Interpreting profiling data in behavioral health care for a continuous quality improvement cycle

OUTCOME MEASUREMENT SYSTEM: PsychSentinel, a symptom reduction measure, uses 20 diagnostically defined symptom checklists derived from the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV). Symptoms are enumerated and are assigned weights on the basis of clinical significance, providing an overall assessment of symptom intensity. The availability of multisite benchmark norms makes possible the computation of observed-to-expected ratios.

Six examples, drawn from the experience of a number of behavioral health care programs since 1994, illustrate how outcome data can be used to guide and test changes that will effect improvements over current practices. Example 1: Problem identification is one of the most obvious and immediate applications of outcome data relative to a quality improvement process. Data were presented at a meeting of the hospital medical staff; the data showed that one clinician had significantly poorer outcomes in treating bipolar patients. A review of the medical records for bipolar patients treated by this clinician indicated that this clinician was changing medications too rapidly, a problem that was quickly and easily corrected-with improved outcomes. Example 6: Data revealed that patients who were treated in accordance with the critical pathway showed a greater degree of improvement, even though these patients entered treatment with a 10% greater level of symptom intensity.

Each example provides a sample of variability in outcomes and therefore an opportunity to study the reasons for the variability and institute changes.