Status: Active

Older patients with fractures face a high hospital readmission risk, which can lead to additional health complications and increase costs for both patients and healthcare systems. To inform preventative care, we studied readmission risk using data from 1,202 Duke patients in Durham and surrounding areas who experienced upper extremity fractures at age 50 or above. Our goals are to identify key risk factors, uncover geographic pattern of readmission, and highlight high-risk regions where targeted interventions could have the greatest impact.

We developed a Bayesian competing risk survival model with spatially varying coefficients. This approach uses patient-level location information to capture fine-scale spatial patterns in readmission risk while adjusting for unmeasured neighborhood factors. Importantly, our model examines how the relationship between comorbidity burden and readmission risk changes across the region. By incorporating spatial random effects, we can detect clusters of high readmission risk and provide uncertainty measures to guide clinical and policy decisions. To our knowledge, this is the first point-level spatial competing risks model in this setting, as prior work has primarily relied on aggregated, areal-level methods.