Can Patient-Reported Outcomes Measurement Information System® (PROMIS) measures accurately enhance understanding of acceptable symptoms and functioning in primary care?

• Main Authors: Ryan P. Jacobson, Daniel Kang, Jeff Houck
• Format: Article
• Language: English
• Published: SpringerOpen 2020-05-01
• Series: Journal of Patient-Reported Outcomes
• Subjects: Patient-reported outcomes; Primary care; PROMIS; PASS scores; Reference values
• Online Access: http://link.springer.com/article/10.1186/s41687-020-00206-9

Abstract Background Value-based healthcare models will require prioritization of the patient’s voice in their own care toward better outcomes. The Patient-Reported Outcomes Measurement Information System® (PROMIS) gives patients a voice and leads providers to actionable treatments across a broad range of diagnoses. However, better interpretation of PROMIS measures is needed. The purpose of this study was to evaluate the accuracy of PROMIS Physical Function (PF), Self-Efficacy for Managing Symptoms (SE), Pain Interference (PI), Fatigue, and Depression measures to discriminate patient acceptable symptom state (PASS) in primary care, determining if that accuracy is stable over time and/or retained when PROMIS score thresholds are set at either ½ or 1 SD worse than the reference population mean. Methods Primary care patients completed the five PROMIS measures and answered the PASS yes/no question at intake (n = 360), 3–14 days follow-up (n = 230), and 45–60 days follow-up (n = 227). Thresholds (optimal, ½ SD, and 1 SD worse than reference values) for PROMIS T-scores associated with PASS were determined through receiver-operator curve analysis. Accuracy was calculated at the three time points for each threshold value. Logistic regression analyses were used to determine combinations of PROMIS measures that best predicted PASS. Results PROMIS PF, SE, PI, and Fatigue optimal score thresholds (maximizing sensitivity and specificity) yielded area under the curve values of 0.77–0.85, with accuracies ranging from 71.7% to 79.1%. Accuracy increased minimally (1.9% to 5.5%) from intake to follow-ups. Thresholds of 1 SD worse than the mean for PROMIS PF and PI measures and ½ SD worse for SE and Fatigue overall retained accuracy versus optimal (+ 1.3% to − 3.6%). Regression models retained SE, PI, and Fatigue as independent predictors of PASS, and minimally increased accuracy to 83.1?%. Conclusions This study establishes actionable PROMIS score thresholds that are stable over time and anchored to patient self-reported health status, increasing interpretability of PF, SE, PI, and Fatigue scores. The findings support the use of these PROMIS measures in primary care toward improving provider-patient communication, prioritizing patient concerns, and optimizing clinical decision making.