Original Study| Volume 22, ISSUE 2, P291-296, February 2021

Download started.


Machine Learning Algorithms to Predict Mortality and Allocate Palliative Care for Older Patients With Hip Fracture

Published:October 29, 2020DOI:



      To evaluate a machine learning model designed to predict mortality for Medicare beneficiaries aged >65 years treated for hip fracture in Inpatient Rehabilitation Facilities (IRFs).


      Retrospective design/cohort analysis of Centers for Medicare & Medicaid Services Inpatient Rehabilitation Facility–Patient Assessment Instrument data.

      Setting and Participants

      A total of 17,140 persons admitted to Medicare-certified IRFs in 2015 following hospitalization for hip fracture.


      Patient characteristics include sociodemographic (age, gender, race, and social support) and clinical factors (functional status at admission, chronic conditions) and IRF length of stay. Outcomes were 30-day and 1-year all-cause mortality. We trained and evaluated 2 classification models, logistic regression and a multilayer perceptron (MLP), to predict the probability of 30-day and 1-year mortality and evaluated the calibration, discrimination, and precision of the models.


      For 30-day mortality, MLP performed well [acc = 0.74, area under the receiver operating characteristic curve (AUROC) = 0.76, avg prec = 0.10, slope = 1.14] as did logistic regression (acc = 0.78, AUROC = 0.76, avg prec = 0.09, slope = 1.20). For 1-year mortality, the performances were similar for both MLP (acc = 0.68, AUROC = 0.75, avg prec = 0.32, slope = 0.96) and logistic regression (acc = 0.68, AUROC = 0.75, avg prec = 0.32, slope = 0.95).

      Conclusion and Implications

      A scoring system based on logistic regression may be more feasible to run in current electronic medical records. But MLP models may reduce cognitive burden and increase ability to calibrate to local data, yielding clinical specificity in mortality prediction so that palliative care resources may be allocated more effectively.


      To read this article in full you will need to make a payment

      Purchase one-time access:

      Academic & Personal: 24 hour online accessCorporate R&D Professionals: 24 hour online access
      One-time access price info
      • For academic or personal research use, select 'Academic and Personal'
      • For corporate R&D use, select 'Corporate R&D Professionals'


      Subscribe to Journal of the American Medical Directors Association
      Already a print subscriber? Claim online access
      Already an online subscriber? Sign in
      Institutional Access: Sign in to ScienceDirect


      1. National and regional estimates on hospital use for all patients from the HCUP Nationwide Inpatient Sample (NIS).
        (Available at:)
        Date accessed: February 6, 2020
        • Haentjens P.
        • Magaziner J.
        • Colon-Emeric C.S.
        • et al.
        Meta-analysis: Excess mortality after hip fracture among older women and men.
        Ann Intern Med. 2010; 152: 380-390
        • Chang W.
        • Lv H.
        • Feng C.
        • et al.
        Preventable risk factors of mortality after hip fracture surgery: Systematic review and meta-analysis.
        Int J Surg. 2018; 52: 320-328
        • Becker D.J.
        • Arora T.
        • Kilgore M.L.
        • et al.
        Trends in the utilization and outcomes of Medicare patients hospitalized for hip fracture, 2000-2008.
        J Aging Health. 2014; 26: 360-379
        • Ali A.M.
        • Gibbons C.E.
        Predictors of 30-day hospital readmission after hip fracture: A systematic review.
        Injury. 2017; 48: 243-252
        • Cenzer I.S.
        • Tang V.
        • Boscardin W.J.
        • et al.
        One-year mortality after hip fracture: Development and validation of a prognostic index.
        J Am Geriatr Soc. 2016; 64: 1863-1868
        • De Lima L.
        • Radbruch L.
        The international association for hospice and palliative care: Advancing hospice and palliative care worldwide.
        J Pain Symptom Manage. 2018; 55: S96-S103
        • Barawid E.
        • Covarrubias N.
        • Tribuzio B.
        • Liao S.
        The benefits of rehabilitation for palliative care patients.
        Am J Hosp Palliat Care. 2015; 32: 34-43
        • Kamal A.H.
        • Wolf S.P.
        • Troy J.
        • et al.
        Policy changes key to promoting sustainability and growth of the specialty palliative care workforce.
        Health Aff (Millwood). 2019; 38: 910-918
        • Lau F.
        • Downing G.M.
        • Lesperance M.
        • et al.
        Use of Palliative Performance Scale in end-of-life prognostication.
        J Palliat Med. 2006; 9: 1066-1075
        • Simmons C.P.L.
        • McMillan D.C.
        • McWilliams K.
        • et al.
        Prognostic tools in patients with advanced cancer: A systematic review.
        J Pain Symptom Manage. 2017; 53: 962-970.e10
        • Fischer S.M.
        • Gozansky W.S.
        • Sauaia A.
        • et al.
        A practical tool to identify patients who may benefit from a palliative approach: The CARING criteria.
        J Pain Symptom Manage. 2006; 31: 285-292
        • Yourman L.C.
        • Lee S.J.
        • Schonberg M.A.
        • et al.
        Prognostic indices for older adults: A systematic review.
        JAMA. 2012; 307: 182-192
        • Hosny A.
        • Parmar C.
        • Quackenbush J.
        • et al.
        Artificial intelligence in radiology.
        Nat Rev Cancer. 2018; 18: 500-510
        • Esteva A.
        • Kuprel B.
        • Novoa R.A.
        • et al.
        Dermatologist-level classification of skin cancer with deep neural networks.
        Nature. 2017; 542: 115-118
        • Mobadersany P.
        • Yousefi S.
        • Amgad M.
        • et al.
        Predicting cancer outcomes from histology and genomics using convolutional networks.
        Proc Natl Acad Sci U S A. 2018; 115: E2970-E2979
        • Karnuta J.M.
        • Navarro S.M.
        • Haeberle H.S.
        • et al.
        Bundled care for hip fractures: A machine-learning approach to an untenable patient-specific payment model.
        J Orthop Trauma. 2019; 33: 324-330
        • Lund J.L.
        • Kuo T.M.
        • Brookhart M.A.
        • et al.
        Development and validation of a 5-year mortality prediction model using regularized regression and Medicare data.
        Pharmacoepidemiol Drug Saf. 2019; 28: 584-592
        • Cary Jr., M.P.
        • Prvu Bettger J.
        • Jarvis J.M.
        • et al.
        Successful community discharge following postacute rehabilitation for Medicare beneficiaries: Analysis of a patient-centered quality measure.
        Health Serv Res. 2018; 53: 2470-2482
        • Ottenbacher K.J.
        • Hsu Y.
        • Granger C.V.
        • Fiedler R.C.
        The reliability of the functional independence measure: A quantitative review.
        Arch Phys Med Rehabil. 1996; 77: 1226-1232
        • Goodman R.A.
        • Ling S.M.
        • Briss P.A.
        • et al.
        Multimorbidity patterns in the United States: Implications for research and clinical practice.
        J Gerontol A Biol Sci Med Sci. 2016; 71: 215-220
        • Stineman M.G.
        • Shea J.A.
        • Jette A.
        • et al.
        The functional independence measure: Tests of scaling assumptions, structure, and reliability across 20 diverse impairment categories.
        Arch Phys Med Rehabil. 1996; 77: 1101-1108
        • Cohen I.
        • Goldszmidt M.
        Properties and Benefits of Calibrated Classifiers.
        Springer-Verlag, Berlin2004
        • Fluss R.
        • Faraggi D.
        • Reiser B.
        Estimation of the Youden Index and its associated cutoff point.
        Biometrical J Biometrische Zeitschrift. 2005; 47: 458-472
        • Pugely A.J.
        • Martin C.T.
        • Gao Y.
        • et al.
        A risk calculator for short-term morbidity and mortality after hip fracture surgery.
        J Orthop Trauma. 2014; 28: 63-69
        • Dodd A.C.
        • Bulka C.
        • Jahangir A.
        • et al.
        Predictors of 30-day mortality following hip/pelvis fractures.
        Orthop Traumatol Surg Res. 2016; 102: 707-710
        • Heyes G.J.
        • Tucker A.
        • Marley D.
        • Foster A.
        Predictors for 1-year mortality following hip fracture: A retrospective review of 465 consecutive patients.
        Eur J Trauma Emerg Surg. 2017; 43: 113-119
        • Novoa-Parra C.D.
        • Hurtado-Cerezo J.
        • Morales-Rodriguez J.
        • et al.
        Factors predicting one-year mortality of patients over 80 years operated after femoral neck fracture.
        Rev Esp Cir Ortop Traumatol. 2019; 63: 202-208
        • Beam A.L.
        • Kohane I.S.
        Big data and machine learning in health care.
        JAMA. 2018; 319: 1317-1318
        • Maxwell M.J.
        • Moran C.G.
        • Moppett I.K.
        Development and validation of a preoperative scoring system to predict 30 day mortality in patients undergoing hip fracture surgery.
        Br J Anaesth. 2008; 101: 511-517
        • Lee C.
        • Yoon J.
        • Schaar M.V.
        Dynamic-Deephit: A deep learning approach for dynamic survival analysis with competing risks based on longitudinal data.
        IEEE Trans Biomed Eng. 2020; 67: 122-133
        • Chapfuwa P.
        • Tao C.
        • Li C.
        • et al.
        (Proceedings of the 35th International Conference on Machine Learning, Stockholm, Sweden)Adversarial time-to-event modeling. 80. 2018: 735-744