Artificial Intelligence for Healthcare Interdisciplinary Partnerships for Analytics-driven Improvements in a Post-COVID World
Book contents
4 - Proceed with Care
Integrating Predictive Analytics with Patient Decision Making
from Part I - Personalized Medicine
Published online by Cambridge University Press: 21 April 2022
Summary
To address the challenge of limited training data for machine learning models in the healthcare domain, we advocate for human-in-the-loop machine learning, which involves domain experts in an inter- active process of developing predictive models. Interpretability offers a promising way to facilitate this interaction. We describe an approach that offers a simple decision tree interpretation for any complex blackbox machine learning model. In a case study with physicians, we find that they were able to use the interpretation to discover an unexpected causal issue in a personalized patient risk score trained on electronic medical record data. To account for dynamics in disease progression, we advocate for building decision models that integrate predictions of the disease progression at the individual patient level with system models capturing the dynamic operational environments. We describe a case study on hospital inpatient management, showing how to build a Markov decision framework that leverages predictive analytics on patient readmission risk and prescribes the optimal set of patients to be discharged each day.
Keywords
- Type
- Chapter
- Information
- Artificial Intelligence for HealthcareInterdisciplinary Partnerships for Analytics-driven Improvements in a Post-COVID World, pp. 60 - 80Publisher: Cambridge University PressPrint publication year: 2022
References
Bastani, Hamsa. Predicting with proxies: transfer learning in high dimension. Management Science, 67(5), 2020.Google Scholar
Nestor, Bret, McDermott, Matthew, Boag, Willie, et al. Feature robustness in non-stationary health records: caveats to deployable model performance in common clinical machine learning tasks. arXiv preprint arXiv:1908.00690, 2019.Google Scholar
Caruana, Rich, Lou, Yin, Gehrke, Johannes, et al. Intelligible models for healthcare: predicting pneumonia risk and hospital 30-day readmission. Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 1721–1730. 2015.Google Scholar
Bastani, Osbert, Kim, Carolyn, and Bastani, Hamsa. Interpreting blackbox models via model extraction. arXiv preprint arXiv:1705.08504, 2017.Google Scholar
Breiman, Leo, Friedman, Jerome, Stone, Charles J, and Olshen, Richard A. Classification and Regression Trees. CRC Press, 1984.Google Scholar
Bastani, Hamsa, Bastani, Osbert, and Kim, Carolyn. Interpreting predictive models for human-in-the-loop analytics. arXiv preprint arXiv:1705.08504, 1–45, 2018. https://hamsabastani.github.io/interp.pdf.Google Scholar
Colwell, Janet. Length of stay: timing it right. Strategies for achieving efficient, high-quality care. ACP Hospitalist, 2014. www.acphospitalist.org/archives/2014/10/los.htm.Google Scholar
Adepoju, Temidayo, Tucker, Anita, Jin, Helen, and Manasseh, Chris. The impact of two managerial responses on hospital occupancy crises. SSRN Electronic Journal, 2019.Google Scholar
Frenz, David. Not too long, not too short, just right. Today’s Hospitalist, 2014. www.todayshospitalist.com/not-too-long-not-too-short-just-right/.Google Scholar
Shi, Pengyi, Helm, Jonathan, Deglise-Hawkinson, Jivan, and Pan, Julian. Timing it right: balancing inpatient congestion versus readmission risk at discharge. Operations Research, 2020. Forthcoming.Google Scholar
Bardhan, Indranil, Jeong-ha, Oh, Zheng, Zhiqiang, and Kirksey, Kirk. Predictive analytics for readmission of patients with congestive heart failure. Information Systems Research, 26(1):19–39, 2014.Google Scholar
Thompson, Steven, Nunez, Manuel, Garfinkel, Robert, and Dean, Matthew D. Or practice – efficient short-term allocation and reallocation of patients to floors of a hospital during demand surges. Operations Research, 57(2):261–273, 2009.Google Scholar
Yang, Hongyu, Rudin, Cynthia, and Seltzer, Margo. Scalable Bayesian rule lists. Proceedings of the 34th International Conference on Machine Learning, 70:3921–3930, 2017.Google Scholar
Friedman, Jerome H. Greedy function approximation: a gradient boosting machine. Annals of Statistics, 29(5):1189–1232, 2001.Google Scholar
Bertsimas, Dimitris and Dunn, Jack. Optimal classification trees. Machine Learning, 106(7):1039–1082, 2017.Google Scholar
Ribeiro, Marco Tulio, Singh, Sameer, and Guestrin, Carlos. Why should I trust you? Explaining the predictions of any classifier. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 1135–1144. 2016.Google Scholar
Simonyan, Karen, Vedaldi, Andrea, and Zisserman, Andrew. Deep inside convolutional networks: visualising image classification models and saliency maps. arXiv preprint arXiv:1312.6034, 2013.Google Scholar
Lundberg, Scott M and Lee, Su-In. A unified approach to interpreting model predictions. In Advances in Neural Information Processing Systems, 4765–4774, 2017.Google Scholar
Lakkaraju, Himabindu and Bastani, Osbert. “How do I fool you?”: manipulating user trust via misleading black box explanations. AIES, 2020.Google Scholar
Wei Koh, Pang and Liang, Percy. Understanding black-box predictions via influence functions. In Proceedings of the 34th International Conference on Machine Learning. Volume 70, 1885–1894. JMLR. org, 2017.Google Scholar
Berk, Emre and Moinzadeh, Kamran. The impact of discharge decisions on health care quality. Management Science, 44(3):400–415, 1998.Google Scholar
Chan, Carri W, Farias, Vivek F, Bambos, Nicholas, and Escobar, Gabriel. Optimizing intensive care unit discharge decisions with patient readmissions. Operations Research, 60(6):1323–1341, 2012.CrossRefGoogle Scholar
Ouyang, Huiyin, Argon, Nilay Tanik, and Ziya, Serhan. Allocation of intensive care unit beds in periods of high demand. Operations Research, 2019. Forthcoming.Google Scholar
Bavafa, Hessam, Ormeci, Lerzan, Savin, Sergei, and Virudachalam, Vanitha. Surgical case-mix and discharge decisions: does within-hospital coordination matter? Working Paper, 1–40, 2019.Google Scholar
Bertsekas, Dimitri P.. Dynamic Programming and Optimal Control: Approximate Dynamic Programming. Volume II. Belmont, MA: Athena Scientific, 2012.Google Scholar
Powell, Warren B.. Approximate Dynamic Programming: Solving the Curses of Dimensionality. Wiley Series in Probability and Statistics. Hoboken, NJ: Wiley-Interscience, 2011.CrossRefGoogle Scholar
Sutton, Richard S.. Learning to predict by the methods of temporal differences. Machine Learning, 3(1):9–44, 1988.Google Scholar
Christopher, J.C.H. Watkins and Peter Dayan. Technical note: Q-learning. Machine Learning, 8(3):279–292, 1992.Google Scholar
de Farias, D. P. and Van Roy, B.. The linear programming approach to approximate dynamic programming. Operations Research, 51(6):850–865, 2003.Google Scholar
Adelman, Daniel and Mersereau, Adam J.. Relaxations of weakly coupled stochastic dynamic programs. Operations Research, 56(3):712–727, 2008.Google Scholar
Schulman, John, Levine, Sergey, Abbeel, Pieter, Jordan, Michael, and Moritz, Philipp. Trust region policy optimization. In International Conference on Machine Learning, 1889–1897, 2015.Google Scholar
Schulman, John, Wolski, Filip, Dhariwal, Prafulla, Radford, Alec, and Klimov, Oleg. Proximal policy optimization algorithms. arXiv preprint arXiv:1707.06347, 2017.Google Scholar