Development of a Mortality Prediction Score for Patients With Hip Fractures

Abstract

Purpose: This study developed and validated a context-specific Hip Fracture Mortality Score (HFMS) to predict 1-year mortality in Thai patients. Methods: A retrospective cohort analysis evaluated consecutive hip fracture admissions at Nan Hospital (January 2019–December 2023), including 1,263 patients aged ≥60 years. Twenty-four candidate predictors were analyzed using univariate logistic regression; variables with p < 0.20 entered multivariate analysis. Beta coefficients were transformed into point values (rounded to the nearest 0.5) for clinical implementation. Results: The overall 1-year mortality rate was 20.3%. Five independent predictors were identified: end-stage renal disease (adjusted odds ratio [aOR] 4.64, 95% confidence interval [CI] 2.097–10.263; 4 points), non-operative management (aOR 2.64, 95% CI 1.667–4.192; 2.5 points), anemia (aOR 2.02, 95% CI 1.451–2.807; 2 points), male sex (aOR 1.80, 95% CI 1.289–2.504; 1.5 points), and walking with aids (aOR 1.43, 95% CI 1.049–1.941; 1 point). The HFMS showed moderate discrimination (area under the curve [AUC] 0.701, 95% CI 0.666–0.737; optimism-corrected AUC 0.700 on bootstrap validation). A 2.5-point threshold (Youden’s index: 0.296) yielded a sensitivity of 56.3%, specificity of 73.4%, positive predictive value of 35.0%, and negative predictive value (NPV) of 86.8%. Low-risk patients (<2.5 points) had mortality rates of 7.97–20% versus 27.84–53.85% in high-risk patients (≥2.5 points). Conclusions: The HFMS is a practical prognostic tool using five ascertainable parameters, avoiding complex algorithms. It addresses the limitations of Western-derived scoring systems in resource-constrained Thai healthcare settings. Despite moderate sensitivity, a high NPV effectively identifies low-risk patients. External validation and prospective evaluation warrant further investigation.