ORTHOPOD: Linking ambulatory future trauma injury distribution from fragility proximal femur fracture caseload

•NHFD numbers can potentially be used to predict ankle and distal radius fractures, but further model refinement required for other fracture types.•Model could be used to aid making informed decisions on restricting and re-locating resources in the trauma network, to improve service and patient care.•Managing ankle and distal radius fractures through dedicated day-case pathways, and a more centralised approach for less common fracture types could form part of this restructuring.•Fragility proximal femur, ankle and distal radius fractures dominate trauma burden. The age of those experiencing traumatic injury and requiring surgery increases. The majority of this increase seen in older patients having operations after accidents is in fragility proximal femur fractures (FPFF). This study designed a model to predict the distribution of fractures suitable for ambulatory trauma list provision based on the number of FPFF patients. The study utilized two datasets which both had data from 64 hospitals. One derived from the ORTHOPOD study dataset, and the other from National Hip Fracture Database. The model tested the predictability of 12 common fracture types based on FPFF data from the two datasets, using linear regression and K-fold cross-validation. The predictive model showed some promise. Evaluation of the model with mean RMSE and Std RMSE demonstrated good predictive performance for some fracture types, although the r-squared values showed that large variation in these fracture types was not always captured by the model. The study highlighted the dominance of FPFFs, and the strong correlation between these and numbers of ankle and distal radius fractures at a given unit. It is possible to model the numbers of ankle and distal radius fractures based off the number of patients admitted with hip fractures. This has great significance given the drive for increased day case utilisation and bed pressures across health services. While the model's current predictability was limited, with methodological improvements and additional data, a more robust predictive model could be developed to aid in the restructuring of trauma networks and improvement of patient care and surgical outcomes.