Grant: Catalyst Grant: Digital Health

When a patient arrives in the Emergency Department (ED) with an extremity fracture (e.g., arm, leg, wrist), doctors must treat the pain effectively. While simple pain relievers work for some, many fracture patients require opioid prescriptions upon discharge. However, the ongoing opioid crisis has led ED doctors to be more cautious, sometimes resulting in undertreated pain that delays recovery and mobility. The risk of an opioid prescription leading to Opioid Use Disorder (OUD) or long-term opioid use varies greatly depending on the individual patient and the prescription details. This creates a "prescriber's dilemma": how to provide adequate pain relief without putting the patient at unnecessary risk. This project has two parallel aims. First, we will use Machine Learning (ML) to build a risk prediction model using Alberta's electronic health records, combining patient history, demographics, and socio-economic factors to predict an individual's risk of OUD or prolonged opioid use after an ED fracture visit. We will rigorously test the model's fairness across groups defined by age, sex, gender, rurality, and socio-economic status to ensure it does not perpetuate health inequities. Second, we will conduct interviews with patients and ED physicians to understand how they think about opioid risk, what information they need for shared decision-making, and what barriers exist to having these conversations in a busy ED. Patient partners with lived experience will collaborate throughout-helping design the study, review the model for potential biases, co-conduct interviews, and shape how findings are shared. Together, these two streams will provide the foundation for a future clinical decision support tool embedded in the electronic medical record, helping doctors and patients make safer, more informed prescribing decisions at the bedside. The long-term impact will be more equitable, patient-centered pain management for fracture patients across urban and rural Alberta.