Postoperative bleeding risk prediction for patients undergoing colorectal surgery

Research output: Contribution to journalArticle

Abstract

Background: There is limited consensus regarding risk factors for postoperative bleeding. The objective of this work was to investigate the capability of machine learning techniques in combination with practice-based longitudinal electronic medical record data for identifying potential new risk factors for postoperative bleeding and predicting patients at high risk of postoperative bleeding. Methods: A retrospective study was conducted for patients who underwent colorectal surgery 1998–2015 at a single tertiary referral center. Various predictors were extracted from electronic medical record. The outcome of interest was the occurrence of postoperative bleeding within 7 days of surgery. Logistic regression and gradient boosting machine models were trained. Area under the receiver operating curve and area under the precision recall curve were used to evaluate the performance to different models. Results: Of 13,399 cases undergoing colorectal resection, 1,680 (12.5%) experienced postoperative bleeding. A total of 299 variables were evaluated. Logistic regression and gradient boosting machine models returned an area under the receiver operating curve of 0.735 and 0.822 and area under the precision recall curve of 0.287 and 0.423, respectively. In addition to well-known risk factors for postoperative bleeding, nutrition (ranked third), weakness (ranked fifth), patient mobility (ranked sixth), and activity level (ranked eighth) were found to be novel predictors in the gradient boosting machine model based on permutation importance. Conclusion: The study identified measures of functional capacity of patient as novel predictors of postoperative bleeding. The study found that risk of postoperative bleeding can be assessed, allowing for better use of human resources in addressing this important adverse event after surgery.

Original languageEnglish (US)
JournalSurgery (United States)
DOIs
StateAccepted/In press - Jan 1 2018

Fingerprint

Colorectal Surgery
Hemorrhage
Electronic Health Records
Logistic Models
Ambulatory Surgical Procedures
Tertiary Care Centers
Area Under Curve
Retrospective Studies

ASJC Scopus subject areas

  • Surgery

Cite this

@article{6ff5bc2da2974a029b85c2d518a54112,
title = "Postoperative bleeding risk prediction for patients undergoing colorectal surgery",
abstract = "Background: There is limited consensus regarding risk factors for postoperative bleeding. The objective of this work was to investigate the capability of machine learning techniques in combination with practice-based longitudinal electronic medical record data for identifying potential new risk factors for postoperative bleeding and predicting patients at high risk of postoperative bleeding. Methods: A retrospective study was conducted for patients who underwent colorectal surgery 1998–2015 at a single tertiary referral center. Various predictors were extracted from electronic medical record. The outcome of interest was the occurrence of postoperative bleeding within 7 days of surgery. Logistic regression and gradient boosting machine models were trained. Area under the receiver operating curve and area under the precision recall curve were used to evaluate the performance to different models. Results: Of 13,399 cases undergoing colorectal resection, 1,680 (12.5{\%}) experienced postoperative bleeding. A total of 299 variables were evaluated. Logistic regression and gradient boosting machine models returned an area under the receiver operating curve of 0.735 and 0.822 and area under the precision recall curve of 0.287 and 0.423, respectively. In addition to well-known risk factors for postoperative bleeding, nutrition (ranked third), weakness (ranked fifth), patient mobility (ranked sixth), and activity level (ranked eighth) were found to be novel predictors in the gradient boosting machine model based on permutation importance. Conclusion: The study identified measures of functional capacity of patient as novel predictors of postoperative bleeding. The study found that risk of postoperative bleeding can be assessed, allowing for better use of human resources in addressing this important adverse event after surgery.",
author = "David Chen and Naveed Afzal and Sunghwan Sohn and Habermann, {Elizabeth B} and Naessens, {James M} and David Larson and Liu, {Hongfang D}",
year = "2018",
month = "1",
day = "1",
doi = "10.1016/j.surg.2018.05.043",
language = "English (US)",
journal = "Surgery (United States)",
issn = "0039-6060",
publisher = "Mosby Inc.",

}

TY - JOUR

T1 - Postoperative bleeding risk prediction for patients undergoing colorectal surgery

AU - Chen, David

AU - Afzal, Naveed

AU - Sohn, Sunghwan

AU - Habermann, Elizabeth B

AU - Naessens, James M

AU - Larson, David

AU - Liu, Hongfang D

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Background: There is limited consensus regarding risk factors for postoperative bleeding. The objective of this work was to investigate the capability of machine learning techniques in combination with practice-based longitudinal electronic medical record data for identifying potential new risk factors for postoperative bleeding and predicting patients at high risk of postoperative bleeding. Methods: A retrospective study was conducted for patients who underwent colorectal surgery 1998–2015 at a single tertiary referral center. Various predictors were extracted from electronic medical record. The outcome of interest was the occurrence of postoperative bleeding within 7 days of surgery. Logistic regression and gradient boosting machine models were trained. Area under the receiver operating curve and area under the precision recall curve were used to evaluate the performance to different models. Results: Of 13,399 cases undergoing colorectal resection, 1,680 (12.5%) experienced postoperative bleeding. A total of 299 variables were evaluated. Logistic regression and gradient boosting machine models returned an area under the receiver operating curve of 0.735 and 0.822 and area under the precision recall curve of 0.287 and 0.423, respectively. In addition to well-known risk factors for postoperative bleeding, nutrition (ranked third), weakness (ranked fifth), patient mobility (ranked sixth), and activity level (ranked eighth) were found to be novel predictors in the gradient boosting machine model based on permutation importance. Conclusion: The study identified measures of functional capacity of patient as novel predictors of postoperative bleeding. The study found that risk of postoperative bleeding can be assessed, allowing for better use of human resources in addressing this important adverse event after surgery.

AB - Background: There is limited consensus regarding risk factors for postoperative bleeding. The objective of this work was to investigate the capability of machine learning techniques in combination with practice-based longitudinal electronic medical record data for identifying potential new risk factors for postoperative bleeding and predicting patients at high risk of postoperative bleeding. Methods: A retrospective study was conducted for patients who underwent colorectal surgery 1998–2015 at a single tertiary referral center. Various predictors were extracted from electronic medical record. The outcome of interest was the occurrence of postoperative bleeding within 7 days of surgery. Logistic regression and gradient boosting machine models were trained. Area under the receiver operating curve and area under the precision recall curve were used to evaluate the performance to different models. Results: Of 13,399 cases undergoing colorectal resection, 1,680 (12.5%) experienced postoperative bleeding. A total of 299 variables were evaluated. Logistic regression and gradient boosting machine models returned an area under the receiver operating curve of 0.735 and 0.822 and area under the precision recall curve of 0.287 and 0.423, respectively. In addition to well-known risk factors for postoperative bleeding, nutrition (ranked third), weakness (ranked fifth), patient mobility (ranked sixth), and activity level (ranked eighth) were found to be novel predictors in the gradient boosting machine model based on permutation importance. Conclusion: The study identified measures of functional capacity of patient as novel predictors of postoperative bleeding. The study found that risk of postoperative bleeding can be assessed, allowing for better use of human resources in addressing this important adverse event after surgery.

UR - http://www.scopus.com/inward/record.url?scp=85050118631&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85050118631&partnerID=8YFLogxK

U2 - 10.1016/j.surg.2018.05.043

DO - 10.1016/j.surg.2018.05.043

M3 - Article

C2 - 30033185

AN - SCOPUS:85050118631

JO - Surgery (United States)

JF - Surgery (United States)

SN - 0039-6060

ER -