Preparing a Clinical Support Model for Silent Mode in General Internal Medicine

Bret Nestor, Liam G. McCoy, Amol Verma, Chloe Pou-Prom, Joshua Murray, Sebnem Kuzulugil, David Dai, Muhammad Mamdani, Anna Goldenberg, Marzyeh Ghassemi
Proceedings of the 5th Machine Learning for Healthcare Conference, PMLR 126:950-972, 2020.

Abstract

The general internal medicine (GIM) ward oversees the recovery of ill patients, excluding those who require intensive attention. Clinicians provide full recoveries, or when appropriate, end-of-life care. We hope to eliminate unexpected deaths in the GIM ward, promptly transfer patients who require escalated care to the intensive care unit, and proactively address deteriorating health to minimise ICU transfers. We describe a clinical decision support system which accesses labs, vitals, administered medications, clinical orders, and specialty consults. Using an ensemble of linear, gated recurrent unit (GRU) and GRU-decay (GRU-D) models, we are able to achieve a positive predictive value of 0.71 while successfully identifying 40% of patients who will experience a future adverse event. We believe that this tool will be useful in shift scheduling and discharging patients, in addition to warning clinicians of risk of decompensation. We note the lessons we learned in transitioning from a high performing model to deployment in silent mode, and all results reported in this paper report on data immediately preceding silent mode.

Cite this Paper


BibTeX
@InProceedings{pmlr-v126-nestor20a, title = {Preparing a Clinical Support Model for Silent Mode in General Internal Medicine}, author = {Nestor, Bret and McCoy, Liam G. and Verma, Amol and Pou-Prom, Chloe and Murray, Joshua and Kuzulugil, Sebnem and Dai, David and Mamdani, Muhammad and Goldenberg, Anna and Ghassemi, Marzyeh}, booktitle = {Proceedings of the 5th Machine Learning for Healthcare Conference}, pages = {950--972}, year = {2020}, editor = {Doshi-Velez, Finale and Fackler, Jim and Jung, Ken and Kale, David and Ranganath, Rajesh and Wallace, Byron and Wiens, Jenna}, volume = {126}, series = {Proceedings of Machine Learning Research}, month = {07--08 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v126/nestor20a/nestor20a.pdf}, url = {http://proceedings.mlr.press/v126/nestor20a.html}, abstract = {The general internal medicine (GIM) ward oversees the recovery of ill patients, excluding those who require intensive attention. Clinicians provide full recoveries, or when appropriate, end-of-life care. We hope to eliminate unexpected deaths in the GIM ward, promptly transfer patients who require escalated care to the intensive care unit, and proactively address deteriorating health to minimise ICU transfers. We describe a clinical decision support system which accesses labs, vitals, administered medications, clinical orders, and specialty consults. Using an ensemble of linear, gated recurrent unit (GRU) and GRU-decay (GRU-D) models, we are able to achieve a positive predictive value of 0.71 while successfully identifying 40% of patients who will experience a future adverse event. We believe that this tool will be useful in shift scheduling and discharging patients, in addition to warning clinicians of risk of decompensation. We note the lessons we learned in transitioning from a high performing model to deployment in silent mode, and all results reported in this paper report on data immediately preceding silent mode.} }
Endnote
%0 Conference Paper %T Preparing a Clinical Support Model for Silent Mode in General Internal Medicine %A Bret Nestor %A Liam G. McCoy %A Amol Verma %A Chloe Pou-Prom %A Joshua Murray %A Sebnem Kuzulugil %A David Dai %A Muhammad Mamdani %A Anna Goldenberg %A Marzyeh Ghassemi %B Proceedings of the 5th Machine Learning for Healthcare Conference %C Proceedings of Machine Learning Research %D 2020 %E Finale Doshi-Velez %E Jim Fackler %E Ken Jung %E David Kale %E Rajesh Ranganath %E Byron Wallace %E Jenna Wiens %F pmlr-v126-nestor20a %I PMLR %P 950--972 %U http://proceedings.mlr.press/v126/nestor20a.html %V 126 %X The general internal medicine (GIM) ward oversees the recovery of ill patients, excluding those who require intensive attention. Clinicians provide full recoveries, or when appropriate, end-of-life care. We hope to eliminate unexpected deaths in the GIM ward, promptly transfer patients who require escalated care to the intensive care unit, and proactively address deteriorating health to minimise ICU transfers. We describe a clinical decision support system which accesses labs, vitals, administered medications, clinical orders, and specialty consults. Using an ensemble of linear, gated recurrent unit (GRU) and GRU-decay (GRU-D) models, we are able to achieve a positive predictive value of 0.71 while successfully identifying 40% of patients who will experience a future adverse event. We believe that this tool will be useful in shift scheduling and discharging patients, in addition to warning clinicians of risk of decompensation. We note the lessons we learned in transitioning from a high performing model to deployment in silent mode, and all results reported in this paper report on data immediately preceding silent mode.
APA
Nestor, B., McCoy, L.G., Verma, A., Pou-Prom, C., Murray, J., Kuzulugil, S., Dai, D., Mamdani, M., Goldenberg, A. & Ghassemi, M.. (2020). Preparing a Clinical Support Model for Silent Mode in General Internal Medicine. Proceedings of the 5th Machine Learning for Healthcare Conference, in Proceedings of Machine Learning Research 126:950-972 Available from http://proceedings.mlr.press/v126/nestor20a.html.

Related Material