Factors Associated with a Recommendation for Operative Treatment for Fracture of the Distal Radius.

Background  Evidence suggests that there is substantial and unexplained surgeon-to-surgeon variation in recommendation of operative treatment for fractures of the distal radius. We studied (1) what factors are associated with recommendation for operative treatment of a fracture of the distal radius and (2) which factors are rated as the most influential on recommendation of operative treatment. Methods  One-hundred thirty-one upper extremity and fracture surgeons evaluated 20 fictitious patient scenarios with randomly assigned factors (e.g., personal, clinical, and radiologic factors) for patients with a fracture of the distal radius. They addressed the following questions: (1) Do you recommend operative treatment for this patient (yes/no)? We determined the influence of each factor on this recommendation using random forest algorithms. Also, participants rated the influence of each factor-excluding age and sex- on a scale from 0 (not at all important) to 10 (extremely important). Results  Random forest algorithms determined that age and angulation were having the most influence on recommendation for operative treatment of a fracture of the distal radius. Angulation on the lateral radiograph and presence or absence of lunate subluxation were rated as having the greatest influence and smoking status and stress levels the lowest influence on advice to patients. Conclusions  The observation that-other than age-personal factors have limited influence on surgeon recommendations for surgery may reflect how surgeon cognitive biases, personal preferences, different perspectives, and incentives may contribute to variations in care. Future research can determine whether decision aids-those that use patient-specific probabilities based on predictive analytics in particular-might help match patient treatment choices to what matters most to them, in part by helping to neutralize the influence of common misconceptions as well as surgeon bias and incentives. Level of Evidence  There is no level of evidence for the study.