AI Predicts One-Year Risk After Hip Surgery in Seniors

The study examined health records from a national Korean insurance plan. It focused on 17,290 individuals who were 65 years old or older and had hip surgery in 2019. The goal was to see what factors most influence the chance of dying within one year after the operation.

Researchers built a predictive model using many possible health indicators. They ranked the importance of each factor with a forest‑based algorithm and then used an explanatory tool to detail each factor's impact. The strongest signals were higher age, needing a blood transfusion, male gender, dementia, living in a low‑income area, having hidden solid tumors, past congestive heart failure, chronic kidney disease, regular statin use, and prior peripheral artery disease. SHAP analysis showed that age had the biggest impact (0.17), followed by blood transfusion (0.10), male gender (0.09), dementia (0.04), low‑income status (0.03), hidden tumors (0.06), past heart failure (0.04), chronic kidney disease (0.05), general anesthesia (0.02), and iron supplementation (0.02).

The model achieved an area‑under‑the‑curve score of about 75%, meaning it correctly identified roughly three‑quarters of the outcomes. Patients who are very old, have several health conditions, or are likely to need a blood transfusion were considered high risk.

Doctors are urged to tailor care for these high‑risk seniors, perhaps by watching them more closely or adjusting treatment plans.