BACKGROUND: Although effective treatment of depressed patients requires regular follow-up contacts and symptom monitoring, an efficient method for assessing treatment outcome is lacking. We investigated responsiveness to treatment, reproducibility, and minimal clinically important difference of the Patient Health Questionnaire-9 (PHQ-9), a standard instrument for diagnosing depression in primary care. METHODS: This study included 434 intervention subjects from the IMPACT study, a multisite treatment trial of late-life depression (63% female, mean age 71 years). Changes in PHQ-9 scores over the course of time were evaluated with respect to change scores of the SCL-20 depression scale as well as 2 independent structured diagnostic interviews for depression during a 6-month period. Test-retest reliability and minimal clinically important difference were assessed in 2 subgroups of patients who completed the PHQ-9 twice exactly 7 days apart. RESULTS: The PHQ-9 responsiveness as measured by effect size was significantly greater than the SCL-20 at 3 months (-1.3 versus -0.9) and equivalent at 6 months (-1.3 versus -1.2). With respect to structured diagnostic interviews, both the PHQ-9 and the SCL-20 change scores accurately discriminated patients with persistent major depression, partial remission, and full remission. Test-retest reliability of the PHQ-9 was excellent, and its minimal clinically important difference for individual change, estimated as 2 standard errors of measurement, was 5 points on the 0 to 27 point PHQ-9 scale. CONCLUSIONS: Well-validated as a diagnostic measure, the PHQ-9 has now proven to be a responsive and reliable measure of depression treatment outcomes. Its responsiveness to treatment coupled with its brevity makes the PHQ-9 an attractive tool for gauging response to treatment in individual patient care as well as in clinical research.
