Machine learning–based survival models for predicting rehospitalization of older hip fracture patients: a retrospective cohort study

Abstract Purpose To evaluate machine learning–based survival model roles in predicting rehospitalization after hip fractures to improve reduce the burden on the healthcare system. Methods This retrospective cohort study examined 718 patients with hip fractures hospitalized at the Daejeon Eulji Medic...

Full description

Saved in:
Bibliographic Details
Main Authors: Juahn Oh, Minah Park, Yonghan Cha, Jae-Hyun Kim, Seung Hoon Kim
Format: Article
Language:English
Published: BMC 2025-05-01
Series:BMC Musculoskeletal Disorders
Subjects:
Online Access:https://doi.org/10.1186/s12891-025-08710-z
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850114684336734208
author Juahn Oh
Minah Park
Yonghan Cha
Jae-Hyun Kim
Seung Hoon Kim
author_facet Juahn Oh
Minah Park
Yonghan Cha
Jae-Hyun Kim
Seung Hoon Kim
author_sort Juahn Oh
collection DOAJ
description Abstract Purpose To evaluate machine learning–based survival model roles in predicting rehospitalization after hip fractures to improve reduce the burden on the healthcare system. Methods This retrospective cohort study examined 718 patients with hip fractures hospitalized at the Daejeon Eulji Medical Center between January 2020 and June 2022. Demographic and clinical variables, and rehospitalization data were collected at 6 weeks and 3, 6, 12, and 24 months. Cox proportional hazards (CoxPH), random survival forest (RSF), gradient boosting (GB), and fast survival support vector machine (SVM) models were developed. Model performance was assessed using the concordance index (c-index), area under the curve (AUC), and Kaplan–Meier survival curves. Feature importance was analyzed using permutation importance, with the best model selected based on overall performance. Results Hyperparameter tuning optimized the models. The GB model had the highest mean AUC of 0.868, followed by the RSF (0.785), SVM (0.763), and CoxPH (0.736) models. Feature importance analysis highlighted femoral neck T-score, age, body mass index, operation time, compression fracture, and total calcium as significant predictors. Feature selection improved the c-index for the RSF model from 0.742 to 0.874 and CoxPH model from 0.717 to 0.915; the GB and SVM models exhibited a c-index decline post-feature selection. The GB and RSF models predicted lower rehospitalization probabilities than Kaplan–Meier estimates; the CoxPH model’s predictions were closely aligned with the observed data. Conclusions The effect of feature selection on model performance highlights the need for comprehensive variable selection and model evaluation strategies to improve predictive accuracy.
format Article
id doaj-art-1bf464f4d2f34fd483f92eae28ef33c8
institution OA Journals
issn 1471-2474
language English
publishDate 2025-05-01
publisher BMC
record_format Article
series BMC Musculoskeletal Disorders
spelling doaj-art-1bf464f4d2f34fd483f92eae28ef33c82025-08-20T02:36:49ZengBMCBMC Musculoskeletal Disorders1471-24742025-05-0126111410.1186/s12891-025-08710-zMachine learning–based survival models for predicting rehospitalization of older hip fracture patients: a retrospective cohort studyJuahn Oh0Minah Park1Yonghan Cha2Jae-Hyun Kim3Seung Hoon Kim4Eulji University School of MedicineDepartment of Ophthalmology, Soonchunhyang University Hospital Cheonan, Soonchunhyang University College of MedicineDepartment of Orthopedic Surgery, Daejeon Eulji Medical Center, Eulji University School of MedicineInstitute for Digital Life Convergence, Dankook UniversityDepartment of Ophthalmology, Soonchunhyang University Hospital Cheonan, Soonchunhyang University College of MedicineAbstract Purpose To evaluate machine learning–based survival model roles in predicting rehospitalization after hip fractures to improve reduce the burden on the healthcare system. Methods This retrospective cohort study examined 718 patients with hip fractures hospitalized at the Daejeon Eulji Medical Center between January 2020 and June 2022. Demographic and clinical variables, and rehospitalization data were collected at 6 weeks and 3, 6, 12, and 24 months. Cox proportional hazards (CoxPH), random survival forest (RSF), gradient boosting (GB), and fast survival support vector machine (SVM) models were developed. Model performance was assessed using the concordance index (c-index), area under the curve (AUC), and Kaplan–Meier survival curves. Feature importance was analyzed using permutation importance, with the best model selected based on overall performance. Results Hyperparameter tuning optimized the models. The GB model had the highest mean AUC of 0.868, followed by the RSF (0.785), SVM (0.763), and CoxPH (0.736) models. Feature importance analysis highlighted femoral neck T-score, age, body mass index, operation time, compression fracture, and total calcium as significant predictors. Feature selection improved the c-index for the RSF model from 0.742 to 0.874 and CoxPH model from 0.717 to 0.915; the GB and SVM models exhibited a c-index decline post-feature selection. The GB and RSF models predicted lower rehospitalization probabilities than Kaplan–Meier estimates; the CoxPH model’s predictions were closely aligned with the observed data. Conclusions The effect of feature selection on model performance highlights the need for comprehensive variable selection and model evaluation strategies to improve predictive accuracy.https://doi.org/10.1186/s12891-025-08710-zHip fractureRehospitalizationMachine learningSurvival analysisGradient boostingCox proportional hazards
spellingShingle Juahn Oh
Minah Park
Yonghan Cha
Jae-Hyun Kim
Seung Hoon Kim
Machine learning–based survival models for predicting rehospitalization of older hip fracture patients: a retrospective cohort study
BMC Musculoskeletal Disorders
Hip fracture
Rehospitalization
Machine learning
Survival analysis
Gradient boosting
Cox proportional hazards
title Machine learning–based survival models for predicting rehospitalization of older hip fracture patients: a retrospective cohort study
title_full Machine learning–based survival models for predicting rehospitalization of older hip fracture patients: a retrospective cohort study
title_fullStr Machine learning–based survival models for predicting rehospitalization of older hip fracture patients: a retrospective cohort study
title_full_unstemmed Machine learning–based survival models for predicting rehospitalization of older hip fracture patients: a retrospective cohort study
title_short Machine learning–based survival models for predicting rehospitalization of older hip fracture patients: a retrospective cohort study
title_sort machine learning based survival models for predicting rehospitalization of older hip fracture patients a retrospective cohort study
topic Hip fracture
Rehospitalization
Machine learning
Survival analysis
Gradient boosting
Cox proportional hazards
url https://doi.org/10.1186/s12891-025-08710-z
work_keys_str_mv AT juahnoh machinelearningbasedsurvivalmodelsforpredictingrehospitalizationofolderhipfracturepatientsaretrospectivecohortstudy
AT minahpark machinelearningbasedsurvivalmodelsforpredictingrehospitalizationofolderhipfracturepatientsaretrospectivecohortstudy
AT yonghancha machinelearningbasedsurvivalmodelsforpredictingrehospitalizationofolderhipfracturepatientsaretrospectivecohortstudy
AT jaehyunkim machinelearningbasedsurvivalmodelsforpredictingrehospitalizationofolderhipfracturepatientsaretrospectivecohortstudy
AT seunghoonkim machinelearningbasedsurvivalmodelsforpredictingrehospitalizationofolderhipfracturepatientsaretrospectivecohortstudy