abstract
- A self-report symptom inventory, the Hopkins Symptom Checklist, was used as a screening test for psychiatric disorder in a group of 82 new patients in a university hospital outpatient medical clinic, and the results were compared with interviewer diagnoses. The prevalence of psychiatric disorder in the group was high (83%). Both parametric (discriminant function analysis) and nonparametric (contingency table) methods produced screening results from the patient self-ratings that were statistically significant but of limited accuracy in separating psychiatrically ill from well patients. Comparison of patient and interviewer ratings of symptoms indicated substantial agreement, suggesting that the screening accuracy of the symptom inventory is limited by the absence of historical and observational data.