Responsiveness of Patient-Reported Outcome Measures After Shoulder Instability Surgery: A Systematic Review and Meta-Analysis

OBJECTIVES: Patient reported outcomes (PROs) are increasingly used in orthopaedics as a tool to objectively assess subjective data and provide a sense of responsiveness to treatment. Unfortunately, there are several PROs and little data as to which outcome scores are most useful. The purpose of this study was to evaluate the utilization and responsiveness of PROs reported in the literature after shoulder instability surgery. METHODS: We performed a systematic review of the PubMed, SportDiscus, Cochrane, and CINHAHL databases according to PRISMA guidelines to identify studies published in the last 10 years which reported PROs after shoulder instability surgery. Articles were excluded if they were not primary research (case reports and review articles), were published only in abstract form, were not available in English, or did not report pre- and post-operative mean PRO values. The specific PROs utilized, number of patients, mean follow-up time, and preoperative and postoperative means and standard deviations were recorded for each article, For studies including preoperative and postoperative means and standard deviations of two or more PROs, and the comparative responsiveness (CR) of each PRO was assessed (CR was calculated by squaring PRO change score t-value ratio). RESULTS: Abstracts from 112 studies were identified for full text review, and 29 studies ultimately met inclusion criteria. Sixteen different PROs were reported in various combinations in the included studies. Mean follow-up was 25.17 months (SD = 15.01) and mean sample size was 47.34 (SD = 59.06). The majority of studies (72.4%) utilized more than one PRO. The most commonly used PROs were the ASES (13 studies, 44.8%), Rowe (10, 34.5%), WOSI (8, 27.6%), VAS-pain (7, 24.1%), UCLA (7, 24.1%), and Constant (6, 20.7%). The remaining PRO tools appeared in 3 or fewer studies. Responsiveness of the different PRO tools was evaluated with a subset of 4 articles that included sample size, pre- and postoperative means, and pre- and post-operative standard deviations for PROs reported in this subset were the ASES, Rowe, WOSI, VAS-pain, and Constant scores. The Rowe score was much more responsive than both the ASES (RE = 22.8) and Constant scores (RE = 33.17). On the contrary, the VAS-pain was the least responsive, with RE = 0.57 when compared to the ASES, and RE = 0.32 when compared to the WOSI. ASES remained more responsive than the Constant (RE = 1.75), VAS-pain (RE = 1.75), and WOSI (RE = 0.97). CONCLUSION: Despite being less frequently utilized, the Rowe score was considerably more responsive than the ASES and Constant scores. ASES, Constant, and WOSI were similar to each other in terms of responsiveness, and the shoulder-specific scores were more responsive than the VAS-pain score. When assessing patient outcomes related to shoulder instability surgery, surgeons may want to consider employing the more sensitive, and instability-specific, Rowe score rather than other commonly used shoulder PROs.