Abstract
Predicting post-discharge complications in cardiothoracic surgery is of utmost importance to improve clinical outcomes. Machine Learning (ML) techniques have been successfully applied in similar tasks, aiming at short time windows and in specific surgical conditions. However, as the target horizon is extended and the impact of unpredictable external factors rises, the complexity of the task increases, and traditional predictive models struggle to reproduce good performances. This study presents a two-step hybrid learning methodology to address this problem. Building up from identifying unique sub-groups of patients with shared characteristics, we then train individual supervised classification models for each sub-group, aiming at improved prediction accuracy and a more granular understanding of each decision. Our results show that specific sub-groups demonstrate substantially better performance when compared to the baseline model without sub-divisions, while others do not benefit from specialised models. Strategies such as the one presented may catalyse the success of applied ML solutions by contributing to a better understanding of their behaviour in different regions of the data space, leading to an informed decision-making process.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
References
Allyn, J., Allou, N., Augustin, P., Philip, I., Martinet, O., Belghiti, M., Provenchère, S., Montravers, P., Ferdynus, C.: A comparison of a machine learning model with Euroscore II in predicting mortality after elective cardiac surgery: a decision curve analysis. Plos One 12 (2017)
Benedetto, U., Dimagli, A., Sinha, S., Cocomello, L., Gibbison, B., Caputo, M., Gaunt, T.R., Lyon, M., Holmes, C.C., Angelini, G.D.: Machine learning improves mortality risk prediction after cardiac surgery: systematic review and meta-analysis. J. Thorac. Cardiovasc. Surg. 163(6) (2020)
Bertsimas, D., Zhuo, D., Dunn, J., Levine, J., Zuccarelli, E., Smyrnakis, N., Tobota, Z., Maruszewski, B., Fragata, J., Sarris, G.E.: Adverse outcomes prediction for congenital heart surgery: a machine learning approach. World J. Pediatric Congenit. Heart Surg. 12, 453–460 (2021)
Breiman, L.: Random forests. Mach. Learn. 45, 5–32 (2001)
Campello, R.J., Moulavi, D., Sander, J.: Density-based clustering based on hierarchical density estimates. In: Advances in Knowledge Discovery and Data Mining: 17th Pacific-Asia Conference, PAKDD 2013, Gold Coast, Australia, Proceedings, Part II 17, pp. 160–172. Springer (2013)
Caruso, E., Zadra, A.R.: The trade-off between costs and outcome after cardiac surgery. evidence from an Italian administrative registry. Health Policy 124(12), 1345–1353 (2020)
Cerqueira, V., Torgo, L., Branco, P., Bellinger, C.: Automated imbalanced classification via layered learning. Mach. Learn. 1–22 (2022)
Curioso, I., Santos, R., Ribeiro, B., Carreiro, A., Coelho, P., Fragata, J., Gamboa, H.: Addressing the curse of missing data in clinical contexts: a novel approach to correlation-based imputation. J. King Saud Univ.-Comput. Inf. Sci. 35(6), 101562 (2023)
Efthymiou, C.A., O’regan, D.J.: Postdischarge complications: what exactly happens when the patient goes home? Interact. Cardiovasc. Thorac. Surg. 12(2), 130–134 (2011)
Fan, Y., Dong, J., Wu, Y., Shen, M., Zhu, S., He, X., Jiang, S., Shao, J., Song, C.: Development of machine learning models for mortality risk prediction after cardiac surgery. Cardiovasc. Diagnos. Therapy 12(1), 12–23 (2022)
Fränti, P., Sieranoja, S., Wikström, K., Laatikainen, T.: Clustering diagnoses from 58 million patient visits in Finland between 2015 and 2018. JMIR Med. Inform. 10(5), e35422 (2022)
Fry, D.E., Pine, M.B., Nedza, S.M., Locke, B.D.G., Reband, B.A.M., Ba, Pine, G.: Inpatient and 90-day postdischarge outcomes in cardiac surgery. Am. J. Manag. Care 4 (2016)
Gordon, M.M., Moser, A.M., Rubin, E.: Unsupervised analysis of classical biomedical markers: robustness and medical relevance of patient clustering using bioinformatics tools. Plos One 7 (2012)
Head, S.J., Howell, N.J., Osnabrugge, R.L., Bridgewater, B., Keogh, B.E., Kinsman, R., Walton, P., Gummert, J.F., Pagano, D., Kappetein, A.P.: The European association for cardio-thoracic surgery (EACTS) database: an introduction. Eur. J. Cardiothorac. Surg. 44(3), e175–e180 (2013)
Jawitz, O.K., Gulack, B.C., Brennan, J.M., Thibault, D.P., Wang, A., O’Brien, S.M., Schroder, J.N., Gaca, J.G., Smith, P.K.: Association of postoperative complications and outcomes following coronary artery bypass grafting. Am. Heart J. 222, 220–228 (2020)
Kaushik, K., Kapoor, D., Varadharajan, V., Nallusamy, R.: Disease management: clustering-based disease prediction. Int. J. Collabor. Enter. 4(1–2), 69–82 (2014)
Khoury, H., Ragalie, W.S., Sanaiha, Y., Boutros, H., Rudasill, S.E., Shemin, R.J., Benharash, P.: Readmission following surgical aortic valve replacement in the United States. Ann. Thorac. Surg. 110(3), 849–855 (2020)
Kortlandt, F.A., van ’t Klooster, C.C., Bakker, A., Swaans, M.J., Kelder, J.C., de Kroon, T.L., Rensing, B.J., Eefting, F.D., van der Heyden, J.A., Post, M.C.: The predictive value of conventional surgical risk scores for periprocedural mortality in percutaneous mitral valve repair. Netherlands Heart J. 24, 475–480 (2016)
McInnes, L., Healy, J., Saul, N., Grossberger, L.: Umap: uniform manifold approximation and projection. J. Open Sour. Softw. 3(29), 861 (2018)
Mortazavi, B., Desai, N.R., Zhang, J., Coppi, A., Warner, F., Krumholz, H.M., Negahban, S.N.: Prediction of adverse events in patients undergoing major cardiovascular procedures. IEEE J. Biomed. Health Inform. 21, 1719–1729 (2017)
Nashef, S.A.M., Roques, F., Sharples, L.D., Nilsson, J., Smith, C., Goldstone, A.R., Lockowandt, U.: Euroscore II. Eur. J. Cardiothorac. Surg.: Official J. Eur. Assoc. Cardio-thorac. Surg. 41(4), 734–44 (2012)
Pudil, P., Novovicová, J., Kittler, J.: Floating search methods in feature selection. Pattern Recognit. Lett. 15, 1119–1125 (1994)
Sanchez, C.E., Hermiller, J.B., Pinto, D.S., Chetcuti, S.J., Arshi, A., Forrest, J.K., Huang, J., Yakubov, S.J.: Predictors and risk calculator of early unplanned hospital readmission following contemporary self-expanding transcatheter aortic valve replacement from the STS/ACC TVT-registry. Cardiovasc. Revascularization Med.: Incl. Mol. Interv. 21(3), 263–270 (2020)
Seese, L.M., Sultan, I.S., Gleason, T.G., Navid, F., Wang, Y., Thoma, F.W., Kilic, A.: The impact of major postoperative complications on long-term survival after cardiac surgery. Ann. Thorac. Surg. 110(1), 128–135 (2019)
Shahian, D.M., Jacobs, J.P., Badhwar, V., Kurlansky, P.A., Furnary, A.P., Cleveland, J.C., Lobdell, K.W., Vassileva, C.M., von Ballmoos, M.C.W., Thourani, V.H., Rankin, J.S., Edgerton, J.R., D’Agostino, R.S., Desai, N.D., Feng, L., He, X., O’Brien, S.M.: The society of thoracic surgeons 2018 adult cardiac surgery risk models: Part 1-background, design considerations, and model development. Ann. Thorac. Surg. 105(5), 1411–1418 (2018)
Silaschi, M., Conradi, L., Seiffert, M., Schnabel, R.B., Schön, G., Blankenberg, S., Reichenspurner, H.C., Diemert, P., Treede, H.: Predicting risk in transcatheter aortic valve implantation: comparative analysis of Euroscore II and established risk stratification tools. Thorac. Cardiovasc. Surg. 63, 472–478 (2014)
Sinha, S., Dimagli, A., Dixon, L., Gaudino, M.F., Caputo, M., Vohra, H.A., Angelini, G.D., Benedetto, U.: Systematic review and meta-analysis of mortality risk prediction models in adult cardiac surgery. Interact. Cardiovasc. Thorac. Surg. 33, 673–686 (2021)
Wang, C., Jin, L., Qiao, F., Xue, Q., Zhang, X., Han, L.: Performance of the society of thoracic surgeons 2008 cardiac risk models for major postoperative complications after heart valve surgery in a Chinese population: a multicenter study. Heart Surg. Forum 21(4), E281–E285 (2018)
Wang, T.K.M., Choi, D.H.M., Haydock, D.A., Gamble, G.D., Stewart, R.A., Ruygrok, P.N.: Comparison of risk scores for prediction of complications following aortic valve replacement. Heart Lung Circul. 24(6), 595–601 (2015)
Zhong, Z., Yuan, X., Liu, S., Yang, Y., Liu, F.: Machine learning prediction models for prognosis of critically ill patients after open-heart surgery. Sci. Rep. 11 (2021)
Acknowledgements
This work was conducted under the project “CardioFollow.AI: An intelligent system to improve patients’ safety and remote surveillance in follow-up for cardiothoracic surgery”, supported by national funds through ‘FCT—Portuguese Foundation for Science and Technology, I.P.’, with the reference DSAIPA/AI/0094/2020.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Ribeiro, B. et al. (2023). Unravelling Heterogeneity: A Hybrid Machine Learning Approach to Predict Post-discharge Complications in Cardiothoracic Surgery. In: Moniz, N., Vale, Z., Cascalho, J., Silva, C., Sebastião, R. (eds) Progress in Artificial Intelligence. EPIA 2023. Lecture Notes in Computer Science(), vol 14116. Springer, Cham. https://doi.org/10.1007/978-3-031-49011-8_24
Download citation
DOI: https://doi.org/10.1007/978-3-031-49011-8_24
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-49010-1
Online ISBN: 978-3-031-49011-8
eBook Packages: Computer ScienceComputer Science (R0)