{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,1]],"date-time":"2026-06-01T05:01:19Z","timestamp":1780290079552,"version":"3.54.0"},"reference-count":42,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,6,1]],"date-time":"2026-06-01T00:00:00Z","timestamp":1780272000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Med."],"abstract":"<jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Hip fractures are associated with significant mortality, especially among elderly patients. Accurate prediction of mortality risk is crucial for optimising perioperative care and resource allocation. Recent advances in machine learning (ML) and deep learning (DL) offer promising methods to enhance clinical risk prediction models; however, their clinical implementation often remains limited due to the complexity of these techniques.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>This retrospective cohort study included 2,604 elderly patients (\u226565\u202fyears) undergoing urgent hip fracture surgery at Sheba Medical Center, Israel, between January 2017 and November 2023. Multiple ML and DL algorithms were evaluated for predicting one-year all-cause mortality using a comprehensive set of clinical, demographic, perioperative, and laboratory variables. Models were rigorously developed and validated through stratified 5-fold cross-validation, addressing class imbalance with the Synthetic Minority Oversampling Technique (SMOTE). Additionally, an automated ML pipeline, generated using a large language model (LLM) coupled with the Tree-based Pipeline Optimisation Tool (TPOT), was benchmarked against manually optimised models. Model performances were assessed using area under the receiver operating characteristic curve (AUC), accuracy, precision, recall, F1-score, false-positive rate, and true-negative rate, supplemented by permutation importance and SHapley Additive exPlanations (SHAP) for interpretability.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>Among all models evaluated, the manually optimised Extreme Gradient Boosting (XGB) algorithm demonstrated superior predictive performance (AUC\u202f=\u202f0.846, accuracy\u202f=\u202f0.791, F1-score\u202f=\u202f0.667, precision\u202f=\u202f0.773, NPV\u202f=\u202f0.798). Important predictors identified included baseline serum albumin and urea levels, patient age, intraoperative hypothermia, and the number of chronic diseases. The automated ML model, generated via LLM and TPOT frameworks, showed comparable performance to the XGB model (AUC\u202f=\u202f0.844), with a higher recall but slightly lower precision.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Discussion<\/jats:title>\n                    <jats:p>ML-based models, particularly the XGB algorithm, significantly enhance predictive accuracy for one-year mortality among elderly hip fracture patients. Crucially, an automated ML framework leveraging large language models provides a practical, clinically accessible alternative, effectively democratising advanced predictive analytics in healthcare settings.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fmed.2026.1804645","type":"journal-article","created":{"date-parts":[[2026,6,1]],"date-time":"2026-06-01T04:19:02Z","timestamp":1780287542000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Comparing manual vs. automated machine learning and deep learning models for predicting one-year mortality in elderly hip fracture patients"],"prefix":"10.3389","volume":"13","author":[{"given":"Adi","family":"Shuchami","sequence":"first","affiliation":[{"name":"Department of Mathematics, Ariel University","place":["Ariel, Israel"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Maxim","family":"Glebov","sequence":"additional","affiliation":[{"name":"Department of Anesthesiology, Sheba Medical Center","place":["Ramat Gan, Israel"]},{"name":"Department of Information Systems, University of Haifa","place":["Haifa, Israel"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Maksim","family":"Katsin","sequence":"additional","affiliation":[{"name":"Department of Anesthesiology, Sheba Medical Center","place":["Ramat Gan, Israel"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yotam","family":"Portnoy","sequence":"additional","affiliation":[{"name":"Faculty of Medicine, Tel-Aviv University","place":["Tel Aviv, Israel"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Haim","family":"Berkenstadt","sequence":"additional","affiliation":[{"name":"Department of Anesthesiology, Sheba Medical Center","place":["Ramat Gan, Israel"]},{"name":"Faculty of Medicine, Tel-Aviv University","place":["Tel Aviv, Israel"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dina","family":"Orkin","sequence":"additional","affiliation":[{"name":"Department of Anesthesiology, Sheba Medical Center","place":["Ramat Gan, Israel"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Teddy","family":"Lazebnik","sequence":"additional","affiliation":[{"name":"Department of Information Systems, University of Haifa","place":["Haifa, Israel"]},{"name":"Department of Computing, J\u00f6nk\u00f6ping University","place":["J\u00f6nk\u00f6ping, Sweden"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,6,1]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"407","DOI":"10.1007\/pl00004148","article-title":"World-wide projections for hip fracture","volume":"7","author":"Gullberg","year":"1997","journal-title":"Osteoporos Int"},{"key":"ref2","doi-asserted-by":"publisher","first-page":"e74586","DOI":"10.7759\/cureus.74586","article-title":"Outcome and cost evaluation of hip fractures in elderly patients at a tertiary Care Hospital in the Caribbean","volume":"16","author":"Sanderson-Jerome","year":"2024","journal-title":"Cureus"},{"key":"ref3","doi-asserted-by":"publisher","first-page":"1633","DOI":"10.1007\/s00198-009-0920-3","article-title":"Excess mortality following hip fracture: a systematic epidemiological review","volume":"20","author":"Abrahamsen","year":"2009","journal-title":"Osteoporos Int"},{"key":"ref4","doi-asserted-by":"publisher","first-page":"1159","DOI":"10.1016\/j.bja.2024.08.007","article-title":"Machine learning-augmented interventions in perioperative care: a systematic review and meta-analysis","volume":"133","author":"Mehta","year":"2024","journal-title":"Br J Anaesth"},{"key":"ref5","doi-asserted-by":"publisher","first-page":"1579","DOI":"10.1016\/S0140-6736(20)30226-9","article-title":"Artificial intelligence and the future of global health","volume":"395","author":"Schwalbe","year":"2020","journal-title":"Lancet"},{"key":"ref6","doi-asserted-by":"publisher","first-page":"291","DOI":"10.1016\/j.jamda.2020.09.025","article-title":"Machine learning algorithms to predict mortality and allocate palliative Care for Older Patients with hip Fracture","volume":"22","author":"Cary","year":"2021","journal-title":"J Am Med Dir Assoc"},{"key":"ref7","doi-asserted-by":"publisher","first-page":"1487","DOI":"10.1016\/j.injury.2020.12.008","article-title":"A novel machine-learning algorithm for predicting mortality risk after hip fracture surgery","volume":"52","author":"Li","year":"2021","journal-title":"Injury"},{"key":"ref8","doi-asserted-by":"publisher","first-page":"6787","DOI":"10.3390\/app10196787","article-title":"Development of a mortality risk model in elderly hip fracture patients by different analytical approaches","volume":"10","author":"Lo","year":"2020","journal-title":"Appl Sci"},{"key":"ref9","doi-asserted-by":"publisher","first-page":"m441","DOI":"10.1136\/bmj.m441","article-title":"Calculating the sample size required for developing a clinical prediction model","volume":"368","author":"Riley","year":"2020","journal-title":"BMJ"},{"key":"ref10","doi-asserted-by":"publisher","first-page":"137","DOI":"10.1007\/s11222-009-9153-8","article-title":"Estimation of prediction error by using K-fold cross-validation","volume":"21","author":"Fushiki","year":"2011","journal-title":"Stat Comput"},{"key":"ref11","doi-asserted-by":"publisher","first-page":"160","DOI":"10.1007\/s10916-014-0160-8","article-title":"The nurse scheduling problem in real-life","volume":"39","author":"Legrain","year":"2015","journal-title":"J Med Syst"},{"key":"ref12","volume-title":"A Brief Study of the Nurse Scheduling Problem (NSP)","author":"Augustine","year":"2009"},{"key":"ref13","doi-asserted-by":"publisher","first-page":"6563498","DOI":"10.1155\/2017\/6563498","article-title":"Directed bee colony optimization algorithm to solve the nurse rostering problem","volume":"2017","author":"Rajeswari","year":"2017","journal-title":"Comput Intell Neurosci"},{"key":"ref14","volume-title":"Logistic Regression","author":"Kleinbaum","year":"2002"},{"key":"ref15","first-page":"713","article-title":"Na\u00efve Bayes","volume":"15","author":"Webb","year":"2010","journal-title":"Encycl Mach Learn"},{"key":"ref16","doi-asserted-by":"publisher","first-page":"1883","DOI":"10.4249\/scholarpedia.1883","article-title":"K-Nearest Neighbor","volume":"4","author":"Peterson","year":"2009","journal-title":"Scholarpedia"},{"key":"ref17","doi-asserted-by":"crossref","first-page":"207","DOI":"10.1007\/978-1-4899-7641-3_9","volume-title":"Machine Learning Models and Algorithms for Big Data Classification","author":"Suthaharan","year":"2016"},{"key":"ref18","doi-asserted-by":"publisher","first-page":"102173","DOI":"10.1016\/j.datak.2023.102173","article-title":"Decision tree post-pruning without loss of accuracy using the SAT-PP algorithm with an empirical evaluation on clinical data","volume":"145","author":"Lazebnik","year":"2023","journal-title":"Data Knowl Eng"},{"key":"ref19","doi-asserted-by":"publisher","first-page":"19905","DOI":"10.1007\/s11042-019-7327-8","article-title":"A healthcare monitoring system using random forest and internet of things (IoT)","volume":"78","author":"Kaur","year":"2019","journal-title":"Multimed Tools Appl"},{"key":"ref20","first-page":"785","author":"Chen","year":"2016"},{"key":"ref21","first-page":"848","article-title":"Application of multilayer perceptron neural networks and support vector machines in classification of healthcare data","volume":"2016","author":"Naraei","year":"2016","journal-title":"IEEE FTC"},{"key":"ref22","first-page":"6679","article-title":"Tabnet: attentive interpretable tabular learning","volume":"35","author":"Arik","year":"2021","journal-title":"AAAI Conf Artif Intell"},{"key":"ref23","article-title":"Tabtransformer: Tabular data modeling using contextual embeddings","volume-title":"arXiv","author":"Huang","year":"2020"},{"key":"ref24","first-page":"1551","article-title":"Hyperparameter optimization. IEEE congress","volume":"2021","author":"Alibrahim","year":"2021","journal-title":"Evol Comput"},{"key":"ref25","doi-asserted-by":"publisher","first-page":"321","DOI":"10.1613\/jair.953","article-title":"SMOTE: synthetic minority over-sampling technique","volume":"16","author":"Chawla","year":"2002","journal-title":"J Artif Intell Res"},{"key":"ref26","doi-asserted-by":"publisher","first-page":"1340","DOI":"10.1093\/bioinformatics\/btq134","article-title":"Permutation importance","volume":"26","author":"Altmann","year":"2010","journal-title":"Bioinformatics"},{"key":"ref27","doi-asserted-by":"publisher","first-page":"106584","DOI":"10.1016\/j.cmpb.2021.106584","article-title":"Explanation of ML models using SHAP","volume":"214","author":"Nohara","year":"2022","journal-title":"Comput Methods Prog Biomed"},{"key":"ref28","volume-title":"LLM Attribution for GPT Models","author":"Rosenfeld","year":"2024"},{"key":"ref29","article-title":"ChatGPT-o1 in medical ethics","volume-title":"medRxiv","author":"Soffer","year":"2024"},{"key":"ref30","first-page":"66","author":"Olson","year":"2016"},{"key":"ref31","doi-asserted-by":"publisher","first-page":"11232","DOI":"10.3390\/su151411232","article-title":"Benchmarking biologically-inspired automatic machine learning for economic tasks","volume":"15","author":"Lazebnik","year":"2023","journal-title":"Sustainability"},{"key":"ref32","doi-asserted-by":"publisher","first-page":"686","DOI":"10.3390\/sym13040686","article-title":"A continuous cuffless blood pressure estimation using tree-based pipeline optimization tool","volume":"13","author":"Fati","year":"2021","journal-title":"Symmetry"},{"key":"ref33","article-title":"A prompt pattern catalog to enhance prompt engineering with Chatgpt","volume-title":"arXiv","author":"White","year":"2023"},{"key":"ref34","article-title":"A comprehensive benchmark of machine and deep learning across diverse tabular datasets","volume-title":"arXiv","author":"Shmuel","year":"2024"},{"key":"ref35","doi-asserted-by":"publisher","first-page":"847206","DOI":"10.3389\/fcvm.2022.847206","article-title":"Predicting ICU mortality in rheumatic heart disease: comparison of XGBoost and logistic regression","volume":"9","author":"Xu","year":"2022","journal-title":"Front Cardiovasc Med"},{"key":"ref36","doi-asserted-by":"publisher","first-page":"162","DOI":"10.1097\/00005131-199704000-00004","article-title":"Hip fractures in the elderly: predictors of one year mortality","volume":"11","author":"Aharonoff","year":"1997","journal-title":"J Orthop Trauma"},{"key":"ref37","doi-asserted-by":"publisher","first-page":"571","DOI":"10.1080\/00365513.2022.2150982","article-title":"Hypoalbuminemia is associated with 30-day mortality in hip fracture patients independently of body mass index","volume":"82","author":"Borge","year":"2022","journal-title":"Scand J Clin Lab Invest"},{"key":"ref38","doi-asserted-by":"publisher","first-page":"698","DOI":"10.1016\/j.injury.2006.04.121","article-title":"Mortality and serum urea and electrolytes on admission for hip fracture patients","volume":"37","author":"Lewis","year":"2006","journal-title":"Injury"},{"key":"ref39","doi-asserted-by":"publisher","first-page":"343","DOI":"10.1097\/BOT.0000000000002332","article-title":"Perioperative hypothermia is associated with increased 30-day mortality in hip fracture patients in the United Kingdom: \u03b1 systematic review and Meta-analysis","volume":"36","author":"Mroczek","year":"2022","journal-title":"J Orthop Trauma"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1145\/3767326","article-title":"Explainable surgical procedures recommender system leveraging large language models","author":"Solomon","year":"2025","journal-title":"ACM Trans Recom Syst"},{"key":"ref41","doi-asserted-by":"publisher","first-page":"1614","DOI":"10.1038\/s41598-026-35327-6","article-title":"Knowledge integration for physics-informed symbolic regression using pre-trained large language models","volume":"16","author":"Taskin","year":"2026","journal-title":"Sci Rep"},{"key":"ref42","first-page":"127","author":"Silva","year":"2025"}],"container-title":["Frontiers in Medicine"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2026.1804645\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,1]],"date-time":"2026-06-01T04:19:04Z","timestamp":1780287544000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2026.1804645\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,6,1]]},"references-count":42,"alternative-id":["10.3389\/fmed.2026.1804645"],"URL":"https:\/\/doi.org\/10.3389\/fmed.2026.1804645","relation":{},"ISSN":["2296-858X"],"issn-type":[{"value":"2296-858X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,6,1]]},"article-number":"1804645"}}