The General Health Questionnaire (GHQ) is a widely used method of assessing psychiatric distress in clinical settings and in the community. The authors administered a 20-item interview version of the GHQ and a battery of survey items on use of ambulatory health care services to 3,389 respondents in eastern Baltimore as part of the Epidemiologic Catchment Area survey. In addition, a subsample of these respondents (n = 810) was examined by research psychiatrists who provided a reference mental disorder diagnosis for comparison with GHQ results. Psychiatrists determined that 102 respondents had American Psychiatric Association's Diagnostic and Statistical Manual, third edition, (DSM-III) disorders suitable for detection by the GHQ. These respondents were categorized by the health service use they reported over the previous 6 months. A comparison indicated that 82% of the diagnosed cases with recent specialty care also had elevated GHQ scores. Among cases with recent general medical care, 52% had elevated GHQ scores. Among diagnosed cases reporting no recent ambulatory health care, only 38% had elevated GHQ scores. Receiver operating characteristic analysis demonstrated a similar trend. Cases of depression were separated into less severe cases (adjustment disorder with depressed mood) and more severe cases (major depression). There were relatively more cases of severe depression than cases of mild depression in those respondents reporting recent health service use compared with those respondents reporting no recent service use. This may account for the higher sensitivity of the GHQ in detecting psychiatric disorders for those persons utilizing health services compared with individuals using no health services.(ABSTRACT TRUNCATED AT 250 WORDS)