Prediction of patient-reported outcomes after proximal humerus fractures in elderly patients does not appear to be a credible option

The aim of this study was to investigate which patient- and fracture-related factors predict functional outcome at two years in patients aged 60 years or older with a proximal humerus fracture, and to assess the possible evidence in favour of treatment effect heterogeneity in these patients. This was a secondary analysis of a prior randomized controlled trial. We included patients from a Nordic multicentre trial including two trial cohorts (two-part and multipart): 88 patients with a two-part fracture (nonoperative vs locking plate), and 160 patients with a multipart fracture (nonoperative vs locking plate vs hemiarthroplasty) recruited between February 2011 and December 2019. The outcomes were Disabilities of Arm, Shoulder and Hand (DASH) questionnaire score, Oxford Shoulder Score (OSS), and Constant-Murley score (CS) at two years. Ordinary least squares regression was used to estimate 24 months of functional outcomes. We used the methodology described in the Predictive Approaches to Treatment effect Heterogeneity (PATH) statement for the assessment of treatment effect heterogeneity. A total of 248 patients were included in the trials. Dizziness frequency was the most important predictor for DASH and OSS at 24 months. Dominance of the injured arm was the most important predictor for Constant score. R2, describing the variation explained by the baseline variables, showed moderate model fit for all outcomes, i.e. DASH (0.395), OSS (0.363), and CS (0.387). When operative treatments were compared with nonoperative treatments, the interactions between predicted outcome and treatment assignment were small and had high uncertainty, implying a lack of heterogeneous treatment effect. We found that dizziness at baseline is the strongest predictor for DASH and OSS at 24 months. Our results showed that patient selection, based on the covariates available to us, does not seem to be a credible possibility in older patients with proximal humerus fractures, and thus nonoperative treatment should remain the gold standard. ...