Academic Journal
Development and Temporal External Validation of a Parsimonious, Interpretable Machine Learning Model for Predicting 6-Month Mortality in Long-Term Care Facilities: A Retrospective Cohort Study.
| Τίτλος: | Development and Temporal External Validation of a Parsimonious, Interpretable Machine Learning Model for Predicting 6-Month Mortality in Long-Term Care Facilities: A Retrospective Cohort Study. |
|---|---|
| Συγγραφείς: | Tsai YC; PecuLab LLC, Seattle, WA, United States., Wu YL; Department of Nursing, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan., Yu YC; Department of Nursing, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan., Chen YW; Department of Nursing, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan.; Department of Nursing, College of Medicine, National Sun Yat-sen University, Kaohsiung, Taiwan., Chen YW; Department of Business Development, Smart Ageing Tech Co., Ltd., New Taipei City, Taiwan., Yang YC; Department of Family Medicine, College of Medicine, National Cheng Kung University, Tainan, Taiwan.; Department of Geriatrics and Gerontology, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, No.138, Sheng Li Road, Tainan, 704, Taiwan, 886 6-2353535 ext 3828, 886 6-2083714., Lo YT; Department of Geriatrics and Gerontology, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, No.138, Sheng Li Road, Tainan, 704, Taiwan, 886 6-2353535 ext 3828, 886 6-2083714.; Department of Public Health, College of Medicine, National Cheng Kung University, Tainan, Taiwan. |
| Πηγή: | JMIR aging [JMIR Aging] 2026 Sep 16; Vol. 9, pp. e94567. Date of Electronic Publication: 2026 Sep 16. |
| Τύπος έκδοσης: | Journal Article; Validation Study |
| Γλώσσα: | English |
| Στοιχεία περιοδικού: | Publisher: JMIR Publications Inc Country of Publication: Canada NLM ID: 101740387 Publication Model: Electronic Cited Medium: Internet ISSN: 2561-7605 (Electronic) Linking ISSN: 25617605 NLM ISO Abbreviation: JMIR Aging Subsets: MEDLINE |
| Imprint Name(s): | Original Publication: [Toronto, ON] : JMIR Publications Inc., [2018]- |
| Ιατρικοί όροι (MeSH): | Long-Term Care*/statistics & numerical data , Mortality*/trends , Machine Learning*, Taiwan/epidemiology ; Humans ; Retrospective Studies ; Predictive Learning Models ; Male ; Aged ; Female ; Aged, 80 and over ; Boosting Machine Learning Algorithms ; Nursing Home Residents ; Random Forest |
| Περίληψη: | Background: Early mortality after long-term care facility (LTCF) admission is common; yet, prognostic tools are often derived from Western minimum dataset-based cohorts or require hospital electronic health record linkages that are unavailable at intake in many LTCFs. There is also limited evidence on explainable, admission-feasible machine learning-based prognostication in Asian LTCF settings. Objective: The aim of the study is to develop and temporally externally validate an interpretable machine learning model for predicting 6-month all-cause mortality among older adults newly admitted to LTCFs in Taiwan using routinely collected LTCF assessment data. Methods: We conducted a retrospective cohort study using the JUBO Long-Term Care Database, a nationwide private administrative registry covering 636 LTCFs (37.4% of the national LTCFs) in Taiwan. We included residents with first-time LTCF admission and prespecified nonoverlapping cohorts for temporal validation: development (January 1, 2020, to December 31, 2023; n=23,901) and external validation (January 1 to December 31, 2024; n=6216). The outcome measure was death within 180 days of admission. We compared a nonlinear ensemble model (hybrid of extreme gradient boosting and random forest [HybridXGBRF]) with 7 other algorithms, including tree-based and linear benchmarks. Discrimination (area under the receiver operating characteristic curve [AUROC]), classification metrics (accuracy, precision, recall, and F Results: In the development cohort, 5272 of 23,901 (22.1%) residents died within 180 days. In the 2024 temporal validation cohort, 1781 of 6216 (28.7%) residents died. In internal cross-validation, HybridXGBRF had the highest AUROC among the evaluated models (0.89, 95% CI 0.88-0.89). In temporal validation, HybridXGBRF maintained strong discrimination (AUROC 0.90, 95% CI 0.89-0.91), with an accuracy of 0.85 and an F Conclusions: An interpretable machine learning model using routinely collected Taiwanese LTCF assessment data achieved strong discrimination, acceptable calibration, and stable temporal validation performance without requiring hospital-based electronic health record linkage. The HybridXGBRF had the highest AUROC among the evaluated models, but performance differences from extreme gradient boosting were small. The model may serve as a risk-stratification tool to help identify residents who could benefit from structured care review, advance care planning, or palliative care assessment. Prospective implementation studies could determine its impact on care processes and resident- or family-centered outcomes. (© Yun-Cheng Tsai, Yi-Lin Wu, Yung-Chen Yu, Yi-Wen Chen, Yi-Wen Chen, Yi-Ching Yang, Yu-Tai Lo. Originally published in JMIR Aging (https://aging.jmir.org).) |
| References: | Nat Mach Intell. 2020 Jan;2(1):56-67. (PMID: 32607472) BMJ. 2024 Apr 16;385:e078378. (PMID: 38626948) EGEMS (Wash DC). 2013 Dec 17;1(3):1035. (PMID: 25848578) Med Care. 2013 Aug;51(8):673-81. (PMID: 23703648) J Am Med Dir Assoc. 2024 Sep;25(9):105142. (PMID: 38986685) J Am Med Dir Assoc. 2015 Jun 1;16(6):470-4. (PMID: 25659622) Comput Methods Programs Biomed. 2025 May;263:108653. (PMID: 39970690) J Am Med Dir Assoc. 2022 Aug;23(8):1389-1395.e4. (PMID: 35218731) J Am Geriatr Soc. 2021 Apr;69(4):938-945. (PMID: 33155268) JAMA Netw Open. 2023 Apr 3;6(4):e238795. (PMID: 37071421) Arch Gerontol Geriatr. 2012 Jul-Aug;55(1):138-42. (PMID: 21764150) Md State Med J. 1965 Feb;14:61-5. (PMID: 14258950) Arch Gerontol Geriatr. 2012 Jul-Aug;55(1):22-4. (PMID: 21764151) J Clin Nurs. 2008 Jan;17(1):99-108. (PMID: 18088262) J Am Geriatr Soc. 2003 Feb;51(2):213-21. (PMID: 12558718) EClinicalMedicine. 2022 Aug 11;52:101610. (PMID: 36034409) BMC Med Res Methodol. 2022 May 28;22(1):154. (PMID: 35643437) Arch Intern Med. 2008 Feb 25;168(4):382-9. (PMID: 18299493) Age Ageing. 2024 May 1;53(5):. (PMID: 38773946) Am Fam Physician. 2021 Jul 1;104(1):34-40. (PMID: 34264616) JMIR Med Inform. 2025 Nov 14;13:e79307. (PMID: 41237368) J Am Geriatr Soc. 2019 Apr;67(4):665-673. (PMID: 30663782) J Am Geriatr Soc. 2018 May;66(5):976-981. (PMID: 29500822) Med Decis Making. 2006 Nov-Dec;26(6):565-74. (PMID: 17099194) Clin Gerontol. 2021 Mar-Apr;44(2):83-96. (PMID: 33164674) BMC Res Notes. 2010 Jul 16;3:200. (PMID: 20637076) JAMA. 2010 Nov 3;304(17):1929-35. (PMID: 21045099) J Am Med Dir Assoc. 2022 Aug;23(8):1403-1408. (PMID: 35227666) J Aging Health. 2018 Apr;30(4):584-604. (PMID: 28553803) Health Sci Rep. 2023 Jun 01;6(6):e1309. (PMID: 37275670) JAMA. 2004 Jun 9;291(22):2734-40. (PMID: 15187055) Int Rev Psychiatry. 2025 Feb-Mar;37(2):142-156. (PMID: 40627515) JAMA. 2012 Jan 11;307(2):182-92. (PMID: 22235089) Health Policy. 2019 Oct;123(10):912-916. (PMID: 31455563) Acad Emerg Med. 2011 Oct;18(10):1099-104. (PMID: 21996075) J Nutr Health Aging. 2009 Nov;13(9):782-8. (PMID: 19812868) J Am Geriatr Soc. 2018 Dec;66(12):2353-2359. (PMID: 30335184) J Palliat Med. 2015 Feb;18(2):100-6. (PMID: 25380219) Int J Environ Res Public Health. 2020 Aug 27;17(17):. (PMID: 32867250) J Am Med Dir Assoc. 2022 Nov;23(11):1800-1806. (PMID: 35760091) J Am Geriatr Soc. 2003 Jan;51(1):96-100. (PMID: 12534853) J Am Med Dir Assoc. 2013 Sep;14(9):649-55. (PMID: 23639716) Adm Policy Ment Health. 2011 Mar;38(2):65-76. (PMID: 20957426) Lancet Neurol. 2014 Aug;13(8):844-54. (PMID: 25030516) N Engl J Med. 2023 Aug 31;389(9):833-838. (PMID: 37646680) J Am Geriatr Soc. 2008 Oct;56(10):1926-31. (PMID: 18811613) J Gerontol A Biol Sci Med Sci. 2005 Apr;60(4):491-8. (PMID: 15933390) |
| Contributed Indexing: | Keywords: AI; Taiwan; long-term care facilities; machine learning; mortality prediction |
| Entry Date(s): | Date Created: 20260916 Date Completed: 20260916 Latest Revision: 20260918 |
| Update Code: | 20260918 |
| PubMed Central ID: | PMC13581260 |
| DOI: | 10.2196/94567 |
| PMID: | 42748398 |
| Βάση Δεδομένων: | MEDLINE |
| ISSN: | 2561-7605 |
|---|---|
| DOI: | 10.2196/94567 |