THE RELIABILITY OF SEVERELY MENTALLY-ILL PATIENTS REPORT OF TREATMENT SATISFACTION

Patient satisfaction has been identified as an important variable in the evaluation and development of programs for severely mentally ill adults. Measuring this variable assumes, however, that patients can reliably report satisfaction with components of treatment programs. Cognitive deficits characteristic of this population could diminish reliability of satisfaction report. To test this issue, the Patient Satisfaction Interview (PSI) was administered at three epochs during six months of treatment to 30 patients participating in a partial hospitalization program; at each epoch, the test was administered twice to determine its test-retest reliability. Test-retest reliability assessed as raw agreement was found to correlate significantly wth levels of psychotic symptoms and intelligence. Level of symptoms was determined at each epoch by completion of the Brief Psychiatric Rating Scale and verbal intelligence was estimated from performance on the vocabulary subscale of the WAIS-R. Pattern analyses showed that earlier items on the PSI elicited more reports of dissatisfaction with treatment than later items. These results suggest that subject and test factors may interact with reliability of patient reports of satisfaction.