Comparing PROMIS computer-adaptive tests to the Brief Symptom Inventory in patients with prostate cancer.

ABSTRACT

PURPOSE: This study assessed whether the Patient-Reported Outcomes Measurement Information System (PROMIS) computer-adaptive tests (CATs) provided results similar to those of the Brief Symptom Inventory (BSI) with a low patient burden. METHODS: Secondary data analysis of 136 prostate cancer patients who completed the 53-item BSI and the PROMIS CATs assessing depression, anxiety, and hostility. RESULTS: The PROMIS CATs and BSI correlated significantly in measures of depression (.85), anxiety (.76), and anger/hostility (.66; p < .001 for all). Using our BSI cutoff points for depression, anxiety, and anger/hostility, ROC analysis yielded areas under the curve of .966 [standard error (SE) = .014, p < .001], .975 (SE = .012, p < .001), and .952 (SE = .027, p < .001), respectively. CONCLUSIONS: PROMIS CATs were highly correlated with the BSI subscales, indicating that the CATs performed well compared with the BSI, a widely used psychosocial measure.