To develop a robust hip fracture risk prediction model for elderly patients that accommodates incomplete clinical data, thereby enhancing clinical decision-making.
Key Findings:
Global model accuracy: 84.67%, AUC: 0.8064, indicating a solid baseline for prediction.
Key predictors: age, sex, BMD, cholesterol, with section modulus of BMD being significant, suggesting areas for targeted intervention.
MMPro-HIP model accuracy: 90.94%, AUC: 0.9423, outperforming the global model, highlighting its potential for clinical application.
Interpretation:
The MMPro-HIP model effectively predicts hip fracture risk in older adults, even with incomplete data, emphasizing the critical role of BMD and demographic factors in clinical assessments.
Limitations:
Single-center study limits generalizability, as findings may not be applicable to broader populations.
Need for external validation of the model to confirm its reliability across different settings.
Conclusion:
The MMPro-HIP model presents a promising approach for hip fracture risk assessment in elderly patients, particularly in settings with incomplete data.
