Effect of Artificial Intelligence or Machine Learning on Prediction of Hip Fracture Risk: Systematic Review

Effect of Artificial Intelligence or Machine Learning on Prediction of Hip Fracture Risk: Systematic Review | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Effect of Artificial Intelligence or Machine Learning on Prediction of Hip Fracture Risk: Systematic Review Yonghan Cha, Jung-Taek Kim, Jin-Woo Kim, Sang Yeob Lee, Jun-Il Yoo, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-2351785/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Introduction: Worldwide, 158 million people over the age of 50 are estimated to have high risk of osteoporotic fractures. It is important to prevent the occurrence of primary fractures by diagnosing and treating osteoporosis at an early stage. Dual energy absorptiometry (DEXA) is one of the preferred modality for screening or diagnosis of osteoporosis and can predict the risk of hip fracture to some extent. However, the DEXA test may be difficult to implement easily in some developing countries and fractures have been observed before patients underwent DEXA. The purpose of this systematic review is to search for studies that predict the risk of hip fracture using AI or ML, organize the results of each study, and analyze the usefulness of this technology. Methods: PubMed Central, OVID Medline, Cochrane Collaboration Library, Web of Science, EMBASE, and AHRQ databases were searched to identify relevant studies published up to June 2022 with English language restriction. The following search terms were used ("hip fractures"[MeSH Terms] OR ("hip"[All Fields] AND "fractures"[All Fields]) OR "hip fractures"[All Fields] OR ("hip"[All Fields] AND "fracture"[All Fields]) OR "hip fracture"[All Fields]) AND ("artificial intelligence"[MeSH Terms] OR ("artificial"[All Fields] AND "intelligence"[All Fields]) OR "artificial intelligence"[All Fields]). Results: 7 studies are included in this study. The total number of subjects included in the 7 studies was 330,099. There were 3 studies that included only women, and 4 studies included both men and women. One study conducted AI training after 1:1 matching between fractured and non-fractured patients. The AUC of AI prediction model for hip fracture risk was 0.39–0.96. The accuracy of AI prediction model for hip fracture risk was 70.26–90%. Conclusion: We believe that predicting the risk of hip fracture by the AI model will help select patients with high fracture risk among osteoporosis patients. However, in order to apply the AI model to the prediction of hip fracture risk in clinical situations, it is necessary to identify the characteristics of the dataset and AI model and use it after performing appropriate validation. hip fracture artificial intelligence machine learning diagnosis prediction Figures Figure 1 Introduction Worldwide, 158 million people over the age of 50 are estimated to have high risk of osteoporotic fractures.( 1 ) Thus, 1 in 5 men and 1 in 3 women over the age of 50 will experience an osteoporotic fracture.( 2 ) If the high risk of osteoporotic fractures due to an aging population continues, it is predicted that the number of osteoporotic fractures will double by 2045.( 1 ) After osteoporotic fractures, patients suffer from reduced quality of life due to chronic pain, functional disability and dependence, high morbidity and mortality.( 3 ) Especially, elderly hip fractures cause socioeconomic burden in developed countries.( 4 ) Therefore, various programs to properly treat these patients and prevent re-fracture are being implemented in various countries.( 5 ) However, it may be more important to prevent the occurrence of primary fractures by diagnosing and treating osteoporosis, the main cause of these fractures, at an early stage. Dual energy absorptiometry (DEXA) is one of the preferred modality for screening or diagnosis of osteoporosis and can predict the risk of hip fracture to some extent.( 6 , 7 ) However, a study by Hsieh et al found that 80% of patients between the ages of 40 and 90 who had visited their institution and had pelvis or spine radiographs did not have a DEXA test.( 7 ) Also, the DEXA test may be difficult to implement easily in some developing countries.( 1 ) Although the fracture risk assessment tool (FRAX) can predict the risk of hip fracture, there are symptomatic lumbar fractures or occult hip fractures, and fractures have been observed before patients underwent DEXA.( 1 ) Therefore, if there is no hassle of performing additional tests and a method of predicting the risk of hip fracture only by taking radiographs is provided, it will be possible to reduce the radiation exposure of patients and reduce additional costs. The artificial intelligence (AI) or machine learning (ML) is a computational modeling tool that has become widely accepted for modeling complex real-world health problems.( 8 ) AI has already been used in many fields of medicine, such as nephrology, microbiology, and radiology, and is being studied in various fields such as diagnosis of fractures and prediction of clinical course in the orthopedics.( 9 ) Although it is possible to predict the risk of hip fracture only with BMD or clinical factors, the prediction model using AI can compensate for the shortcomings of existing examination methods and handle large numbers of input variables simultaneously. Also, if an automated system of AI is contructed, there is an advantage that the hassle of checking examinations can be solved.( 10 ) However, it seems that little is known about the assessment of hip fracture risk by AI yet. Therefore, the purpose of this systematic review is to search for studies that predict the risk of hip fracture using AI or ML, organize the results of each study, and analyze the usefulness of this technology. Methods 2.1 Study eligibility criteria Studies were selected based on the following inclusion criteria: 1) studies using AI or ML techniques for prediction of hip fracture risk, such as femoral neck fracture, intertrochanteric fracture, or subtrochanteric fracture; and 2) studies reporting on statistical analysis of area under the ROC (receiver operating characteristic) curve (AUC) or accuracy for prediction of hip fracture risk. Studies were excluded if they failed to meet the above criteria. 2.2 Search methods for identification of studies PubMed Central, OVID Medline, Cochrane Collaboration Library, Web of Science, EMBASE, and AHRQ databases were searched to identify relevant studies published up to June 2022 with English language restriction. The following search terms were used ("hip fractures"[MeSH Terms] OR ("hip"[All Fields] AND "fractures"[All Fields]) OR "hip fractures"[All Fields] OR ("hip"[All Fields] AND "fracture"[All Fields]) OR "hip fracture"[All Fields]) AND ("artificial intelligence"[MeSH Terms] OR ("artificial"[All Fields] AND "intelligence"[All Fields]) OR "artificial intelligence"[All Fields]). Manual search was also conducted for possibly related references. Two of authors, Yonghan Cha and Jung-Taek Kim, reviewed the titles, abstracts, and full texts of all potentially relevant studies independently, as recommended by the Cochrane Collaboration. Any disagreement was resolved by the third reviewer, Jun-Il Yoo. We assessed full-text articles of the remaining studies according to the previously defined inclusion and exclusion criteria, and then selected eligible articles. The reviewers were not blinded to authors, institutions, or the publication. 2.3 Data extraction The following information was extracted from the included articles: authors, publication year, study period, number of patients, sex, age, AI algorithm variables for AI training, AUC or accuracy for prediction of hip fracture risk. Results The initial search identified 123 references from the selected databases. Eighty-two references were excluded by screening the abstracts and titles for duplicates, unrelated articles, case reports, systematic reviews. The remaining 45 studies underwent full-text reviews and subsequently, 38 studies were excluded. Finally, 7 studies are included in this study.( 11 – 17 ) The details of the identification of relevant studies are shown in the flow chart of the study selection process.(Fig. 1 ) The total number of subjects included in the 7 studies was 330,099.(Table 1 .) There were 3 studies that included only women,( 11 – 13 ) and 4 studies included both men and women.( 14 – 17 ) One study conducted AI training after 1:1 matching between fractured and non-fractured patients.( 17 ) Table 1 Study, study period, demographic data of included studies Author Publication year Study period Subjects Total n of patients n of hip fracture patients Age (mean ± sd or range) n of Female (%) Hsieh 2021 2006.–2020. Aged 40–90 years who underwent hip and spine radiographs 23,339 (hip 5164, spine 18175) No description Hip: 72.2 ± 11.2 Spine: 67.1 ± 10.6 Hip: 3,997 (77.4%) Spine: 14,469 (79.6%) Engels 2020 2008.4.–2014.3 over 65 years of age 288,086 7,644 (2.7%) 75.67 ± 6.2 140,709 (48.8%) Villamor 2020 No description Postmenopausal women 137 89 (65%) 81.4 ± 6.95 137 (100%) Ho-Le 2017 No description Women over 60 years of age 1,167 90 (7.7%) No Fx.: 69.1 ± 6.4 Hip Fx.: 76.8 ± 7.5 1,167 (100%) Kruse 2017 1996.–2006. Danish national patient registry Men: 717 Women: 4,722 340 (6.3%, men: 47, women: 293) Men No Fx.: 61.8 Hip Fx.: 69.3 Women No Fx.: 59.7 Hip Fx.: 74.5 4,722 (86.8%) Jiang 2015 No description Postmenopausal women 11,497 186 (1.6%) AI training group No Fx.: 62 ± 7 Hip Fx.: 68.8 ± 6.9 AI validation group No Fx.: 62 ± 7.2 Hip Fx.: 69.5 ± 5.7 11,497 (100%) Tseng 2013 2004.4.–2006.1. Over 60 years of age 434 217 (50%, men 68, women 149) Men No Fx.: 78.4 ± 7.9 Hip Fx.: 70 ± 7.4 Women No Fx.: 77.8 ± 6.8 Hip Fx.: 80.7 ± 7.8 298 (68.7%) n: number, sd: standard deviation, Fx.: fracture, AI: artificial intelligence The AI algorithm used for prediction of hip fracture risk was very diverse, such as Artificial Neural Network, support vector machine, and K-nearest neighbors et al.(Table 2 ) In addition, the variables used for AI training included not only demographic factors such as age, sex, body mass index, and past medical history, but also socioeconomic factors such as income level and education level. Also, Villamor et al. used geometrical factors of femur from finite element analysis with patient`s demographic factors for AI training.( 11 ) Hsieh et al. used pelvis and lumbar spine radiographs and DEXA as variables for hip fracture risk prediction. The AUC of AI prediction model for hip fracture risk was 0.39–0.96. The accuracy of AI prediction model for hip fracture risk was 70.26–90%. Table 2 Prediction model of artificial intelligence and results of prediction for hip fracture risk in included studies Author AI algorithm Training variables for AI AUC for prediction of hip fracture risk Accuracy for prediction of hip fracture risk Hsieh No description Pelvis and lumbar spine radiographs and DEXA 10-years risk: 0.96 10-years risk: 90 Engels LR, Random Forest, SVM, RUSBoost, Superlearner, XGBoost Age, gender, prior fracture history, medication use within administrative claims data 4-years risk LR: 0.695–0.704 Random Forest: 0.685 SVM: 0.650 RUSBoost: 0.702 Superlearner: 0.698 XGBoost: 0.703 No description Villamor SVM, LR, ANN, Random Forest Age, height, weight, BMI, BMD, geometrical factors of femur from FEA No description SVM: 78.35 LR: 73.09 ANN: 70.26 Random Forest: 73.34 Ho-Le LR, ANN, KNN, SVM BMD, fracture history, frequency of falls during the previous 12 months, calcium intakes, alcohol consumption, cigarette, metabolic equivalent index, Height, weight No description 10-years hip fracture risk ANN: 87.3 LR: 81.5 KNN: 79.4 SVM: 81.5 Kruse Xtreme Gradient Boosting, Conditional Inference Random Forest, Generalized Additive Model, ADABoost, Random Forest, Generalized Linear Model, Bagged Multivariate Adaptive Regression Splines, Bayesian Generalized Linear Model, Bagged Flexible Discriminant, Bagged Tree, LR, Classification Tree, Stochastic Gradient Boosting, KNN Medication use, total medication costs, ICD-10 codes, maximum length in years from occurrence to the scan date, CCI, Yearly income, Primary medical visit count and costs during both the prior and post periods, education level, job, ethnicity, age, sex, height, BMI, DEXA 5-years hip fracture risk in women Xtreme Gradient Boosting: 0.92 Random Forest: 0.91 Bagged Flexible Discriminant: 0.91 Bagged Multivariate Adaptive Regression Splines: 0.91 Generalized Additive Model: 0.89 Conditional Inference Random Forest: 0.86 Bagged Tree: 0.87 LR: 0.86 Generalized Linear Model: 0.85 KNN: 0.83 5-years hip fracture risk in men Xtreme Gradient Boosting: 0.89 Conditional Inference Random Forest: 0.88 Generalized Additive Model: 0.89 ADABoost: 0.84 Random Forest: 0.81 Generalized Linear Model: 0.83 Bagged Multivariate Adaptive Regression Splines: 0.80 Bayesian Generalized Linear Model: 0.77 Bagged Flexible Discriminant: 0.74 Bagged Tree: 0.68 LR: 0.58 Classification Tree: 0.57 Stochastic Gradient Boosting; 0.39 No description Jiang SVM Ethnicity, self-reported health, fracture history, physical activity, smoking status, parent broke hip, corticosteroid use, diabetes treatment, age, height, weight, BMD, hip geometry 0.881 No description Tseng ANN Monthly income, weight, height, leisure-time physical activity, MMSE score, peak expiratory flow rate, hand grip strength, BMD 0.868 No description AI: artificial intelligence, ML: machine learning, DEXA: dual energy absorptiometry, BMI: body mass index, BMD: bone mineral density, FEA: finite element analysis, LR: Logistic regression, ANN: Artificial Neural Network, SVN: support vector machine, KNN: K-nearest neighbors, CCI: Charlson`s comorbidity score, ICD: International Classification of Diseases, AUC: area under the receiver operating characteristic curve Discussions The identification of high-risk individuals for hip fracture is clinically and socioeconomically important, because it could facilitate early intervention to reduce the burden of hip fracture in the general population.( 12 ) However, osteoporosis is a silent disease that progresses before osteoporotic fractures.( 18 ) After fracture has occurred, it increases mortality and morbidity in affected patients. Therefore, population-based screening is essential to identifying at-risk patients and implementing preventive services. But, Prediction of hip fracture is very difficult, because it is influenced by multiple risk factors. Known risk factors for hip fracture are low bone mineral density (BMD), previous history of hip fracture, female, advanced age, lower body weight and physical activity, sarcopenia, alcohol consumption, and smoking et al.( 19 ) Although the most clinically important risk factor is low BMD, considering BMD and other clinical factors together for predicting fracture risk can increase the accuracy.( 12 ) However, as in FRAX, the assessment of hip fracture risk using conventional methods may not include several important factors.( 10 ) On the other hand, hip fracture prediction using ML can handle large numbers of input variables simultaneously and consider invisible relationships between variables.( 10 ) Also, as Kruse et al. showed in their study, there is an advantage of not having to go through input work by clinicians if the system is built to automatically analyze clinical data or image data in the AI model.( 16 ) Considering the results of the studies included in this review, AUC and accuracy for prediction of hip fracture risk by AI ranged from low to high. This seems to be because the variables and AI algorithms used are very diverse. These reports also make it difficult for clinicians to determine which model is the best. Therefore, when evaluating the results of studies on the prediction of hip fracture risk using the AI model, the following should be considered. The first is to check whether external validation is present. Kruse et al. argued that it can be very dangerous to report the results after training using one AI model or dataset.( 16 ) They also said that some studies have reported the results of prediction models without external validation, and these problems are not well known. The second factor to pay attention to in hip fracture prediction analysis using AI model is overfitting. Ho-Le et al. reported that overfitting could be a problem in any AI models, and that the number of hip fractures per risk factors was > 10 to prevent overfitting.( 12 ) They argued that the consistency between training and test results of AI algorithm models is important to determine whether over-fitting is occurring. A third consideration is the problem of handling missing data in the dataset used for AI training. A lot of data is required for AI training, and especially big database facilitates this. However, not all patients in the database have all the data. Therefore, it is advisable to check whether techniques for missing data such as imputation are used. However, Jiang et al. reported that these techniques were not necessary because the purpose of their study was primarily to demonstrate the potential increase in predictive ability by combining clinical and computational data.( 13 ) There are several limitations in our study. First, we did not consider the degree of training of AI algorism. Second, we did not consider how many variables were used in the prediction model in the interpretation of the study results. Third, we have not been able to conclude which AI model is the most accurate for prediction of hip fracture yet. Conclusions We believe that predicting the risk of hip fracture by the AI model will help select patients with high fracture risk among osteoporosis patients. However, in order to apply the AI model to the prediction of hip fracture risk in clinical situations, it is necessary to identify the characteristics of the dataset and AI model and use it after performing appropriate validation. Declarations Competing interests All authors confirmed that there is no conflict of interest. Authors’ contributions Yonghan Cha, Jun-Il Yoo conceived and designed the articles. Yonghan Cha, Jung-Taek Kim, Sung Hyo Seo, Jin-Woo Kim, Sang Yeob Lee and Jun-Il Yoo performed the searching and screening. Yonghan Cha, Jung-Taek Kim, and Jin-Woo Kim analyzed and interpreted the data. Yonghan Cha wrote the paper. The authors read and approved the final manuscript. Funding This research was supported by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number : HI22C0494). Acknowledgement Not applicable Availability of data and materials All data generated or analyzed during this study are included in this published article. Ethics approval and consent to participate Not applicable Consent for publication - Not applicable References Johnell O, Kanis JA. An estimate of the worldwide prevalence and disability associated with osteoporotic fractures. Osteoporos Int J Establ Result Coop Eur Found Osteoporos Natl Osteoporos Found USA. 2006;17:1726–33. doi: 10.1007/s00198-006-0172-4 . Sànchez-Riera L, Carnahan E, Vos T, Veerman L, Norman R, Lim SS, Hoy D, Smith E, Wilson N, Nolla JM, et al. The global burden attributable to low bone mineral density. Ann Rheum Dis. 2014;73:1635–45. doi: 10.1136/annrheumdis-2013-204320 . Saito T, Sterbenz JM, Malay S, Zhong L, MacEachern MP, Chung KC. Effectiveness of anti-osteoporotic drugs to prevent secondary fragility fractures: systematic review and meta-analysis. Osteoporos Int J Establ Result Coop Eur Found Osteoporos Natl Osteoporos Found USA. 2017;28:3289–300. doi: 10.1007/s00198-017-4175-0 . Cha YH, Ha Y-C, Lim J-Y, Kim W-S. Introduction of the Cost-Effectiveness Studies of Fracture Liaison Service in Other Countries. J Bone Metab. 2020;27:79–83. doi: 10.11005/jbm.2020.27.2.79 . Cha YH, Ha Y-C, Lim J-Y. Establishment of Fracture Liaison Service in Korea: Where Is It Stand and Where Is It Going? J Bone Metab. 2019;26:207–11. doi: 10.11005/jbm.2019.26.4.207 . LeBoff MS, Greenspan SL, Insogna KL, Lewiecki EM, Saag KG, Singer AJ, Siris ES. The clinician’s guide to prevention and treatment of osteoporosis. Osteoporos Int J Establ Result Coop Eur Found Osteoporos Natl Osteoporos Found USA. 2022;33:2049–102. doi: 10.1007/s00198-021-05900-y . Cheng C-T, Wang Y, Chen H-W, Hsiao P-M, Yeh C-N, Hsieh C-H, Miao S, Xiao J, Liao C-H, Lu L. A scalable physician-level deep learning algorithm detects universal trauma on pelvic radiographs. Nat Commun. 2021;12:1066. doi: 10.1038/s41467-021-21311-3 . Basheer IA, Hajmeer M. Artificial neural networks: fundamentals, computing, design, and application. J Microbiol Methods. 2000;43:3–31. doi: 10.1016/S0167-7012(00)00201-3 . Patel JL, Goyal RK. Applications of artificial neural networks in medical science. Curr Clin Pharmacol. 2007;2:217–26. doi: 10.2174/157488407781668811 . Vries BCS de, Hegeman JH, Nijmeijer W, Geerdink J, Seifert C, Groothuis-Oudshoorn CGM. Comparing three machine learning approaches to design a risk assessment tool for future fractures: predicting a subsequent major osteoporotic fracture in fracture patients with osteopenia and osteoporosis. Osteoporos Int. 2021;32:437–49. doi: 10.1007/s00198-020-05735-z . Villamor E, Monserrat C, Del Río L, Romero-Martín JA, Rupérez MJ. Prediction of osteoporotic hip fracture in postmenopausal women through patient-specific FE analyses and machine learning. Comput Methods Programs Biomed. 2020;193:105484. doi: 10.1016/j.cmpb.2020.105484 . Ho-Le TP, Center JR, Eisman JA, Nguyen TV, Nguyen HT. Prediction of hip fracture in post-menopausal women using artificial neural network approach. Annu Int Conf IEEE Eng Med Biol Soc IEEE Eng Med Biol Soc Annu Int Conf. 2017;2017:4207–10. doi: 10.1109/EMBC.2017.8037784 . Jiang P, Missoum S, Chen Z. Fusion of clinical and stochastic finite element data for hip fracture risk prediction. J Biomech. 2015;48:4043–52. doi: 10.1016/j.jbiomech.2015.09.044 . Hsieh C-I, Zheng K, Lin C, Mei L, Lu L, Li W, Chen F-P, Wang Y, Zhou X, Wang F, et al. Automated bone mineral density prediction and fracture risk assessment using plain radiographs via deep learning. Nat Commun. 2021;12:5472. doi: 10.1038/s41467-021-25779-x . Engels A, Reber KC, Lindlbauer I, Rapp K, Büchele G, Klenk J, Meid A, Becker C, König H-H. Osteoporotic hip fracture prediction from risk factors available in administrative claims data – A machine learning approach. PLoS ONE. 2020;15:e0232969. doi: 10.1371/journal.pone.0232969 . Kruse C, Eiken P, Vestergaard P. Machine Learning Principles Can Improve Hip Fracture Prediction. Calcif Tissue Int. 2017;100:348–60. doi: 10.1007/s00223-017-0238-7 . Tseng W-J, Hung L-W, Shieh J-S, Abbod MF, Lin J. Hip fracture risk assessment: artificial neural network outperforms conditional logistic regression in an age- and sex-matched case control study. BMC Musculoskelet Disord. 2013;14:207. doi: 10.1186/1471-2474-14-207 . Nazrun AS, Tzar MN, Mokhtar SA, Mohamed IN. A systematic review of the outcomes of osteoporotic fracture patients after hospital discharge: morbidity, subsequent fractures, and mortality. Ther Clin Risk Manag. 2014;10:937–48. doi: 10.2147/TCRM.S72456 . Marks R. Hip fracture epidemiological trends, outcomes, and risk factors, 1970–2009. Int J Gen Med. 2010;3:1–17. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-2351785","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":162526425,"identity":"437eaeb4-87f8-4b09-b42b-d89047a290f4","order_by":0,"name":"Yonghan Cha","email":"","orcid":"","institution":"Eulji University Hospital","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Yonghan","middleName":"","lastName":"Cha","suffix":""},{"id":162526426,"identity":"b9eb2b73-3ee0-4ed0-8baf-957dac11f1ef","order_by":1,"name":"Jung-Taek Kim","email":"","orcid":"","institution":"Ajou University School of Medicine, Ajou Medical Center","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Jung-Taek","middleName":"","lastName":"Kim","suffix":""},{"id":162526427,"identity":"ac01a599-8f76-41c9-92b0-ae1b4d41944d","order_by":2,"name":"Jin-Woo Kim","email":"","orcid":"","institution":"Nowon Eulji Medical Center, Eulji University","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Jin-Woo","middleName":"","lastName":"Kim","suffix":""},{"id":162526428,"identity":"01d57658-c26e-4bcb-8b06-603a310ffac5","order_by":3,"name":"Sang Yeob Lee","email":"","orcid":"","institution":"Gyeongsang national university hospital","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Sang","middleName":"Yeob","lastName":"Lee","suffix":""},{"id":162526429,"identity":"3e7163aa-14d4-4ed6-bb0f-cbfd49b1968a","order_by":4,"name":"Jun-Il Yoo","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAv0lEQVRIiWNgGAWjYBAC9gYGhgMMFQfAnAMPiNHCcwCk5QxUSwJIhBgtDIxtEC0MxGmRyH146Oa8O/LmYocfAm25I2dPWEu6weHcbc8Md85OMwBqeWZM0BZ7iTQGoJbDjBtuJ4C0HE7sIWwLSMucw/Ybbqd/AGmpJ1JLw+HEDbdzwLYkEPY+zzOGwznHDicDtRQcSDA4bNhzgJAW9jTmzzk1h22BDtv84UPFYXlQ5JICDEhTPgpGwSgYBaMABwAAnXBJBZPNfwEAAAAASUVORK5CYII=","orcid":"","institution":"Gyeongsang national university hospital","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Jun-Il","middleName":"","lastName":"Yoo","suffix":""},{"id":162526431,"identity":"2b3f938e-affd-438c-bb72-90371173e663","order_by":5,"name":"Sung Hyo Seo","email":"","orcid":"","institution":"Gyeongsang national university hospital","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Sung","middleName":"Hyo","lastName":"Seo","suffix":""}],"badges":[],"createdAt":"2022-12-07 01:44:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-2351785/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-2351785/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":30898496,"identity":"a226a1b6-9490-4acd-a1d1-b1a448873391","added_by":"auto","created_at":"2022-12-29 18:23:13","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":661149,"visible":true,"origin":"","legend":"\u003cp\u003eThe flow chart of the study selection process.\u003c/p\u003e","description":"","filename":"Onlinefigure1.png","url":"https://assets-eu.researchsquare.com/files/rs-2351785/v1/fb6b602200703c1cbc2cf6f6.png"},{"id":35930049,"identity":"38395e49-7dd6-4938-a1dc-1c866209a680","added_by":"auto","created_at":"2023-04-18 11:14:39","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":544143,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-2351785/v1/e356aa93-9e1b-43e8-b869-2ec001d24ea7.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Effect of Artificial Intelligence or Machine Learning on Prediction of Hip Fracture Risk: Systematic Review","fulltext":[{"header":"Introduction","content":"\u003cp\u003eWorldwide, 158\u0026nbsp;million people over the age of 50 are estimated to have high risk of osteoporotic fractures.(\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) Thus, 1 in 5 men and 1 in 3 women over the age of 50 will experience an osteoporotic fracture.(\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) If the high risk of osteoporotic fractures due to an aging population continues, it is predicted that the number of osteoporotic fractures will double by 2045.(\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) After osteoporotic fractures, patients suffer from reduced quality of life due to chronic pain, functional disability and dependence, high morbidity and mortality.(\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) Especially, elderly hip fractures cause socioeconomic burden in developed countries.(\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e) Therefore, various programs to properly treat these patients and prevent re-fracture are being implemented in various countries.(\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e) However, it may be more important to prevent the occurrence of primary fractures by diagnosing and treating osteoporosis, the main cause of these fractures, at an early stage.\u003c/p\u003e \u003cp\u003eDual energy absorptiometry (DEXA) is one of the preferred modality for screening or diagnosis of osteoporosis and can predict the risk of hip fracture to some extent.(\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e) However, a study by Hsieh et al found that 80% of patients between the ages of 40 and 90 who had visited their institution and had pelvis or spine radiographs did not have a DEXA test.(\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e) Also, the DEXA test may be difficult to implement easily in some developing countries.(\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) Although the fracture risk assessment tool (FRAX) can predict the risk of hip fracture, there are symptomatic lumbar fractures or occult hip fractures, and fractures have been observed before patients underwent DEXA.(\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) Therefore, if there is no hassle of performing additional tests and a method of predicting the risk of hip fracture only by taking radiographs is provided, it will be possible to reduce the radiation exposure of patients and reduce additional costs.\u003c/p\u003e \u003cp\u003eThe artificial intelligence (AI) or machine learning (ML) is a computational modeling tool that has become widely accepted for modeling complex real-world health problems.(\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e) AI has already been used in many fields of medicine, such as nephrology, microbiology, and radiology, and is being studied in various fields such as diagnosis of fractures and prediction of clinical course in the orthopedics.(\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e) Although it is possible to predict the risk of hip fracture only with BMD or clinical factors, the prediction model using AI can compensate for the shortcomings of existing examination methods and handle large numbers of input variables simultaneously. Also, if an automated system of AI is contructed, there is an advantage that the hassle of checking examinations can be solved.(\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e) However, it seems that little is known about the assessment of hip fracture risk by AI yet.\u003c/p\u003e \u003cp\u003eTherefore, the purpose of this systematic review is to search for studies that predict the risk of hip fracture using AI or ML, organize the results of each study, and analyze the usefulness of this technology.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Study eligibility criteria\u003c/h2\u003e \u003cp\u003eStudies were selected based on the following inclusion criteria: 1) studies using AI or ML techniques for prediction of hip fracture risk, such as femoral neck fracture, intertrochanteric fracture, or subtrochanteric fracture; and 2) studies reporting on statistical analysis of area under the ROC (receiver operating characteristic) curve (AUC) or accuracy for prediction of hip fracture risk. Studies were excluded if they failed to meet the above criteria.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Search methods for identification of studies\u003c/h2\u003e \u003cp\u003ePubMed Central, OVID Medline, Cochrane Collaboration Library, Web of Science, EMBASE, and AHRQ databases were searched to identify relevant studies published up to June 2022 with English language restriction. The following search terms were used (\"hip fractures\"[MeSH Terms] OR (\"hip\"[All Fields] AND \"fractures\"[All Fields]) OR \"hip fractures\"[All Fields] OR (\"hip\"[All Fields] AND \"fracture\"[All Fields]) OR \"hip fracture\"[All Fields]) AND (\"artificial intelligence\"[MeSH Terms] OR (\"artificial\"[All Fields] AND \"intelligence\"[All Fields]) OR \"artificial intelligence\"[All Fields]). Manual search was also conducted for possibly related references. Two of authors, Yonghan Cha and Jung-Taek Kim, reviewed the titles, abstracts, and full texts of all potentially relevant studies independently, as recommended by the Cochrane Collaboration. Any disagreement was resolved by the third reviewer, Jun-Il Yoo. We assessed full-text articles of the remaining studies according to the previously defined inclusion and exclusion criteria, and then selected eligible articles. The reviewers were not blinded to authors, institutions, or the publication.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Data extraction\u003c/h2\u003e \u003cp\u003eThe following information was extracted from the included articles: authors, publication year, study period, number of patients, sex, age, AI algorithm variables for AI training, AUC or accuracy for prediction of hip fracture risk.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003eThe initial search identified 123 references from the selected databases. Eighty-two references were excluded by screening the abstracts and titles for duplicates, unrelated articles, case reports, systematic reviews. The remaining 45 studies underwent full-text reviews and subsequently, 38 studies were excluded. Finally, 7 studies are included in this study.(\u003cspan additionalcitationids=\"CR12 CR13 CR14 CR15 CR16\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e) The details of the identification of relevant studies are shown in the flow chart of the study selection process.(Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e)\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe total number of subjects included in the 7 studies was 330,099.(Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.) There were 3 studies that included only women,(\u003cspan additionalcitationids=\"CR12\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e) and 4 studies included both men and women.(\u003cspan additionalcitationids=\"CR15 CR16\" citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e) One study conducted AI training after 1:1 matching between fractured and non-fractured patients.(\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e)\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eStudy, study period, demographic data of included studies\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"8\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAuthor\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePublication\u003c/p\u003e \u003cp\u003eyear\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eStudy period\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSubjects\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eTotal n of patients\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003en of hip fracture patients\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eAge (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;sd or range)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003en of Female (%)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHsieh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2006.\u0026ndash;2020.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAged 40\u0026ndash;90 years who underwent hip and spine radiographs\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e23,339\u003c/p\u003e \u003cp\u003e(hip 5164, spine 18175)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNo description\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eHip: 72.2\u0026thinsp;\u0026plusmn;\u0026thinsp;11.2\u003c/p\u003e \u003cp\u003eSpine: 67.1\u0026thinsp;\u0026plusmn;\u0026thinsp;10.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eHip: 3,997 (77.4%)\u003c/p\u003e \u003cp\u003eSpine: 14,469 (79.6%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEngels\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2008.4.\u0026ndash;2014.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eover 65 years of age\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e288,086\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e7,644 (2.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e75.67\u0026thinsp;\u0026plusmn;\u0026thinsp;6.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e140,709 (48.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVillamor\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo description\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePostmenopausal women\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e137\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e89 (65%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e81.4\u0026thinsp;\u0026plusmn;\u0026thinsp;6.95\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e137 (100%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHo-Le\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2017\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo description\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWomen over 60 years of age\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1,167\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e90 (7.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNo Fx.: 69.1\u0026thinsp;\u0026plusmn;\u0026thinsp;6.4\u003c/p\u003e \u003cp\u003eHip Fx.: 76.8\u0026thinsp;\u0026plusmn;\u0026thinsp;7.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1,167 (100%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKruse\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2017\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1996.\u0026ndash;2006.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDanish national patient registry\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eMen: 717\u003c/p\u003e \u003cp\u003eWomen: 4,722\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e340 (6.3%, men: 47, women: 293)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eMen\u003c/p\u003e \u003cp\u003eNo Fx.: 61.8\u003c/p\u003e \u003cp\u003eHip Fx.: 69.3\u003c/p\u003e \u003cp\u003eWomen\u003c/p\u003e \u003cp\u003eNo Fx.: 59.7\u003c/p\u003e \u003cp\u003eHip Fx.: 74.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e4,722 (86.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eJiang\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2015\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNo description\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePostmenopausal women\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e11,497\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e186 (1.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eAI training group\u003c/p\u003e \u003cp\u003eNo Fx.: 62\u0026thinsp;\u0026plusmn;\u0026thinsp;7\u003c/p\u003e \u003cp\u003eHip Fx.: 68.8\u0026thinsp;\u0026plusmn;\u0026thinsp;6.9\u003c/p\u003e \u003cp\u003eAI validation group\u003c/p\u003e \u003cp\u003eNo Fx.: 62\u0026thinsp;\u0026plusmn;\u0026thinsp;7.2\u003c/p\u003e \u003cp\u003eHip Fx.: 69.5\u0026thinsp;\u0026plusmn;\u0026thinsp;5.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e11,497 (100%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTseng\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2013\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2004.4.\u0026ndash;2006.1.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eOver 60 years of age\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e434\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e217 (50%, men 68, women 149)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eMen\u003c/p\u003e \u003cp\u003eNo Fx.: 78.4\u0026thinsp;\u0026plusmn;\u0026thinsp;7.9\u003c/p\u003e \u003cp\u003eHip Fx.: 70\u0026thinsp;\u0026plusmn;\u0026thinsp;7.4\u003c/p\u003e \u003cp\u003eWomen\u003c/p\u003e \u003cp\u003eNo Fx.: 77.8\u0026thinsp;\u0026plusmn;\u0026thinsp;6.8\u003c/p\u003e \u003cp\u003eHip Fx.: 80.7\u0026thinsp;\u0026plusmn;\u0026thinsp;7.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e298 (68.7%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"8\"\u003en: number, sd: standard deviation, Fx.: fracture, AI: artificial intelligence\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe AI algorithm used for prediction of hip fracture risk was very diverse, such as Artificial Neural Network, support vector machine, and K-nearest neighbors et al.(Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) In addition, the variables used for AI training included not only demographic factors such as age, sex, body mass index, and past medical history, but also socioeconomic factors such as income level and education level. Also, Villamor et al. used geometrical factors of femur from finite element analysis with patient`s demographic factors for AI training.(\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e) Hsieh et al. used pelvis and lumbar spine radiographs and DEXA as variables for hip fracture risk prediction. The AUC of AI prediction model for hip fracture risk was 0.39\u0026ndash;0.96. The accuracy of AI prediction model for hip fracture risk was 70.26\u0026ndash;90%.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePrediction model of artificial intelligence and results of prediction for hip fracture risk in included studies\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAuthor\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAI algorithm\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTraining variables for AI\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAUC for prediction of hip fracture risk\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eAccuracy for prediction of hip fracture risk\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHsieh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo description\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePelvis and lumbar spine radiographs and DEXA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e10-years risk: 0.96\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e10-years risk: 90\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEngels\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLR, Random Forest, SVM, RUSBoost, Superlearner, XGBoost\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAge, gender, prior fracture history, medication use within administrative claims data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4-years risk\u003c/p\u003e \u003cp\u003eLR: 0.695\u0026ndash;0.704\u003c/p\u003e \u003cp\u003eRandom Forest: 0.685\u003c/p\u003e \u003cp\u003eSVM: 0.650\u003c/p\u003e \u003cp\u003eRUSBoost: 0.702\u003c/p\u003e \u003cp\u003eSuperlearner: 0.698\u003c/p\u003e \u003cp\u003eXGBoost: 0.703\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNo description\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVillamor\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSVM, LR, ANN, Random Forest\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAge, height, weight, BMI, BMD, geometrical factors of femur from FEA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNo description\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSVM: 78.35\u003c/p\u003e \u003cp\u003eLR: 73.09\u003c/p\u003e \u003cp\u003eANN: 70.26\u003c/p\u003e \u003cp\u003eRandom Forest: 73.34\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHo-Le\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLR, ANN, KNN, SVM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBMD, fracture history, frequency of falls during the previous 12 months, calcium intakes, alcohol consumption, cigarette, metabolic equivalent index, Height, weight\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNo description\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e10-years hip fracture risk\u003c/p\u003e \u003cp\u003eANN: 87.3\u003c/p\u003e \u003cp\u003eLR: 81.5\u003c/p\u003e \u003cp\u003eKNN: 79.4\u003c/p\u003e \u003cp\u003eSVM: 81.5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKruse\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eXtreme Gradient Boosting, Conditional Inference Random Forest, Generalized Additive Model, ADABoost, Random Forest, Generalized Linear Model, Bagged Multivariate Adaptive Regression Splines, Bayesian Generalized Linear Model, Bagged Flexible Discriminant, Bagged Tree, LR, Classification Tree, Stochastic Gradient Boosting, KNN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMedication use, total medication costs, ICD-10 codes, maximum length in years from occurrence to the scan date, CCI, Yearly income, Primary medical visit count and costs during both the prior and post periods, education level, job, ethnicity, age, sex, height, BMI, DEXA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5-years hip fracture risk in women\u003c/p\u003e \u003cp\u003eXtreme Gradient Boosting: 0.92\u003c/p\u003e \u003cp\u003eRandom Forest: 0.91\u003c/p\u003e \u003cp\u003eBagged Flexible Discriminant: 0.91\u003c/p\u003e \u003cp\u003eBagged Multivariate Adaptive Regression Splines: 0.91\u003c/p\u003e \u003cp\u003eGeneralized Additive Model: 0.89\u003c/p\u003e \u003cp\u003eConditional Inference Random Forest: 0.86\u003c/p\u003e \u003cp\u003eBagged Tree: 0.87\u003c/p\u003e \u003cp\u003eLR: 0.86\u003c/p\u003e \u003cp\u003eGeneralized Linear Model: 0.85\u003c/p\u003e \u003cp\u003eKNN: 0.83\u003c/p\u003e \u003cp\u003e5-years hip fracture risk in men\u003c/p\u003e \u003cp\u003eXtreme Gradient Boosting: 0.89\u003c/p\u003e \u003cp\u003eConditional Inference Random Forest: 0.88\u003c/p\u003e \u003cp\u003eGeneralized Additive Model: 0.89\u003c/p\u003e \u003cp\u003eADABoost: 0.84\u003c/p\u003e \u003cp\u003eRandom Forest: 0.81\u003c/p\u003e \u003cp\u003eGeneralized Linear Model: 0.83\u003c/p\u003e \u003cp\u003eBagged Multivariate Adaptive Regression Splines: 0.80\u003c/p\u003e \u003cp\u003eBayesian Generalized Linear Model: 0.77\u003c/p\u003e \u003cp\u003eBagged Flexible Discriminant: 0.74\u003c/p\u003e \u003cp\u003eBagged Tree: 0.68\u003c/p\u003e \u003cp\u003eLR: 0.58\u003c/p\u003e \u003cp\u003eClassification Tree: 0.57\u003c/p\u003e \u003cp\u003eStochastic Gradient Boosting; 0.39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNo description\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eJiang\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSVM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEthnicity, self-reported health, fracture history, physical activity, smoking status, parent broke hip, corticosteroid use, diabetes treatment, age, height, weight, BMD, hip geometry\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.881\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNo description\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTseng\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eANN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMonthly income, weight, height, leisure-time physical activity, MMSE score, peak expiratory flow rate, hand grip strength, BMD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.868\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNo description\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003eAI: artificial intelligence, ML: machine learning, DEXA: dual energy absorptiometry, BMI: body mass index, BMD: bone mineral density, FEA: finite element analysis, LR: Logistic regression, ANN: Artificial Neural Network, SVN: support vector machine, KNN: K-nearest neighbors, CCI: Charlson`s comorbidity score, ICD: International Classification of Diseases, AUC: area under the receiver operating characteristic curve\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e"},{"header":"Discussions","content":"\u003cp\u003eThe identification of high-risk individuals for hip fracture is clinically and socioeconomically important, because it could facilitate early intervention to reduce the burden of hip fracture in the general population.(\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e) However, osteoporosis is a silent disease that progresses before osteoporotic fractures.(\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e) After fracture has occurred, it increases mortality and morbidity in affected patients. Therefore, population-based screening is essential to identifying at-risk patients and implementing preventive services. But, Prediction of hip fracture is very difficult, because it is influenced by multiple risk factors. Known risk factors for hip fracture are low bone mineral density (BMD), previous history of hip fracture, female, advanced age, lower body weight and physical activity, sarcopenia, alcohol consumption, and smoking et al.(\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e) Although the most clinically important risk factor is low BMD, considering BMD and other clinical factors together for predicting fracture risk can increase the accuracy.(\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e) However, as in FRAX, the assessment of hip fracture risk using conventional methods may not include several important factors.(\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e) On the other hand, hip fracture prediction using ML can handle large numbers of input variables simultaneously and consider invisible relationships between variables.(\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e) Also, as Kruse et al. showed in their study, there is an advantage of not having to go through input work by clinicians if the system is built to automatically analyze clinical data or image data in the AI model.(\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e)\u003c/p\u003e \u003cp\u003eConsidering the results of the studies included in this review, AUC and accuracy for prediction of hip fracture risk by AI ranged from low to high. This seems to be because the variables and AI algorithms used are very diverse. These reports also make it difficult for clinicians to determine which model is the best. Therefore, when evaluating the results of studies on the prediction of hip fracture risk using the AI model, the following should be considered. The first is to check whether external validation is present. Kruse et al. argued that it can be very dangerous to report the results after training using one AI model or dataset.(\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e) They also said that some studies have reported the results of prediction models without external validation, and these problems are not well known. The second factor to pay attention to in hip fracture prediction analysis using AI model is overfitting. Ho-Le et al. reported that overfitting could be a problem in any AI models, and that the number of hip fractures per risk factors was \u0026gt;\u0026thinsp;10 to prevent overfitting.(\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e) They argued that the consistency between training and test results of AI algorithm models is important to determine whether over-fitting is occurring. A third consideration is the problem of handling missing data in the dataset used for AI training. A lot of data is required for AI training, and especially big database facilitates this. However, not all patients in the database have all the data. Therefore, it is advisable to check whether techniques for missing data such as imputation are used. However, Jiang et al. reported that these techniques were not necessary because the purpose of their study was primarily to demonstrate the potential increase in predictive ability by combining clinical and computational data.(\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e)\u003c/p\u003e \u003cp\u003eThere are several limitations in our study. First, we did not consider the degree of training of AI algorism. Second, we did not consider how many variables were used in the prediction model in the interpretation of the study results. Third, we have not been able to conclude which AI model is the most accurate for prediction of hip fracture yet.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eWe believe that predicting the risk of hip fracture by the AI model will help select patients with high fracture risk among osteoporosis patients. However, in order to apply the AI model to the prediction of hip fracture risk in clinical situations, it is necessary to identify the characteristics of the dataset and AI model and use it after performing appropriate validation.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors confirmed that there is no conflict of interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026rsquo; contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eYonghan Cha, Jun-Il Yoo\u0026nbsp;conceived and designed the articles.\u0026nbsp;Yonghan Cha, Jung-Taek Kim,\u0026nbsp;Sung Hyo Seo,\u0026nbsp;Jin-Woo Kim, Sang Yeob Lee and\u0026nbsp;Jun-Il Yoo\u0026nbsp;performed the searching and screening.\u0026nbsp;Yonghan Cha, Jung-Taek Kim, and\u0026nbsp;Jin-Woo Kim\u0026nbsp;analyzed and interpreted the data.\u0026nbsp;Yonghan Cha\u0026nbsp;wrote the paper. The authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was supported by a grant of the Korea Health Technology R\u0026amp;D Project through the Korea\u003c/p\u003e\n\u003cp\u003eHealth Industry Development Institute (KHIDI), funded by the Ministry of Health \u0026amp; Welfare, Republic\u003c/p\u003e\n\u003cp\u003eof Korea (grant number : HI22C0494).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data generated or analyzed during this study are included in this published article.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u0026nbsp;-\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eJohnell O, Kanis JA. An estimate of the worldwide prevalence and disability associated with osteoporotic fractures. Osteoporos Int J Establ Result Coop Eur Found Osteoporos Natl Osteoporos Found USA. 2006;17:1726\u0026ndash;33. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00198-006-0172-4\u003c/span\u003e\u003cspan address=\"10.1007/s00198-006-0172-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eS\u0026agrave;nchez-Riera L, Carnahan E, Vos T, Veerman L, Norman R, Lim SS, Hoy D, Smith E, Wilson N, Nolla JM, et al. The global burden attributable to low bone mineral density. Ann Rheum Dis. 2014;73:1635\u0026ndash;45. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/annrheumdis-2013-204320\u003c/span\u003e\u003cspan address=\"10.1136/annrheumdis-2013-204320\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSaito T, Sterbenz JM, Malay S, Zhong L, MacEachern MP, Chung KC. Effectiveness of anti-osteoporotic drugs to prevent secondary fragility fractures: systematic review and meta-analysis. Osteoporos Int J Establ Result Coop Eur Found Osteoporos Natl Osteoporos Found USA. 2017;28:3289\u0026ndash;300. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00198-017-4175-0\u003c/span\u003e\u003cspan address=\"10.1007/s00198-017-4175-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCha YH, Ha Y-C, Lim J-Y, Kim W-S. Introduction of the Cost-Effectiveness Studies of Fracture Liaison Service in Other Countries. J Bone Metab. 2020;27:79\u0026ndash;83. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.11005/jbm.2020.27.2.79\u003c/span\u003e\u003cspan address=\"10.11005/jbm.2020.27.2.79\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCha YH, Ha Y-C, Lim J-Y. Establishment of Fracture Liaison Service in Korea: Where Is It Stand and Where Is It Going? J Bone Metab. 2019;26:207\u0026ndash;11. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.11005/jbm.2019.26.4.207\u003c/span\u003e\u003cspan address=\"10.11005/jbm.2019.26.4.207\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeBoff MS, Greenspan SL, Insogna KL, Lewiecki EM, Saag KG, Singer AJ, Siris ES. The clinician\u0026rsquo;s guide to prevention and treatment of osteoporosis. Osteoporos Int J Establ Result Coop Eur Found Osteoporos Natl Osteoporos Found USA. 2022;33:2049\u0026ndash;102. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00198-021-05900-y\u003c/span\u003e\u003cspan address=\"10.1007/s00198-021-05900-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCheng C-T, Wang Y, Chen H-W, Hsiao P-M, Yeh C-N, Hsieh C-H, Miao S, Xiao J, Liao C-H, Lu L. A scalable physician-level deep learning algorithm detects universal trauma on pelvic radiographs. Nat Commun. 2021;12:1066. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41467-021-21311-3\u003c/span\u003e\u003cspan address=\"10.1038/s41467-021-21311-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBasheer IA, Hajmeer M. Artificial neural networks: fundamentals, computing, design, and application. J Microbiol Methods. 2000;43:3\u0026ndash;31. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/S0167-7012(00)00201-3\u003c/span\u003e\u003cspan address=\"10.1016/S0167-7012(00)00201-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePatel JL, Goyal RK. Applications of artificial neural networks in medical science. Curr Clin Pharmacol. 2007;2:217\u0026ndash;26. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2174/157488407781668811\u003c/span\u003e\u003cspan address=\"10.2174/157488407781668811\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVries BCS de, Hegeman JH, Nijmeijer W, Geerdink J, Seifert C, Groothuis-Oudshoorn CGM. Comparing three machine learning approaches to design a risk assessment tool for future fractures: predicting a subsequent major osteoporotic fracture in fracture patients with osteopenia and osteoporosis. Osteoporos Int. 2021;32:437\u0026ndash;49. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00198-020-05735-z\u003c/span\u003e\u003cspan address=\"10.1007/s00198-020-05735-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVillamor E, Monserrat C, Del R\u0026iacute;o L, Romero-Mart\u0026iacute;n JA, Rup\u0026eacute;rez MJ. Prediction of osteoporotic hip fracture in postmenopausal women through patient-specific FE analyses and machine learning. Comput Methods Programs Biomed. 2020;193:105484. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cmpb.2020.105484\u003c/span\u003e\u003cspan address=\"10.1016/j.cmpb.2020.105484\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHo-Le TP, Center JR, Eisman JA, Nguyen TV, Nguyen HT. Prediction of hip fracture in post-menopausal women using artificial neural network approach. Annu Int Conf IEEE Eng Med Biol Soc IEEE Eng Med Biol Soc Annu Int Conf. 2017;2017:4207\u0026ndash;10. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1109/EMBC.2017.8037784\u003c/span\u003e\u003cspan address=\"10.1109/EMBC.2017.8037784\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang P, Missoum S, Chen Z. Fusion of clinical and stochastic finite element data for hip fracture risk prediction. J Biomech. 2015;48:4043\u0026ndash;52. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jbiomech.2015.09.044\u003c/span\u003e\u003cspan address=\"10.1016/j.jbiomech.2015.09.044\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHsieh C-I, Zheng K, Lin C, Mei L, Lu L, Li W, Chen F-P, Wang Y, Zhou X, Wang F, et al. Automated bone mineral density prediction and fracture risk assessment using plain radiographs via deep learning. Nat Commun. 2021;12:5472. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41467-021-25779-x\u003c/span\u003e\u003cspan address=\"10.1038/s41467-021-25779-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEngels A, Reber KC, Lindlbauer I, Rapp K, B\u0026uuml;chele G, Klenk J, Meid A, Becker C, K\u0026ouml;nig H-H. Osteoporotic hip fracture prediction from risk factors available in administrative claims data \u0026ndash; A machine learning approach. PLoS ONE. 2020;15:e0232969. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pone.0232969\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0232969\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKruse C, Eiken P, Vestergaard P. Machine Learning Principles Can Improve Hip Fracture Prediction. Calcif Tissue Int. 2017;100:348\u0026ndash;60. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00223-017-0238-7\u003c/span\u003e\u003cspan address=\"10.1007/s00223-017-0238-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTseng W-J, Hung L-W, Shieh J-S, Abbod MF, Lin J. Hip fracture risk assessment: artificial neural network outperforms conditional logistic regression in an age- and sex-matched case control study. BMC Musculoskelet Disord. 2013;14:207. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/1471-2474-14-207\u003c/span\u003e\u003cspan address=\"10.1186/1471-2474-14-207\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNazrun AS, Tzar MN, Mokhtar SA, Mohamed IN. A systematic review of the outcomes of osteoporotic fracture patients after hospital discharge: morbidity, subsequent fractures, and mortality. Ther Clin Risk Manag. 2014;10:937\u0026ndash;48. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2147/TCRM.S72456\u003c/span\u003e\u003cspan address=\"10.2147/TCRM.S72456\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMarks R. Hip fracture epidemiological trends, outcomes, and risk factors, 1970\u0026ndash;2009. Int J Gen Med. 2010;3:1\u0026ndash;17.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"hip fracture, artificial intelligence, machine learning, diagnosis, prediction","lastPublishedDoi":"10.21203/rs.3.rs-2351785/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-2351785/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eIntroduction: \u003c/strong\u003e\u0026nbsp;Worldwide, 158 million people over the age of 50 are estimated to have high risk of osteoporotic fractures. It is important to prevent the occurrence of primary fractures by diagnosing and treating osteoporosis at an early stage. Dual energy absorptiometry (DEXA) is one of the preferred modality for screening or diagnosis of osteoporosis and can predict the risk of hip fracture to some extent. However, the DEXA test may be difficult to implement easily in some developing countries and fractures have been observed before patients underwent DEXA. The purpose of this systematic review is to search for studies that predict the risk of hip fracture using AI or ML, organize the results of each study, and analyze the usefulness of this technology.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods: \u003c/strong\u003ePubMed Central, OVID Medline, Cochrane Collaboration Library, Web of Science, EMBASE, and AHRQ databases were searched to identify relevant studies published up to June 2022 with English language restriction. The following search terms were used (\"hip fractures\"[MeSH Terms] OR (\"hip\"[All Fields] AND \"fractures\"[All Fields]) OR \"hip fractures\"[All Fields] OR (\"hip\"[All Fields] AND \"fracture\"[All Fields]) OR \"hip fracture\"[All Fields]) AND (\"artificial intelligence\"[MeSH Terms] OR (\"artificial\"[All Fields] AND \"intelligence\"[All Fields]) OR \"artificial intelligence\"[All Fields]).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults: \u003c/strong\u003e7 studies are included in this study. The total number of subjects included in the 7 studies was 330,099. There were 3 studies that included only women, and 4 studies included both men and women. One study conducted AI training after 1:1 matching between fractured and non-fractured patients. The AUC of AI prediction model for hip fracture risk was 0.39–0.96. The accuracy of AI prediction model for hip fracture risk was 70.26–90%.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion: \u003c/strong\u003eWe believe that predicting the risk of hip fracture by the AI model will help select patients with high fracture risk among osteoporosis patients. However, in order to apply the AI model to the prediction of hip fracture risk in clinical situations, it is necessary to identify the characteristics of the dataset and AI model and use it after performing appropriate validation.\u003c/p\u003e","manuscriptTitle":"Effect of Artificial Intelligence or Machine Learning on Prediction of Hip Fracture Risk: Systematic Review","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2022-12-29 18:23:08","doi":"10.21203/rs.3.rs-2351785/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"5a6d512d-7b96-452e-b62e-82f1650324cc","owner":[],"postedDate":"December 29th, 2022","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2023-04-18T11:14:30+00:00","versionOfRecord":[],"versionCreatedAt":"2022-12-29 18:23:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-2351785","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-2351785","identity":"rs-2351785","version":["v1"]},"buildId":"WrCJVZZCHTDjtuVLN7oU0","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}