Kevin A Chen1, Chinmaya U Joisa2, Karyn B Stitzenberg1, Jonathan Stem1, Jose G Guillem1, Shawn M Gomez2, Muneera R Kapadia3. 1. Department of Surgery, University of North Carolina, 100 Manning Drive, Burnett Womack Building, Suite 4038, Chapel Hill, NC, 27599, USA. 2. Joint Department of Biomedical Engineering, University of North Carolina, 10202C Mary Ellen Jones Building, Chapel Hill, NC, 27599, USA. 3. Department of Surgery, University of North Carolina, 100 Manning Drive, Burnett Womack Building, Suite 4038, Chapel Hill, NC, 27599, USA. muneera_kapadia@med.unc.edu.
Abstract
BACKGROUND: Readmission after colorectal surgery is common and often implies complications for patients and costs for hospitals. Previous works have created predictive models using logistic regression for this outcome but have shown limited accuracy. Machine learning has shown promise in improving predictions by identifying non-linear patterns in data. We sought to create a more accurate predictive model for readmission after colorectal surgery using machine learning. METHODS: Patients who underwent colorectal surgery were identified in the National Quality Improvement Program (NSQIP) database including years 2012-2019 and split into training, validation, and test sets. The primary outcome was readmission within 30 days of surgery. Three types of machine learning models were created, including random forest (RF), gradient boosting (XGB), and neural network (NN). A logistic regression (LR) model was also created for comparison. Model performance was evaluated using area under the receiver operating characteristic curve (AUROC). RESULTS: The dataset included 213,827 patients after application of exclusion criteria. A total of 23,083 (10.8%) of patients experienced readmission. NN obtained an AUROC of 0.751 (95% CI 0.743-0.759), compared with 0.684 (95% CI 0.676-0.693) for LR. RF and XGB performed similarly with AUROCs of 0.749 (95% CI 0.741-0.757) and 0.745 (95% CI 0.737-0.753) respectively. Ileus, index admission length of stay, organ-space surgical site infection present at time of surgery, and ostomy placement were identified as the most contributory variables. CONCLUSIONS: Machine learning approaches outperformed traditional statistical methods in the prediction of readmission after colorectal surgery. After external validation, this improved prediction model could be used to target interventions to reduce readmission rate.
BACKGROUND: Readmission after colorectal surgery is common and often implies complications for patients and costs for hospitals. Previous works have created predictive models using logistic regression for this outcome but have shown limited accuracy. Machine learning has shown promise in improving predictions by identifying non-linear patterns in data. We sought to create a more accurate predictive model for readmission after colorectal surgery using machine learning. METHODS: Patients who underwent colorectal surgery were identified in the National Quality Improvement Program (NSQIP) database including years 2012-2019 and split into training, validation, and test sets. The primary outcome was readmission within 30 days of surgery. Three types of machine learning models were created, including random forest (RF), gradient boosting (XGB), and neural network (NN). A logistic regression (LR) model was also created for comparison. Model performance was evaluated using area under the receiver operating characteristic curve (AUROC). RESULTS: The dataset included 213,827 patients after application of exclusion criteria. A total of 23,083 (10.8%) of patients experienced readmission. NN obtained an AUROC of 0.751 (95% CI 0.743-0.759), compared with 0.684 (95% CI 0.676-0.693) for LR. RF and XGB performed similarly with AUROCs of 0.749 (95% CI 0.741-0.757) and 0.745 (95% CI 0.737-0.753) respectively. Ileus, index admission length of stay, organ-space surgical site infection present at time of surgery, and ostomy placement were identified as the most contributory variables. CONCLUSIONS: Machine learning approaches outperformed traditional statistical methods in the prediction of readmission after colorectal surgery. After external validation, this improved prediction model could be used to target interventions to reduce readmission rate.
Authors: Sophia Y Chen; Miloslawa Stem; Susan L Gearhart; Bashar Safar; Sandy H Fang; Nilofer S Azad; Adrian G Murphy; Amol K Narang; Christopher L Wolfgang; Jonathan E Efron Journal: World J Surg Date: 2019-10 Impact factor: 3.352
Authors: Caroline E Jones; Robert H Hollis; Tyler S Wahl; Brad S Oriel; Kamal M F Itani; Melanie S Morris; Mary T Hawn Journal: Am J Surg Date: 2016-06-01 Impact factor: 2.565
Authors: N K Francis; A Luther; E Salib; L Allanby; D Messenger; A S Allison; N J Smart; J B Ockrim Journal: Tech Coloproctol Date: 2015-06-19 Impact factor: 3.781
Authors: Rachelle N Damle; Nicole B Cherng; Julie M Flahive; Jennifer S Davids; Justin A Maykel; Paul R Sturrock; W Brian Sweeney; Karim Alavi Journal: Dis Colon Rectum Date: 2014-12 Impact factor: 4.585
Authors: Sophia Y Chen; Miloslawa Stem; Marcelo Cerullo; Joseph K Canner; Susan L Gearhart; Bashar Safar; Sandy H Fang; Jonathan E Efron Journal: Dis Colon Rectum Date: 2018-12 Impact factor: 4.585
Authors: Karel G M Moons; Douglas G Altman; Johannes B Reitsma; John P A Ioannidis; Petra Macaskill; Ewout W Steyerberg; Andrew J Vickers; David F Ransohoff; Gary S Collins Journal: Ann Intern Med Date: 2015-01-06 Impact factor: 25.391
Authors: Jeremy Albright; Farwa Batool; Robert K Cleary; Andrew J Mullard; Edward Kreske; Jane Ferraro; Scott E Regenbogen Journal: Surg Endosc Date: 2018-08-27 Impact factor: 4.584
Authors: Siavash Bolourani; Mohammad A Tayebi; Li Diao; Ping Wang; Vihas Patel; Frank Manetta; Paul C Lee Journal: J Thorac Cardiovasc Surg Date: 2020-05-29 Impact factor: 5.209