Psychometric properties of NIH PROMIS® instruments in bariatric surgery candidates

Objective: Ascertain the psychometric properties of the Patient-Reported Outcomes Measurement Information System® (PROMIS®) Depression (PROMIS-D®) and Anxiety (PROMIS-A®) Short Forms in a sample of presurgical bariatric patients. Method: This retrospective chart review included 259 patients who completed a presurgical psychological evaluation comprising a semistructured interview and self-report measures. Other data used included demographics, psychiatric diagnoses, and initial psychological clearance for surgery (yes/no). Reliability of the PROMIS-D® and PROMIS-A® instruments was assessed using Cronbach's alpha, interitem correlations, and item-total correlations. Convergent, predictive, and concurrent validity were assessed using various methods (correlations, logistic regressions, and generalized linear models) to determine whether PROMIS® instruments were correlated with similar measures, predicted surgical clearance and psychiatric diagnosis, and whether psychiatric diagnosis status was associated with significantly different PROMIS® scores. Categorical confirmatory factor analyses assessed the factor structure, and multiple-indicator multiple cause models assessed invariance. Results: Patients were predominantly female (78%), white (97%) and middle-aged (M 43.49, SD 11.26). Rates of diagnosed anxiety and depression-related disorders were high (17% and 23%). At least 65% of respondents reported "never" or "rarely" experiencing symptoms across both measures, and the mean T scores for the PROMIS-A® and PROMIS-D® were 48.31 (SD 8.99) and 48.64 (SD 9.49), respectively. Both measures demonstrated good psychometric properties and were essentially unidimensional. Conclusions: Analyses supported using both PROMIS® measures. Additionally, two subgroups were identified: those who reported virtually no symptoms of anxiety or depression, likely because they want surgical approval, and those not receiving initial psychological clearance for surgery based on PROMIS® self-reported scores indicating distress and clinician-diagnosed psychopathology.