Teaching a Machine to Feel Postoperative Pain

Combining High-Dimensional Clinical Data with Machine Learning Algorithms to Forecast Acute Postoperative Pain

Patrick J. Tighe, Chris Harle, Robert W. Hurley, Haldun Aytug, Andre P. Boezaart, Roger B. Fillingim

Research output: Contribution to journalArticle

11 Citations (Scopus)

Abstract

Background: Given their ability to process highly dimensional datasets with hundreds of variables, machine learning algorithms may offer one solution to the vexing challenge of predicting postoperative pain. Methods: Here, we report on the application of machine learning algorithms to predict postoperative pain outcomes in a retrospective cohort of 8,071 surgical patients using 796 clinical variables. Five algorithms were compared in terms of their ability to forecast moderate to severe postoperative pain: Least Absolute Shrinkage and Selection Operator (LASSO), gradient-boosted decision tree, support vector machine, neural network, and k-nearest neighbor (k-NN), with logistic regression included for baseline comparison. Results: In forecasting moderate to severe postoperative pain for postoperative day (POD) 1, the LASSO algorithm, using all 796 variables, had the highest accuracy with an area under the receiver-operating curve (ROC) of 0.704. Next, the gradient-boosted decision tree had an ROC of 0.665 and the k-NN algorithm had an ROC of 0.643. For POD 3, the LASSO algorithm, using all variables, again had the highest accuracy, with an ROC of 0.727. Logistic regression had a lower ROC of 0.5 for predicting pain outcomes on POD 1 and 3. Conclusions: Machine learning algorithms, when combined with complex and heterogeneous data from electronic medical record systems, can forecast acute postoperative pain outcomes with accuracies similar to methods that rely only on variables specifically collected for pain outcome prediction.

Original languageEnglish (US)
Pages (from-to)1386-1401
Number of pages16
JournalPain Medicine (United States)
Volume16
Issue number7
DOIs
StatePublished - Jan 1 2015
Externally publishedYes

Fingerprint

Acute Pain
Postoperative Pain
Teaching
Decision Trees
Aptitude
Logistic Models
Pain
Electronic Health Records
Machine Learning

Keywords

  • Algorithm
  • Machine Learning
  • Pain Prediction
  • Postoperative Pain

ASJC Scopus subject areas

  • Clinical Neurology
  • Anesthesiology and Pain Medicine

Cite this

Teaching a Machine to Feel Postoperative Pain : Combining High-Dimensional Clinical Data with Machine Learning Algorithms to Forecast Acute Postoperative Pain. / Tighe, Patrick J.; Harle, Chris; Hurley, Robert W.; Aytug, Haldun; Boezaart, Andre P.; Fillingim, Roger B.

In: Pain Medicine (United States), Vol. 16, No. 7, 01.01.2015, p. 1386-1401.

Research output: Contribution to journalArticle

Tighe, Patrick J. ; Harle, Chris ; Hurley, Robert W. ; Aytug, Haldun ; Boezaart, Andre P. ; Fillingim, Roger B. / Teaching a Machine to Feel Postoperative Pain : Combining High-Dimensional Clinical Data with Machine Learning Algorithms to Forecast Acute Postoperative Pain. In: Pain Medicine (United States). 2015 ; Vol. 16, No. 7. pp. 1386-1401.
@article{f139823e593748c7942a70348e68b7d7,
title = "Teaching a Machine to Feel Postoperative Pain: Combining High-Dimensional Clinical Data with Machine Learning Algorithms to Forecast Acute Postoperative Pain",
abstract = "Background: Given their ability to process highly dimensional datasets with hundreds of variables, machine learning algorithms may offer one solution to the vexing challenge of predicting postoperative pain. Methods: Here, we report on the application of machine learning algorithms to predict postoperative pain outcomes in a retrospective cohort of 8,071 surgical patients using 796 clinical variables. Five algorithms were compared in terms of their ability to forecast moderate to severe postoperative pain: Least Absolute Shrinkage and Selection Operator (LASSO), gradient-boosted decision tree, support vector machine, neural network, and k-nearest neighbor (k-NN), with logistic regression included for baseline comparison. Results: In forecasting moderate to severe postoperative pain for postoperative day (POD) 1, the LASSO algorithm, using all 796 variables, had the highest accuracy with an area under the receiver-operating curve (ROC) of 0.704. Next, the gradient-boosted decision tree had an ROC of 0.665 and the k-NN algorithm had an ROC of 0.643. For POD 3, the LASSO algorithm, using all variables, again had the highest accuracy, with an ROC of 0.727. Logistic regression had a lower ROC of 0.5 for predicting pain outcomes on POD 1 and 3. Conclusions: Machine learning algorithms, when combined with complex and heterogeneous data from electronic medical record systems, can forecast acute postoperative pain outcomes with accuracies similar to methods that rely only on variables specifically collected for pain outcome prediction.",
keywords = "Algorithm, Machine Learning, Pain Prediction, Postoperative Pain",
author = "Tighe, {Patrick J.} and Chris Harle and Hurley, {Robert W.} and Haldun Aytug and Boezaart, {Andre P.} and Fillingim, {Roger B.}",
year = "2015",
month = "1",
day = "1",
doi = "10.1111/pme.12713",
language = "English (US)",
volume = "16",
pages = "1386--1401",
journal = "Pain Medicine",
issn = "1526-2375",
publisher = "Wiley-Blackwell",
number = "7",

}

TY - JOUR

T1 - Teaching a Machine to Feel Postoperative Pain

T2 - Combining High-Dimensional Clinical Data with Machine Learning Algorithms to Forecast Acute Postoperative Pain

AU - Tighe, Patrick J.

AU - Harle, Chris

AU - Hurley, Robert W.

AU - Aytug, Haldun

AU - Boezaart, Andre P.

AU - Fillingim, Roger B.

PY - 2015/1/1

Y1 - 2015/1/1

N2 - Background: Given their ability to process highly dimensional datasets with hundreds of variables, machine learning algorithms may offer one solution to the vexing challenge of predicting postoperative pain. Methods: Here, we report on the application of machine learning algorithms to predict postoperative pain outcomes in a retrospective cohort of 8,071 surgical patients using 796 clinical variables. Five algorithms were compared in terms of their ability to forecast moderate to severe postoperative pain: Least Absolute Shrinkage and Selection Operator (LASSO), gradient-boosted decision tree, support vector machine, neural network, and k-nearest neighbor (k-NN), with logistic regression included for baseline comparison. Results: In forecasting moderate to severe postoperative pain for postoperative day (POD) 1, the LASSO algorithm, using all 796 variables, had the highest accuracy with an area under the receiver-operating curve (ROC) of 0.704. Next, the gradient-boosted decision tree had an ROC of 0.665 and the k-NN algorithm had an ROC of 0.643. For POD 3, the LASSO algorithm, using all variables, again had the highest accuracy, with an ROC of 0.727. Logistic regression had a lower ROC of 0.5 for predicting pain outcomes on POD 1 and 3. Conclusions: Machine learning algorithms, when combined with complex and heterogeneous data from electronic medical record systems, can forecast acute postoperative pain outcomes with accuracies similar to methods that rely only on variables specifically collected for pain outcome prediction.

AB - Background: Given their ability to process highly dimensional datasets with hundreds of variables, machine learning algorithms may offer one solution to the vexing challenge of predicting postoperative pain. Methods: Here, we report on the application of machine learning algorithms to predict postoperative pain outcomes in a retrospective cohort of 8,071 surgical patients using 796 clinical variables. Five algorithms were compared in terms of their ability to forecast moderate to severe postoperative pain: Least Absolute Shrinkage and Selection Operator (LASSO), gradient-boosted decision tree, support vector machine, neural network, and k-nearest neighbor (k-NN), with logistic regression included for baseline comparison. Results: In forecasting moderate to severe postoperative pain for postoperative day (POD) 1, the LASSO algorithm, using all 796 variables, had the highest accuracy with an area under the receiver-operating curve (ROC) of 0.704. Next, the gradient-boosted decision tree had an ROC of 0.665 and the k-NN algorithm had an ROC of 0.643. For POD 3, the LASSO algorithm, using all variables, again had the highest accuracy, with an ROC of 0.727. Logistic regression had a lower ROC of 0.5 for predicting pain outcomes on POD 1 and 3. Conclusions: Machine learning algorithms, when combined with complex and heterogeneous data from electronic medical record systems, can forecast acute postoperative pain outcomes with accuracies similar to methods that rely only on variables specifically collected for pain outcome prediction.

KW - Algorithm

KW - Machine Learning

KW - Pain Prediction

KW - Postoperative Pain

UR - http://www.scopus.com/inward/record.url?scp=84947133089&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84947133089&partnerID=8YFLogxK

U2 - 10.1111/pme.12713

DO - 10.1111/pme.12713

M3 - Article

VL - 16

SP - 1386

EP - 1401

JO - Pain Medicine

JF - Pain Medicine

SN - 1526-2375

IS - 7

ER -