PROMIS depression measures perform similarly to legacy measures relative to a structured diagnostic interview for depression in cancer patients

Purpose To assess the convergent validity of the Patient-Reported Outcomes Measurement Information System (PROMIS) depression measures relative to legacy measures and criterion validity against a structured diagnostic interview for depression in an oncology sample. Methods 132 oncology/haematology outpatients completed the PROMIS Depression Computer Adaptive Test (PROMIS-D-CAT) and PROMIS Depression Short Form (PROMIS-D-SF) along with seven legacy measures: Beck Depression Inventory (BDI); Centre for Epidemiological Studies Depression (CES-D); Depression, Anxiety and Stress Scale; Hospital Anxiety and Depression Scale; Patient Health Questionnaire; Distress Thermometer and PSYCH-6. Correlations, area under the curve (AUC) and diagnostic accuracy statistics were calculated with Structured Clinical Interview as the gold standard. Results Both PROMIS measures correlated with all legacy measures at p0.800) were comparable. At the cut-off points for mild depression of 53, the PROMIS measures had sensitivity (0.83 for PROMIS-D-CAT and 0.80 for PROMIS-D-SF) similar to or better than 6/7 legacy measures with high negative predictive value (>90%). At cut-off points of 60 for moderate depression, PROMIS measures had specificity >90%, similar to or better than all legacy measures and positive predictive value ≥0.50 (similar to 5/7 legacy measures). Conclusions The convergent and criterion validity of the PROMIS depression measures in cancer populations was confirmed, although the optimal cut-off points are not established. PROMIS measures were briefer than BDI-II and CES-D but do not offer any advance in terms of diagnostic accuracy, reduced response burden or cost over other legacy measures of depression in oncology patients.