Loading

Original Research Open Access
Volume 6 | Issue 1 | DOI: https://doi.org/10.33696/Orthopaedics.6.037

Development and Internal Validation of a Clinical Prediction Model for 90-Day Walking Recovery Following Hip Hemiarthroplasty in Elderly Patients with Neck of Femur Fracture in Nigeria

  • 1Consultant Orthopaedic and Trauma Surgeon, Orthopaedic and Trauma department, National Orthopaedic Hospital, Lagos, Nigeria
  • 2Lecturer and Honorary Consultant Psychiatrist, Department of Psychiatry, LAUTECH and LAUTECH Teaching Hospital, Ogbomoso, Oyo State, Nigeria
  • 3Consultant General Surgeon, Faculty of Clinical Sciences, Osun State University, Osogbo, Osun State, Nigeria
  • 4Consultant Orthopaedic Surgeon, Department of Hospital services, Federal Ministry of Health, Abuja, Nigeria
+ Affiliations - Affiliations

Corresponding Author

Kehinde Adesola Alatishe, medistuff1@yahoo.com, Kehinde.alatishe@npmcn.edu.ng

Received Date: April 25, 2026

Accepted Date: June 23, 2026

Abstract

Background: Early mobilization following hip hemiarthroplasty for neck of femur (NOF) fractures is critical to reducing morbidity and mortality in elderly patients. Objective tools to predict short-term walking recovery are lacking.

Objectives: To develop and internally validate a clinical prediction model for 90-day postoperative walking recovery in elderly patients following hip hemiarthroplasty for neck of femur fracture in Nigeria.

Methods: This retrospective cohort study evaluated elderly patients with NOF fractures treated via hemiarthroplasty (January 2018–December 2022). Patients with unstable intertrochanteric or pathologic fractures, incomplete records, or 90-day mortality were excluded. Predictor variables were screened using univariable logistic regression. Factors with p<0.1 entered a multivariable model using backward stepwise elimination (p<0.05 for retention). Model performance was rigorously evaluated via discrimination (Area Under the ROC Curve [AUC]) and calibration (decile plot and Hosmer-Lemeshow test). Internal validation was performed via 1,000 bootstrap resamples and model utility was assessed using Decision Curve Analysis (DCA).

Results: Out of 286 screened cases, 162 patients were included. Four independent predictors of non-ambulatory outcomes were retained: residual limb weakness-LW (OR=8.68, p<0.001), cognitive impairment-Cog (OR=1.90, p=0.009), pre-injury walking status-pW (OR=1.30, p=0.042), and Charlson Comorbidity Index- CCI (OR=1.17, p=0.018). The model demonstrated strong discrimination (AUC=0.79; 95% CI: [0.70, 0.88]; p<0.001) and good calibration (Hosmer-Lemeshow p=0.664). DCA confirmed substantial clinical net benefit across a 15–75% threshold range. The final prognostic formula is: Log-Odds = 0.203 + (0.265 × pW) + (2.161 × LW) + (0.160 × CCI) + (0.644 × Cog).

Conclusion: This internally validated four-item clinical model effectively predicts 90-day walking recovery following hemiarthroplasty in Nigerian geriatric patients. It serves as a practical bedside tool to enhance preoperative risk stratification, counseling, and personalized postoperative rehabilitation care.

Keywords

Clinical prediction model, Ambulation, Hemiarthroplasty, Geriatrics, Neck of femur fracture, Multivariate analysis, Nigeria

Author Information X