Revolutionizing patient care: Development and external validation of the "Rotterdam hip fracture mortality prediction – 30 days" (RHMP-30) – A comprehensive multidimensional approach to prognosticating 30-day mortality in hip fracture patients

Aim: The aim of this study is to develop an accurate and clinically relevant prediction tool for 30-day mortality after hip fracture surgery. The prediction tool is based on a detailed and careful collected clinical dataset yielding nuanced insights into patient mortality rates.

Patients and methods A prospective hip fracture database of two hospitals was used to obtain data between 2011 and 2021. The prediction model is developed by an adaptive Least Absolute Shrinkage and Selection Operator (LASSO) with the data of the first hospital. The cohort of the second hospital was used to validate the prediction model externally by performing a Receiver Operating Characteristics curve and a calibration plot.

Results: In total, 3523 patients were analyzed of which 302 patients (8.6%) died within 30 days after surgery. The prediction model was developed based on the cohort of the first hospital. After the adaptive LASSO 7 over 26 variables were included in the prediction model: age, gender, ASA 4, dementia, albumin level, KATZ-ADL-score and living in a nursing home. The Area Under the Curve (AUC) of the prediction model in the first cohort was 0.789. The external validation of the prediction model was performed in the second hospital with an AUC of 0.775.

Conclusion: The "Rotterdam Hip Fracture Mortality Prediction – 30 days " (RHMP-30) is developed and externally validated in two up to date, extensive hip fracture cohorts from two different hospitals. The RHMP-30 can support in clinical decision making and informed consent for the most fragile patients in the hip fracture population.

No