Development of Machine Learning Algorithms to Predict Clinically Meaningful Improvement for the Patient-Reported Health State After Total Hip Arthroplasty Kyle N. Kunze, BS a, * , Aditya V. Karhade, BE b , Alex J. Sadauskas, MD a , Joseph H. Schwab, MD, MS b , Brett R. Levine, MD, MS a a Department of Orthopaedic Surgery, Rush University Medical Center, Chicago, IL b Department of Orthopedic Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, MA article info Article history: Received 4 February 2020 Received in revised form 1 March 2020 Accepted 10 March 2020 Available online xxx Keywords: total hip arthroplasty THA machine learning clinical outcomes MCID clinically significant outcome abstract Background: Failure to achieve clinically significant outcome (CSO) improvement after total hip arthro- plasty (THA) imposes a potential cost-to-risk imbalance in the context of bundle payment models. Pa- tient perception of their health state is one component of such risk. The purpose of the current study is to develop machine learning algorithms to predict CSO for the patient-reported health state (PRHS) and build a clinical decision-making tool based on risk factors. Methods: A retrospective review of primary THA patients between 2014 and 2017 was performed. Variables considered for prediction included demographics, medical history, preoperative PRHS, and modified Harris Hip Score. The minimal clinically important difference (MCID) for the PRHS was calculated using a distribution-based method. Five supervised machine learning algorithms were developed and assessed by discrimination, calibration, Brier score, and decision curve analysis. Results: Of 616 patients, a total of 407 (69.2%) achieved the MCID for the PRHS. The random forest al- gorithm achieved the best performance in the independent testing set not used for algorithm devel- opment (c-statistic 0.97, calibration intercept À0.05, calibration slope 1.45, Brier score 0.054). The most important factors for achieving the MCID were preoperative PRHS, preoperative opioid use, age, and body mass index. Individual patient-level explanations were provided for the algorithm predictions and the algorithms were incorporated into an open access digital application available here: https://sorg- apps.shinyapps.io/THA_PRHS_mcid/. Conclusion: The current study created a clinical decision-making tool based on partially modifiable risk factors for predicting CSO after THA. The tool demonstrates excellent discriminative capacity for iden- tifying those at greatest risk for failing to achieve CSO in their current health state and may allow for preoperative health optimization. © 2020 Elsevier Inc. All rights reserved. The demand for total hip arthroplasty (THA) is estimated to grow by 174% to 572,000 annual procedures by the year 2030 [1]. Paralleling this exponential increase in THA procedures is a growing concern for unsustainable healthcare expenditures for the care regarding these procedures. Bundled payment care initiatives are now being employed for cost-control in anticipation of the overwhelming numbers and economic risk. To this end, increasing attention has been turned toward developing predictive models for identifying and optimizing patients who are at risk of readmission and incurring additional cost after surgery [2,3]. Further knowledge of the costs of care throughout the period of risk may offer an op- portunity for cost containment [4,5]. Current methods of understanding and measuring patient improvement after THA are highly dependent on patient-reported outcome measures (PROMs). Administration and interpretation of scores generated by such PROMs may allow for preoperative identification of patients at risk for readmission or adverse out- comes based on the potential need for more complex care or One or more of the authors of this paper have disclosed potential or pertinent conflicts of interest, which may include receipt of payment, either direct or indirect, institutional support, or association with an entity in the biomedical field which may be perceived to have potential conflict of interest with this work. For full disclosure statements refer to https://doi.org/10.1016/j.arth.2020.03.019. * Reprint requests: Kyle N. Kunze, BS, Department of Orthopaedic Surgery, Rush University Medical Center,1611 W. Harrison Street, Suite 300, Chicago, IL 60612. Contents lists available at ScienceDirect The Journal of Arthroplasty journal homepage: www.arthroplastyjournal.org https://doi.org/10.1016/j.arth.2020.03.019 0883-5403/© 2020 Elsevier Inc. All rights reserved. The Journal of Arthroplasty xxx (2020) 1e5