Application of Machine Learning on Health Examination Data for Predicting the Decrease of Bone Mineral Density - Original Article

Aug 2, 2026

Early detection and preventive action for reduced bone density can improve patients’ quality of life while lowering long-term economic burden. This study developed and compared five machine learning algorithms to predict osteoporosis risk using health examination data from adults aged 40+ collected between January 2022 and January 2024. After screening 17 initial variables to 13, model performance was assessed by accuracy, sensitivity, specificity, and AUROC, showing the RF model as the top performer across genders.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages9

Reading Time23 min

AccessFree to download

Abstract

Introduction

Background and clinical need

Diagnostic approaches and limitations

Risk factors and existing tools

Study aims

Methods

Data acquisition

ORIGINAL ARTICLE # Application of Machine Learning on Health ExaminationData for Predicting the Decrease of Bone Mineral Density Bohan Li 1, Dongjin Wu2, Xiaoqian Kong 1, Yan Shi 1, Chunzheng Gao2*, Yixin Li 1* 1Health Management Center, The Second Hospital of Shandong University, Shandong 250033, China. 2Spine Surgery, The Second Hospital of Shandong University, Shandong 250033, China. * Corresponding Author: Yixin Li, MD, Health Management Center, The Second Hospital of Shandong University, 247 Beiyuan Road, Jinan,Shandong 250033, P.R. China. Email: sdey_hmc@163.com. Chunzheng Gao, MD. Spine Surgery, The Second Hospital of Shandong University, 247 Beiyuan Road, Jinan,Shandong 250033, P.R. China. Email: 15153169697@163.com. ## ABSTRACT Background: Early detection and preventive measuresfor reduced bone density can greatly improve patients'quality of life and reduce economic burdens. This study aimed to develop machine learning algorithms that canaccurately predict the risk of bone mineral density loss. Methods: The study included participants aged 40 yearsand older who underwent health evaluations at an affiliated institution from January 2022 to January 2024. Fivemachine learning algorithms were used to predict the risk of osteoporosis: k-nearest neighbor (KNN), randomforest (RF), support vector machine (SVM), artificial neural network (ANN), and logistic regression (LR) . Theperformances were evaluated based on accuracy, sensitivity, specificity, and area under the receiver operatingcharacteristic curve (AUROC) . Results: This study included 11132 patients, of whom 3568 had decreased bonedensity. The initial dataset contains 17 variables. After the data screening, 13 variables were included in themachine learning model. The AUROC for ANN, KNN, LR, RF, and SVM were 0.882, 0.906, 0.684, 0.918, 0.896for males and 0.881, 0.843, 0.784, 0.922, 0.872 for females, respectively. The accuracies ofANN, KNN, LR, RF,and SVM were 0.83, 0.86, 0.75, 0.88, 0.82 for males, and 0.81, 0.77, 0.74, 0.85, 0.79 for females. Conclusion:In this study, we developed five machine learning models to accurately predict bone density reduction. The RFmodel performed best in both male and female populations, with the highest AUROC. Application of machinelearning models in clinical settings can help improve the prevention, detection, and early treatment of bonedensity reduction. Key Words: Machine Learning, KNN, RF, SVM, ANN, LR, Osteoporosis, Osteopenia, Bone Mineral Density ## INTRODUCTION Osteoporosis is a systemic bone disease thatcommonly occurs with aging and is characterizedby low bone mass and fragile bone structure, whichincreases the risk of fractures.1 Approximately 50%of postmenopausal women and 20% of men over 50 worldwide were affected by osteoporosis.2,3 InChina, the prevalence of osteoporosis in adults isapproximately 7%, 22.5% in males aged 50 yearsand above, and 50.1%.4 Another multicenter studyrevealed that the age-standardized prevalence of osteoporosis in men and women aged > 50 yearsin China was 6.46% and 29.13%, respectively.5Acceleration of the aging process has led to anincrease in the incidence of osteoporosis andosteoporotic fractures. These conditions now posea significant public health problem, impacting themedical and economic development of countriesworldwide.6,7 Therefore, preventing osteoporosisor detecting it early, along with effectivelymanaging it, can improve patients'quality of lifeand reduce their financial burden. 332 Acta Med Indones - Indones J Intern Med • Vol 57 • Number 3 • July 2025 With the rapid evolution of imaging technology,an increasing number of techniques for diagnosingosteoporosis have been introduced, such as dual -energy X-ray absorptiometry, quantitative CT, andquantitative ultrasound absorptiometry.8-10 Thegold standard for the diagnosis of osteoporosisis measurement of bone mineral density (BMD)using dual-energy X-ray absorptiometry (DXA) .11According to the recommendat

FAQs

What is the main purpose of this study?

To develop machine learning algorithms that accurately predict the risk of bone mineral density loss and support prevention, early detection, and treatment in clinical practice.

Which machine learning models were evaluated?

Five models were tested: k-nearest neighbor (KNN), random forest (RF), support vector machine (SVM), artificial neural network (ANN), and logistic regression (LR).

How was model performance measured?

Performance was evaluated using accuracy, sensitivity, specificity, and AUROC (area under the receiver operating characteristic curve).

Aug 2, 2026

Upload by Theodora