Do Preoperative PROMIS Scores Predict Postoperative Outcomes in Hallux Rigidus Patients Undergoing Cheilectomy

CATEGORY: Midfoot/Forefoot; Other INTRODUCTION/PURPOSE: Prior studies have shown that preoperative patient reported outcome scores, specifically the Patient- Reported Outcomes Measurement Information System (PROMIS) score, predict improvement in foot and ankle surgery. This has been shown in hallux valgus, progressive collapsing foot deformity (PCFD), and end-stage ankle arthritis. However, no studies have investigated whether preoperative PROMIS scores are predictive of postoperative outcomes following a cheilectomy for hallux rigidus. The purpose of this study was to determine if preoperative PROMIS scores were predictive of which patients undergoing cheilectomy for hallux rigidus would achieve the minimal clinically important difference (MCID). Additionally, we sought to determine if preoperative PROMIS thresholds could be estimated to help guide surgical decision-making. METHODS: This retrospective cohort study included patients > 18 years, who underwent cheilectomy with or without a Moberg osteotomy for treatment of hallux rigidus and had preoperative and ≥2-year postoperative PROMIS physical function scores. Preoperative to 2-year postoperative change in PROMIS scores in the physical function, pain interference, pain intensity, and depression domains were recorded, and receiver operating curve (ROC) analyses were performed to determine if preoperative scores were predictive of achieving the MCID for each domain. Area under the curve (AUC) was used to determine if preoperative PROMIS scores were predictive of achieving the MCID for each domain. 95% sensitivity and specificity thresholds were determined for statistically significant AUCs (AUC0.70). Positive and negative likelihood ratios (LRs) were calculated for the specificity and sensitivity thresholds, respectively. For statistically significant LRs, posttest probabilities were calculated for meeting MCID if patients met the preoperative PROMIS cutoff. RESULTS: Statistically significant AUCs were found for PROMIS physical function (AUC=0.71), pain intensity (AUC=0.70) and depression (AUC=0.79) when using MCIDs defined as ½ of the standard deviation of the preoperative to 2-year postoperative change (Table 1). The AUC for pain interference was not statistically significant (AUC=0.69). The preoperative 95% sensitivity threshold was 53.2 for the physical function domain, the negative LR was 1.36, and the posttest probability of achieving the MCID for physical function with a preoperative t-score greater than 53.2 was 63%. However, the LRs were not statistically significant for the physical function specificity threshold and the pain intensity sensitivity and specificity thresholds, so posttest probabilities could not be calculated. CONCLUSION: Our results suggest that preoperative PROMIS physical function, pain interference, and pain intensity scores had limited predictive value in determining of which patients would achieve meaningful clinical improvement following a cheilectomy. Patients with worse functional limitations were not necessarily more likely to clinically improve after a cheilectomy. The pain interference domain was not predictive of meeting the MCID postoperatively. Additionally, the calculated thresholds for pain intensity, and depression were not reliable predictors of clinical improvement. This is in contrast to prior studies that have found useful procedure-specific thresholds such for bunionectomy, flatfoot reconstruction, and total ankle replacement.