OpenLB
Open Library of Bioscience
Advanced Search
iScience
Jul 19, 2024;27 (7): 110281.

Explainable machine learning for predicting 30-day readmission in acute heart failure patients.

Yang Zhang, Tianyu Xiang, Yanqing Wang, Tingting Shu, Chengliang Yin, Huan Li, Minjie Duan, Mengyan Sun, Binyi Zhao, Kaisaierjiang Kadier, Qian Xu, Tao Ling, Fanqi Kong, Xiaozhu Liu
Author Information
  1. Yang Zhang: College of Medical Informatics, Chongqing Medical University, Chongqing, China.
  2. Tianyu Xiang: Information Center, The University-Town Hospital of Chongqing Medical University, Chongqing, China.
  3. Yanqing Wang: The First Clinical College,Chongqing Medical University, Chongqing 400016, China.
  4. Tingting Shu: Army Medical University (Third Military Medical University), Chongqing, China.
  5. Chengliang Yin: Faculty of Medicine, Macau University of Science and Technology, Macau 999078, China.
  6. Huan Li: Chongqing College of Electronic Engineering, Chongqing, China.
  7. Minjie Duan: College of Medical Informatics, Chongqing Medical University, Chongqing, China.
  8. Mengyan Sun: Harris Manchester College, Oxford, UK.
  9. Binyi Zhao: First Department of Medicine Medical Faculty Mannheim University Medical Centre Mannheim (UMM)University of Heidelberg, Mannheim, Germany.
  10. Kaisaierjiang Kadier: Department of Cardiology, First Affiliated Hospital of Xinjiang Medical University, ��r��mqi, China.
  11. Qian Xu: Collection Development Department of Library, Chongqing Medical University, Chongqing, China.
  12. Tao Ling: Department of Pharmacy, Suqian First Hospital, Suqian, China.
  13. Fanqi Kong: Department of Cardiology, The Third Affiliated Hospital of Wenzhou Medical University, Wenzhou, China.
  14. Xiaozhu Liu: Medical Data Science Academy, Chongqing Medical University, Chongqing, China.
PMID: 39040074 DOI: 10.1016/j.isci.2024.110281

Abstract

We aimed to develop a machine-learning based predictive model to identify 30-day readmission risk in Acute heart failure (AHF) patients. In this study 2232 patients hospitalized with AHF were included. The variance inflation factor value and 5-fold cross-validation were used to select vital clinical variables. Five machine learning algorithms with good performance were applied to develop models, and the discrimination ability was comprehensively evaluated by sensitivity, specificity, and area under the ROC curve (AUC). Prediction results were illustrated by SHapley Additive exPlanations (SHAP) values. Finally, the XGBoost model performs optimally: the greatest AUC of 0.763 (0.703-0.824), highest sensitivity of 0.660, and high accuracy of 0.709. This study developed an optimal XGBoost model to predict the risk of 30-day unplanned readmission for AHF patients, which showed more significant performance compared with traditional logistic regression (LR) model.

Keywords

bioinformatics cardiovascular medicine

References

  1. Heart Fail Rev. 2021 Jan;26(1):57-64 [PMID: 31897907]
  2. Contemp Nurse. 2019 Feb;55(1):15-26 [PMID: 30632457]
  3. Circ Cardiovasc Qual Outcomes. 2008 Sep;1(1):29-37 [PMID: 20031785]
  4. IEEE Trans Neural Netw. 1991;2(2):318-22 [PMID: 18276387]
  5. N Engl J Med. 2013 Jan 10;368(2):100-2 [PMID: 23301730]
  6. Heart Fail Rev. 2019 Mar;24(2):177-187 [PMID: 30488242]
  7. Chest. 2016 Apr;149(4):1021-9 [PMID: 26836926]
  8. Circ J. 2013;77(9):2209-17 [PMID: 23955345]
  9. J Crit Care. 2019 Dec;54:110-116 [PMID: 31408805]
  10. Circulation. 2017 Mar 7;135(10):e146-e603 [PMID: 28122885]
  11. J Card Fail. 2018 Nov;24(11):783-792 [PMID: 30217774]
  12. JACC Heart Fail. 2017 May;5(5):359-366 [PMID: 28285121]
  13. Eur J Heart Fail. 2019 Nov;21(11):1338-1352 [PMID: 31127678]
  14. J Am Coll Cardiol. 2014 Apr 1;63(12):1123-1133 [PMID: 24491689]
  15. ESC Heart Fail. 2021 Feb;8(1):390-398 [PMID: 33232585]
  16. Int J Stroke. 2018 Oct;13(7):717-724 [PMID: 29693505]
  17. Eur Heart J. 2006 Nov;27(22):2725-36 [PMID: 17000631]
  18. Eur Heart J. 2016 Jul 14;37(27):2129-2200 [PMID: 27206819]
  19. JAMA. 2002 Mar 27;287(12):1541-7 [PMID: 11911756]
  20. ESC Heart Fail. 2020 Oct;7(5):3151-3160 [PMID: 32696559]
  21. Circulation. 2005 Dec 20;112(25):3958-68 [PMID: 16365214]
  22. Eur J Heart Fail. 2017 Oct;19(10):1242-1254 [PMID: 28463462]
  23. Clin Cardiol. 2021 Feb;44(2):230-237 [PMID: 33355945]
  24. Eur Heart J. 2015 Aug 7;36(30):1958-66 [PMID: 25998514]
  25. Popul Health Manag. 2013 Apr;16(2):125-31 [PMID: 23113632]
  26. J Transl Med. 2020 Dec 7;18(1):462 [PMID: 33287854]
  27. J Clin Nurs. 2020 Feb;29(3-4):320-329 [PMID: 31698508]
  28. Medicine (Baltimore). 2019 Jul;98(27):e16233 [PMID: 31277137]
  29. N Engl J Med. 2016 Apr 21;374(16):1543-51 [PMID: 26910198]
  30. Eur J Heart Fail. 2016 Jun;18(6):613-25 [PMID: 27324686]
  31. Contemp Nurse. 2017 Feb;53(1):105-120 [PMID: 28151071]
  32. Int J Cardiol. 2017 Apr 1;232:208-215 [PMID: 28087182]
  33. Eur Heart J. 2013 Oct;34(40):3151-9 [PMID: 24048728]
  34. Nat Rev Dis Primers. 2020 Mar 5;6(1):16 [PMID: 32139695]
  35. Diagnostics (Basel). 2021 Jan 29;11(2): [PMID: 33572844]
  36. J Card Fail. 2015 Nov;21(11):885-91 [PMID: 26209002]
  37. JAMA. 2016 Dec 27;316(24):2647-2656 [PMID: 28027367]
  38. JAMA. 2013 Jan 23;309(4):355-63 [PMID: 23340637]
Journal Article

Word Cloud

Created with Highcharts 10.0.0modelpatients030-dayreadmissionAHFdevelopriskheartfailurestudymachinelearningperformancesensitivityAUCXGBoostaimedmachine-learningbasedpredictiveidentifyAcute2232hospitalizedincludedvarianceinflationfactorvalue5-foldcross-validationusedselectvitalclinicalvariablesFivealgorithmsgoodappliedmodelsdiscriminationabilitycomprehensivelyevaluatedspecificityareaROCcurvePredictionresultsillustratedSHapleyAdditiveexPlanationsSHAPvaluesFinallyperformsoptimally:greatest763703-0824highest660highaccuracy709developedoptimalpredictunplannedshowedsignificantcomparedtraditionallogisticregressionLRExplainablepredictingacutebioinformaticscardiovascularmedicine

Similar Articles

Cited By