Explainable machine learning predictions to help anesthesiologists prevent hypoxemia during surgery
Open accessAbstract
Hypoxemia causes serious patient harm, and while anesthesiologists strive to avoid hypoxemia during surgery, anesthesiologists are not reliably able to predict which patients will have intraoperative hypoxemia. Using minute by minute EMR data from fifty thousand surgeries we developed and tested a machine learning based system called Prescience that predicts real-time hypoxemia risk and presents an explanation of factors contributing to that risk during general anesthesia. Prescience improved anesthesiologists’ performance when providing interpretable hypoxemia risks with contributing factors . The results suggest that if anesthesiologists currently anticipate 15% of events, then with Prescience assistance they could anticipate 30% of events or an estimated additional 2.4 million annually in the US, a large portion of which may be preventable because they are attributable to modifiable factors. The prediction explanations are broadly consistent with the literature and anesthesiologists’ prior knowledge. Prescience can also improve clinical understanding of hypoxemia risk during anesthesia by providing general insights into the exact changes in risk induced by certain patient or procedure characteristics. Making predictions of complex medical machine learning models (such as Prescience) interpretable has broad applicability to other data-driven prediction tasks in medicine.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Shapley values are locally accurate, meaning they can explain a prediction for a single attendance.
“One particular strength of Shapley values is that they are locally accurate, i.e, they can explain a prediction for a single attendance.”
From Using explainable machine learning to characterise data drift and detect emergent health risks for emergency department admissions during COVID-19 (2021) · cites this paper as [16]
TreeExplainer, as implemented in the SHAP library, efficiently calculates Shapley values for attributing the fractional contribution of an input feature to admission risk.
“To attribute the fractional contribution of a given input feature to admission risk, we made use of the TreeExplainer algorithm as implemented in the SHAP (SHapley Additive exPlanations) library16–18.”
From Using explainable machine learning to characterise data drift and detect emergent health risks for emergency department admissions during COVID-19 (2021) · cites this paper as [16]
The sum of SHAP values for a given patient attendance equals the patient's admission risk predicted by the model.
“The sum of SHAP values for a given patient attendance must therefore equal the patient’s admission risk predicted by the model.”
From Using explainable machine learning to characterise data drift and detect emergent health risks for emergency department admissions during COVID-19 (2021) · cites this paper as [16]
Shapley values are consistent, meaning the values add up to the actual prediction of the model, so they can be used to check the global importance of a feature.
“In addition, Shapley values are consistent (the values add up to the actual prediction of the model) meaning they can also be used to check the global importance of a feature.”
From Using explainable machine learning to characterise data drift and detect emergent health risks for emergency department admissions during COVID-19 (2021) · cites this paper as [16]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
XGBoost: A Scalable Tree Boosting System
Chen, T., & Guestrin, C. (2016) · Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining · doi:10.1145/2939672.2939785 · Cited by 54,569 · see citing works
Greedy function approximation: A gradient boosting machine
Friedman, J. H. (2001) · The Annals of Statistics · doi:10.1214/aos/1013203451 · Cited by 30,731 · see citing works
"Why Should I Trust You?": Explaining the Predictions of Any Classifier
Ribeiro, M. T., Singh, S., & Guestrin, C. (2016) · Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining · doi:10.1145/2939672.2939778 · Cited by 17,555 · see citing works
Pharmacological interventions for somatoform disorders in adults
Kleinstäuber, M., Witthöft, M., Steffanowski, A., van Marwijk, H., Hiller, W., & Lambert, M. J. (2014) · Cochrane Database of Systematic Reviews · doi:10.1002/14651858.cd010628.pub2 · Cited by 11,361 · see citing works
A surgical safety checklist to reduce morbidity and mortality in a global population
Haynes, A. B., Weiser, T. G., Berry, W. R., Lipsitz, S. R., Breizat, A.-H. S., Dellinger, E. P., Herbosa, T., Joseph, S., Kibatala, P. L., Lapitan, M. C. M., et al. (2009) · New England Journal of Medicine · doi:10.1056/nejmsa0810119 · Cited by 5,688 · see citing works
Deo, R. C. (2015) · Circulation · doi:10.1161/circulationaha.115.001593 · Cited by 3,466 · see citing works
Explaining prediction models and individual predictions with feature contributions
Štrumbelj, E., & Kononenko, I. (2014) · Knowledge and Information Systems · doi:10.1007/s10115-013-0679-x · Cited by 2,286 · see citing works
Cleveland, W. S., & McGill, R. (1984) · Journal of the American Statistical Association · doi:10.1080/01621459.1984.10478080 · Cited by 1,797 · see citing works
The incidence and nature of surgical adverse events in Colorado and Utah in 1992
Gawande, A. A., Thomas, E. J., Zinner, M. J., & Brennan, T. A. (1999) · Surgery · doi:10.1067/msy.1999.98664 · Cited by 977 · see citing works
The Shapley value : essays in honor of Lloyd S. Shapley
Roth, A. E. (1991) · Cited by 700 · see citing works
Surgical data science for next-generation interventions
Maier‐Hein, L., Vedula, S. S., Speidel, S., Navab, N., Kikinis, R., Park, A., Eisenmann, M., Feußner, H., Forestier, G., Giannarou, S., et al. (2017) · Nature Biomedical Engineering · doi:10.1038/s41551-017-0132-7 · Cited by 616 · see citing works
A targeted real-time early warning score (TREWScore) for septic shock.
Henry, K. E., Hager, D. N., Pronovost, P. J., & Saria, S. (2015) · Science Translational Medicine · doi:10.1126/scitranslmed.aab3719 · Cited by 598 · see citing works
Adverse events in surgical patients in Australia
Kable, A. K., Gibberd, R. W., & Spigelman, A. D. (2002) · International Journal for Quality in Health Care · doi:10.1093/intqhc/14.4.269 · Cited by 453 · see citing works
Integrated monitoring and analysis for early warning of patient deterioration
Tarassenko, L., Hann, A., & Young, D. (2006) · British Journal of Anaesthesia · doi:10.1093/bja/ael113 · Cited by 250 · see citing works
Integration of Early Physiological Responses Predicts Later Illness Severity in Preterm Infants
Saria, S., Rajani, A. K., Gould, J. B., Koller, D., & Penn, A. A. (2010) · Science Translational Medicine · doi:10.1126/scitranslmed.3001304 · Cited by 177 · see citing works
The incidence of hypoxemia during surgery: evidence from two institutions
Ehrenfeld, J. M., Funk, L. M., van Schalkwyk, J., Merry, A., Sandberg, W. S., & Gawande, A. A. (2010) · Canadian Journal of Anesthesia/Journal canadien d anesthésie · doi:10.1007/s12630-010-9366-5 · Cited by 136 · see citing works
An unexpected unity among methods for interpreting model predictionsPreprint
Lundberg, S., & Lee, S. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1611.07478 · Cited by 101 · see citing works
Self-citationIntraoperative Clinical Decision Support for Anesthesia: A Narrative Review of Available Systems.
Nair, B., Gabel, E., Hofer, I., Schwid, H., & Cannesson, M. (2017) · Anesthesia and Analgesia · doi:10.1213/ane.0000000000001636 · Cited by 75 · see citing works
Self-citationDunham, C. M., Hileman, B. M., Hutchinson, A. E., Chance, E. A., & Huang, G. S. (2014) · BMC Anesthesiology · doi:10.1186/1471-2253-14-43 · Cited by 41 · see citing works
Sub-Tenon's anaesthesia versus topical anaesthesia for cataract surgery
Guay, J. M. Y., & Sales, K. (2015) · Cochrane Database of Systematic Reviews · doi:10.1002/14651858.cd006291.pub3 · Cited by 41 · see citing works
Multi-Step Ahead Predictions for Critical Levels in Physiological Time Series
ElMoaqet, H., Tilbury, D. M., & Ramachandran, S. K. (2016) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2016.2561974 · Cited by 37 · see citing works
Learning (predictive) risk scores in the presence of censoring due to interventions
Dyagilev, K., & Saria, S. (2015) · Machine Learning · doi:10.1007/s10994-015-5527-7 · Cited by 36 · see citing works
Increasing body mass index and the incidence of intraoperative hypoxemia
Kendale, S. M., & Blitz, J. D. (2016) · Journal of Clinical Anesthesia · doi:10.1016/j.jclinane.2016.03.020 · Cited by 35 · see citing works
Phenotyping of Clinical Time Series with LSTM Recurrent Neural NetworksPreprint
Lipton, Z. C., Kale, D. C., & Wetzel, R. C. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1510.07641 · Cited by 31 · see citing works
Strachan, L. M., & Noble, D. W. (2001) · PubMed · Cited by 23 · see citing works
Lumachi, F., Marzano, B., Fanti, G., Basso, S. M. M., Mazza, F., & Chiara, G. B. (2010) · PubMed · Cited by 17 · see citing works
Flagged first, then by citations
2 of the 28 references couldn’t be matched to a known paper.
Research works printed but not linked: 2
- J. Guay, E. A. Ochroch, in Cochrane Database of Systematic Reviews, J. Guay, Ed. (John Wiley & Sons, Ltd, Chichester, UK, 2015), p. CD011151.
- A. E. Roth, The Shapley Value -Cambridge University PressCambridge Univ. Press (1988) (available at http://www.cambridge.org/catalogue/catalogue.asp?isbn=0511829728).
Cited by
Sorted by citations
From local explanations to global understanding with explainable AI for trees
Lundberg, S. M., Erion, G., Chen, H., DeGrave, A., Prutkin, J. M., Nair, B., Katz, R., Himmelfarb, J., Bansal, N., & Lee, S.-I. (2020) · Nature Machine Intelligence · doi:10.1038/s42256-019-0138-9 · Cited by 10,551 · see citing works
Self-citationZitnik, M., Nguyen, F., Wang, B., Leskovec, J., Goldenberg, A., & Hoffman, M. M. (2019) · Information Fusion · doi:10.1016/j.inffus.2018.09.012 · Cited by 675 · see citing works
Consistent Individualized Feature Attribution for Tree Ensembles
Lundberg, S. M., Erion, G. G., & Lee, S.-I. (2018) · arXiv · doi:10.48550/arxiv.1802.03888 · Cited by 1,430 · see citing works
Self-citationExplainable AI for Trees: From Local Explanations to Global UnderstandingPreprint
Lundberg, S., Erion, G., Chen, H., DeGrave, A. J., Prutkin, J. M., Nair, B. G., Katz, R., Himmelfarb, J., Bansal, N., & Lee, S. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1905.04610 · Cited by 252 · see citing works
Self-citationNeupane, S., Ables, J., Anderson, W. S., Mittal, S., Rahimi, S., Banicescu, I., & Seale, M. A. (2022) · IEEE Access · doi:10.1109/access.2022.3216617 · Cited by 293 · see citing works
Duckworth, C., Chmiel, F. P., Burns, D. K., Zlatev, Z. D., White, N. M., Daniels, T. W. V., Kiuber, M., & Boniface, M. J. (2021) · Scientific Reports · doi:10.1038/s41598-021-02481-y · Cited by 109 · see citing works
Barda, A. J., Horvat, C. M., & Hochheiser, H. (2020) · BMC Medical Informatics and Decision Making · doi:10.1186/s12911-020-01276-x · Cited by 105 · see citing works
Duckworth, C., Guy, M., Kumaran, A., O’Kane, A. A., Ayobi, A., Chapman, A. P., Marshall, P., & Boniface, M. (2022) · Journal of Diabetes Science and Technology · doi:10.1177/19322968221103561 · Cited by 76 · see citing works
Explainable Boosting Machines for Slope Failure Spatial Predictive Modeling
Maxwell, A. E., Sharma, M., & Donaldson, K. A. (2021) · Remote Sensing · doi:10.3390/rs13244991 · Cited by 68 · see citing works
Investigations of Performance and Bias in Human-AI Teamwork in Hiring
Peng, A., Nushi, B., Kıcıman, E., Inkpen, K. M., & Kamar, E. S. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i11.21468 · Cited by 34 · see citing works
Neurophysiological Correlates of Concussion: Deep Learning for Clinical Assessment
Boshra, R., Ruiter, K. I., DeMatteo, C. A., Reilly, J. P., & Connolly, J. F. (2019) · Scientific Reports · doi:10.1038/s41598-019-53751-9 · Cited by 26 · see citing works
Enabling artificial intelligence in high acuity medical environments
Kasparick, M., Andersen, B., Franke, S., Rockstroh, M., Golatowski, F., Timmermann, D., Ingenerf, J., & Neumuth, T. (2019) · Minimally Invasive Therapy & Allied Technologies · doi:10.1080/13645706.2019.1599957 · Cited by 21 · see citing works
Lazzarini, N., Filippoupolitis, A., Manzione, P., & Eleftherohorinou, H. (2022) · PLoS ONE · doi:10.1371/journal.pone.0271227 · Cited by 20 · see citing works
TruVR: Trustworthy Cybersickness Detection using Explainable Machine Learning
Kundu, R. K., Islam, R., Calyam, P. P., & Hoque, K. A. (2022) · Proceedings - International Symposium on Mixed and Augmented Reality, ISMAR · doi:10.1109/ismar55827.2022.00096 · Cited by 21 · see citing works
Anomaly Detection and Analysis in Nuclear Power Plants
Chaudhary, A., Han, J., Kim, S. K., Kim, A., & Choi, S. (2024) · Electronics · doi:10.3390/electronics13224428 · Cited by 18 · see citing works
Al-Hussaini, I., White, B., Varmeziar, A., Mehra, N., Sanchez, M., Lee, J. J., DeGroote, N. P., Miller, T. P., & Mitchell, C. S. (2024) · Journal of Clinical Medicine · doi:10.3390/jcm13061788 · Cited by 15 · see citing works
Surgical data science, an emerging field of medicine
Nagy, D. Á., Rudas, I. J., & Haidegger, T. P. (2017) · IEEE Neumann Colloquium (NC) · doi:10.1109/nc.2017.8263251 · Cited by 14 · see citing works
Diabetic risk prognosis with tree ensembles integrating feature attribution methods
Hansen, J. V. (2021) · Evolutionary Intelligence · doi:10.1007/s12065-021-00663-1 · Cited by 11 · see citing works
Machine and Deep Learning Models for Hypoxemia Severity Triage in CBRNE Emergencies
Nanini, S., Abid, M., Mamouni, Y., Wiedemann, A., Jouvet, P. A., & Bourassa, S. (2024) · Diagnostics · doi:10.3390/diagnostics14232763 · Cited by 12 · see citing works
Zhao, Y. S., He, B., Xu, Z., Zhang, Y., Zhao, X., Huang, Z.-A., Yang, F., Wang, L., Duan, L., Song, J., et al. (2022) · Briefings in Bioinformatics · doi:10.1093/bib/bbac555 · Cited by 7 · see citing works
Hybrid Gradient Boosting Trees and Neural Networks for Forecasting Operating Room DataPreprint
Chen, H., Lundberg, S., & Lee, S. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1801.07384 · Cited by 8 · see citing works
Self-citationUtkin, L. V., Мелдо, А. А., Kovalev, M. S., & Kasimov, E. M. (2021) · Scientific and Technical Information Processing · doi:10.3103/s0147688221050129 · Cited by 6 · see citing works
Anesthesiologist-level forecasting of hypoxemia with only SpO2 data using deep learningPreprint
Erion, G., Chen, H., Lundberg, S., & Lee, S. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1712.00563 · Cited by 5 · see citing works
Self-citationHeading Toward Trusted ATCO-AI Systems: A Literature ReviewPreprint
Zhong, J. Y. (2021) · doi:10.31224/osf.io/t769b · Cited by 5 · see citing works
Applying Explainable Artificial Intelligence Techniques on Linked Open Government Data
Kalampokis, E., Karamanou, A., & Tarabanis, K. A. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-84789-0_18 · Cited by 4 · see citing works
Coalition Feature Interpretation and Attribution in Algorithmic Trading Models
Hansen, J. V. (2020) · Computational Economics · doi:10.1007/s10614-020-10053-x · Cited by 3 · see citing works
Diagonal Discriminant Analysis With Feature Selection for High-Dimensional Data
Romanes, S., Ormerod, J. T., & Yang, J. Y. H. (2019) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2019.1637748 · Cited by 2 · see citing works
Study Program "Robotics and Autonomous Systems" at the University of Lübeck
Herzog, C., Hoffmann, N., Tantau, T., & Rostalski, P. (2018) · European Workshop on Microelectronics Education (EWME) · doi:10.1109/ewme.2018.8629446
Moreira, L. I. (2023) · Digital Library of Theses and Dissertations (Universidade de São Paulo) · doi:10.11606/d.5.2023.tde-04082023-142705
Barda, A. J., Horvat, C. M., & Hochheiser, H. (2020) · Figshare · doi:10.6084/m9.figshare.13071218
Barda, A. J., Horvat, C. M., & Hochheiser, H. (2020) · Figshare · doi:10.6084/m9.figshare.13071224
Barda, A. J., Horvat, C. M., & Hochheiser, H. (2020) · Figshare · doi:10.6084/m9.figshare.13071215
Barda, A. J., Horvat, C. M., & Hochheiser, H. (2020) · Figshare · doi:10.6084/m9.figshare.13071221
Barda, A. J., Horvat, C. M., & Hochheiser, H. (2020) · Figshare · doi:10.6084/m9.figshare.13071227
- Authors
- Lundberg, S. M., Nair, B., Vavilala, M. S., Horibe, M., Eisses, M. J., Adams, T., Liston, D. E., Low, D. K.-W., Newman, S.-F., Kim, J., & Lee, S.-I.
- Affiliations
- University of Washington (US)
- VA Puget Sound Health Care System (US)
- Seattle Children's Hospital (US)
- Corresponding author
- Lee, Su-In
- Venue
- bioRxiv (Cold Spring Harbor Laboratory)
- Publisher
- openRxiv
- Published
- 2017-10-21
- OpenAlex
- W2766560823
- Cited by
- 34
- References
- 28
- Citations by year
- 2024: 3 · 2023: 1 · 2022: 6 · 2021: 6 · 2020: 8 · 2019: 4 · 2018: 4 · 2017: 2
- Type
- posted content
- Language
- en
- Open access
- green
- doi.org
- Locations
- bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion)
- Funders
- National Science Foundation
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning in Healthcare
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
- Keywords
- intraoperative hypoxemia, general anesthesia, explainable machine learning, clinical decision support, anesthesiologists, real-time risk prediction, electronic health records, prescience, modifiable risk factors, patient safety
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
