemployment. We focus here on how individuals with probable risk for mental illness are faring in this robust economy, using data from Health Care for Communities (HCC), a national survey conducted from the fall of 1997 through the end of 1998 (1).

We examine two labor force participation measures. The first, percentage out of the labor force, is the percentage of the total population not employed or seeking a job for reasons such as retirement or disability. The second, unemployment, is the fraction of persons in the total labor force who are without current employment and are looking for a job. In early 1998, the labor force totaled 137 million, of whom about 6.3 million were unemployed.

Measures derived from the Composite Diagnostic Interview were used to assess the likely presence of major depressive disorder, dysthymic disorder, generalized anxiety disorder, panic disorder, and bipolar disorder and to find out whether the respondent ever had an overnight stay for psychotic symptoms or had received a past diagnosis of schizophrenia.

A total of 9,585 persons participated in the HCC survey, of whom 1,876 reported a past or present mental disorder. Responses from those reporting a mental disorder were categorized into three groups: any disorder, depressive or dysthymic disorder within the past year, and bipolar or psychotic disorder at any time.

F1 shows that unemployment rates are three to five times higher among persons with mental disorders than among those with no disorder (t=11.5, df=6,860, p<.01). The higher unemployment rate holds when the data for men and women are considered separately and is not an artifact of a higher percentage of women with a mental disorder. A higher percentage of persons with any mental disorder are out of the labor force compared with persons with no mental disorder.

F2 tracks the probability that a previously employed person with health insurance became unemployed or left the labor force between the HCC survey and its parent survey, the Community Tracking Study (1), conducted 15 to 18 months before. It shows that individuals at risk of mental disorders are significantly more likely than others to leave a job that provided insurance (t=4.52, df=4,142, p<.001 for unemployment; t=3.11, df= 4,321, p<.01 for out of labor force).

The authors are affiliated with Rand, 1700 Main Street, Santa Monica, California 90403 (e-mail, Roland_Sturm@rand.org). Harold Alan Pincus, M.D., and Terri L. Tanielian, M.A., are coeditors of this column.