Academic Journal
View in EDS

Interpretable LightGBM model for predicting postoperative gastrointestinal hemorrhage in elderly hip fracture patients: leveraging systemic inflammation and medication exposures for personalized risk stratification.

Bibliographic Details
Title: Interpretable LightGBM model for predicting postoperative gastrointestinal hemorrhage in elderly hip fracture patients: leveraging systemic inflammation and medication exposures for personalized risk stratification.
Authors: Sai H; Department of Orthopaedic Surgery, Hebei Medical University Third Hospital, No 139 Ziqiang Road, Shijiazhuang, Hebei, 050051, P. R. China., Ren J; Department of Orthopaedic Surgery, Hebei Medical University Third Hospital, No 139 Ziqiang Road, Shijiazhuang, Hebei, 050051, P. R. China., Ma X; Department of Orthopaedic Surgery, Hebei Medical University Third Hospital, No 139 Ziqiang Road, Shijiazhuang, Hebei, 050051, P. R. China., Yang K; Department of Orthopaedic Surgery, Hebei Medical University Third Hospital, No 139 Ziqiang Road, Shijiazhuang, Hebei, 050051, P. R. China., Bai C; Department of Orthopaedic Surgery, Hebei Medical University Third Hospital, No 139 Ziqiang Road, Shijiazhuang, Hebei, 050051, P. R. China., Liu M; Department of Orthopaedic Surgery, Hebei Medical University Third Hospital, No 139 Ziqiang Road, Shijiazhuang, Hebei, 050051, P. R. China., Zhao F; Department of Orthopaedic Surgery, Hebei Medical University Third Hospital, No 139 Ziqiang Road, Shijiazhuang, Hebei, 050051, P. R. China., Chen H; Joint Knee Protection Center, Hebei Medical University Third Hospital, Shijiazhuang, China., Li X; Department of Orthopaedic Surgery, Hebei Medical University Third Hospital, No 139 Ziqiang Road, Shijiazhuang, Hebei, 050051, P. R. China. 13363889051@126.com.
Source: BMC geriatrics [BMC Geriatr] 2026 May 02; Vol. 26 (1). Date of Electronic Publication: 2026 May 02.
Publication Type: Journal Article
Language: English
Journal Info: Publisher: BioMed Central Country of Publication: England NLM ID: 100968548 Publication Model: Electronic Cited Medium: Internet ISSN: 1471-2318 (Electronic) Linking ISSN: 14712318 NLM ISO Abbreviation: BMC Geriatr Subsets: MEDLINE
Imprint Name(s): Original Publication: London : BioMed Central, [2001]-
MeSH Terms: Gastrointestinal Hemorrhage*/etiology , Gastrointestinal Hemorrhage*/diagnosis , Gastrointestinal Hemorrhage*/epidemiology , Hip Fractures*/surgery , Postoperative Hemorrhage*/diagnosis , Postoperative Hemorrhage*/epidemiology , Postoperative Hemorrhage*/etiology , Boosting Machine Learning Algorithms* , Inflammation*, Anti-Inflammatory Agents, Non-Steroidal/adverse effects ; Risk Assessment/methods ; Aged ; Aged, 80 and over ; Female ; Humans ; Male ; Retrospective Studies ; Risk Factors
Abstract: Background: Postoperative gastrointestinal (GI) bleeding is a serious complication after hip fracture surgery in older adults, yet perioperative risk stratification remains limited because commonly used GI-bleeding scores are not tailored to orthopedic settings. This study aimed to develop and internally validate an interpretable model to predict postoperative GI bleeding risk in elderly hip fracture patients, using data routinely available during the perioperative period.
Methods: We retrospectively included 342 elderly patients who underwent hip fracture surgery at the Third Hospital of Hebei Medical University from January to December 2023. The outcome was GI bleeding within 1 month after surgery, confirmed by medical records and/or telephone follow-up. Patients were randomly split into a training set (n = 242) and a validation set (n = 100). Predictors were screened using LASSO with 10-fold cross-validation, followed by multivariable logistic regression to identify independent risk factors. Ten prediction algorithms were trained and compared. Model performance was assessed by AUC, calibration, and decision curve analysis, and interpretability was evaluated using SHAP.
Results: GI bleeding occurred in 38 patients (11.1%). Multivariable analysis identified four independent predictors: alcohol consumption history (OR 8.109, 95% CI 2.463-26.69), glucocorticoid use (OR 4.922, 95% CI 1.055-22.97), NSAID use (OR 6.851, 95% CI 1.811-25.915), and higher systemic immune-inflammation index (SII) (OR 1.001, 95% CI 1.000-1.002). Among the tested models, LightGBM showed the best overall performance, with AUCs of 0.843 (training) and 0.817 (validation), good calibration, and the highest net benefit on decision curve analysis. SHAP results ranked feature importance as SII, NSAID use, alcohol consumption history, and glucocorticoid use, consistent with regression findings.
Conclusions: We developed and validated an interpretable LightGBM model that predicts postoperative GI bleeding risk in elderly hip fracture patients using routinely available clinical data. The final model incorporates only preoperative variables, systemic inflammation, NSAID use, alcohol history, and glucocorticoid use, supporting its application for early risk stratification prior to surgery.
(© 2026. The Author(s).)
Competing Interests: Declarations. Ethics approval and consent to participate: The Ethics Committee of the Third Hospital of Hebei Medical University approved the study protocol (approval number: KS2023-068-1; approval date: 1 August 2023). This study is a retrospective cohort study, exempt from informed consent. Consent for publication: Not applicable. Competing interests: The authors declare no competing interests.
References: Sci Rep. 2025 May 1;15(1):15289. (PMID: 40312416)
Ann R Coll Surg Engl. 2016 Jul;98(6):371-5. (PMID: 27055405)
Aliment Pharmacol Ther. 2007 Feb 1;25(3):297-308. (PMID: 17217452)
Sci Rep. 2019 Aug 12;9(1):11696. (PMID: 31406189)
J Pers Med. 2023 Aug 17;13(8):. (PMID: 37623521)
EFORT Open Rev. 2021 Aug 10;6(8):686-691. (PMID: 34532076)
Front Oncol. 2022 Nov 10;12:1019009. (PMID: 36439437)
Risk Manag Healthc Policy. 2025 May 11;18:1557-1568. (PMID: 40376449)
BMC Complement Med Ther. 2022 Feb 7;22(1):38. (PMID: 35130890)
Nutrients. 2022 Apr 05;14(7):. (PMID: 35406126)
Intern Med J. 2018 Jun;48(6):731-735. (PMID: 29898280)
Exp Ther Med. 2019 Oct;18(4):2648-2652. (PMID: 31572513)
Medicine (Baltimore). 2019 Jun;98(26):e16239. (PMID: 31261585)
BMC Anesthesiol. 2022 Apr 29;22(1):130. (PMID: 35488196)
Cureus. 2023 Oct 30;15(10):e47982. (PMID: 38034276)
Crit Care Med. 2015 Oct;43(10):2076-84. (PMID: 26110488)
J Orthop Surg Res. 2024 May 27;19(1):311. (PMID: 38802945)
Front Med (Lausanne). 2024 May 30;11:1408371. (PMID: 38873200)
Ann Transl Med. 2022 Feb;10(3):130. (PMID: 35284557)
Foods. 2020 May 03;9(5):. (PMID: 32375270)
Dermatol Pract Concept. 2024 Jan 01;14(1):. (PMID: 38364414)
J Orthop Surg Res. 2023 May 10;18(1):347. (PMID: 37165391)
Healthcare (Basel). 2023 Oct 30;11(21):. (PMID: 37957997)
BMJ. 2017 Jan 4;356:i6432. (PMID: 28053181)
Injury. 2024 Aug;55(8):111635. (PMID: 38852528)
Eur J Emerg Med. 2023 Feb 1;30(1):32-39. (PMID: 36542335)
Front Immunol. 2021 Jun 09;12:580989. (PMID: 34177879)
Front Pharmacol. 2024 Jan 08;14:1326224. (PMID: 38259284)
Contributed Indexing: Keywords: Elderly patients; Gastrointestinal bleeding; Hip fracture; Machine learning; Prediction model; Risk factors
Substance Nomenclature: 0 (Anti-Inflammatory Agents, Non-Steroidal)
Entry Date(s): Date Created: 20260502 Date Completed: 20260622 Latest Revision: 20260813
Update Code: 20260813
PubMed Central ID: PMC13285533
DOI: 10.1186/s12877-026-07607-3
PMID: 42070040
Database: MEDLINE
Description
ISSN:1471-2318
DOI:10.1186/s12877-026-07607-3