Dataset Information

Predicting Postoperative Length of Stay for Isolated Coronary Artery Bypass Graft Patients Using Machine Learning

ABSTRACT: Purpose Predictive analytics (PA) is a new trending approach in the field of healthcare that uses machine learning to build a prediction model using supervised learning algorithms. Isolated coronary artery bypass grafting (iCABG), an open-heart surgery, is commonly performed in the treatment of coronary heart disease. Aim The aim of this study was to develop and evaluate a model to predict postoperative length of stay (PLoS) for iCABG patients using supervised machine learning techniques, and to identify the features with the highest contribution to the model. Methods This is a retrospective study that uses historic data of adult patients who underwent isolated CABG (iCABG). After initial data pre-processing, data imputation using the kNN method was applied. The study used five prediction models using Naïve Bayes, Decision Tree, Random Forest, Logistic Regression and k Nearest Neighbor algorithms. Data imbalance was managed using the following widely used methods: oversampling, undersampling, “Both”, and random over-sampling examples (ROSE). The features selection process was conducted using the Boruta method. Two techniques were applied to examine the performance of the models, (70%, 30%) split and cross-validation, respectively. Models were evaluated by comparing their performance using AUC and other metrics. Results In the final dataset, six distinct features and 621 instances were used to develop the models. A total of 20 models were developed using R statistical software. The model generated using Random Forest with “Both” resampling method and cross-validation technique was deemed the best fit (AUC=0.81; F1 score=0.82; and recall=0.82). Attributes found to be highly predictive of PLoS were pulmonary artery systolic, age, height, EuroScore II, intra-aortic balloon pump used, and complications during operation. Conclusion This study demonstrates the significance and effectiveness of building a model that predicts PLoS for iCABG patients using patient specifications and pre-/intra-operative measures.

SUBMITTER: Alshakhs F

PROVIDER: S-EPMC7537993 | biostudies-literature | 2020 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundPreoperative B-type natriuretic peptide (BNP) is known to predict adverse outcomes after cardiac surgery. The value of postoperative BNP for predicting adverse outcomes is less well delineated. The authors hypothesized that peak postoperative plasma BNP (measured postoperative days 1-5) predicts hospital length of stay (HLOS) and mortality in patients undergoing primary coronary artery bypass grafting, even after adjusting for preoperative BNP and perioperative clinical risk factors.MethodsThis study is a prospective longitudinal study of 1,183 patients undergoing primary coronary artery bypass grafting surgery. Mortality was defined as all-cause death within 5 yr after surgery. Cox proportional hazards analyses were conducted to separately evaluate the associations between peak postoperative BNP and HLOS and mortality. Multivariable adjustments were made for patient demographics, preoperative BNP concentration, and clinical risk factors. BNP measurements were log10 transformed before analysis.ResultsOne hundred fifteen deaths (9.7%) occurred in the cohort (mean follow-up = 4.3 yr, range = 2.38-5.0 yr). After multivariable adjustment for preoperative BNP and clinical covariates, peak postoperative BNP predicted HLOS (hazard ratio [HR] = 1.28, 95% CI = 1.002-1.64, P = 0.049) but not mortality (HR = 1.62, CI = 0.71-3.68, P = 0.25), whereas preoperative BNP independently predicted HLOS (HR = 1.09, CI = 1.01-1.18, P = 0.03) and approached being an independent predictor of mortality (HR = 1.36, CI = 0.96-1.94, P = 0.08). When preoperative and peak postoperative BNP were separately adjusted for within the clinical multivariable models, each independently predicted HLOS (preoperative BNP HR = 1.13, CI = 1.05-1.21, P = 0.0007; peak postoperative BNP HR = 1.44, CI = 1.15-1.81, P = 0.001) and mortality (preoperative BNP HR = 1.50, CI = 1.09-2.07, P = 0.01; peak postoperative BNP HR = 2.29, CI = 1.11-4.73, P = 0.02).ConclusionsPreoperative BNP may be better than peak postoperative BNP for predicting HLOS and longer term mortality after primary coronary artery bypass grafting surgery.

Project description:BACKGROUND:To date, studies evaluating outcome improvements associated with participation in physician-led collaboratives have been limited by the absence of a contemporaneous control group. We examined post cardiac surgery pneumonia rates associated with participation in a statewide, quality improvement collaborative relative to a national physician reporting program. METHODS AND RESULTS:We evaluated 911 754 coronary artery bypass operations (July 1, 2011, to June 30, 2017) performed across 1198 hospitals participating in a voluntary national physician reporting program (Society of Thoracic Surgeons [STS]), including 33 that participated in a Michigan-based collaborative (MI-Collaborative). Unlike STS hospitals not participating in the MI-Collaborative (i.e., STSnonMI) that solely received blinded reports, MI-Collaborative hospitals received a multi-faceted intervention starting November 2012 (quarterly in-person meetings showcasing unblinded data, webinars, site visits). Eighteen of the MI-Collaborative hospitals received additional support to implement recommended pneumonia prevention practices ("MI-CollaborativePlus"), whereas 15 did not ("MI-CollaborativeOnly"). We evaluated rates of postoperative pneumonia, adjusting for patient mix and hospital effects. Baseline patient characteristics were qualitatively similar between groups and time. During the preintervention period (Q3/2011 through Q3/2012), there was no statistically significant difference in the adjusted odds of pneumonia for STS hospitals participating in the MI-Collaborative compared to the STS non-MI hospitals. However, during the intervention period (Q4/2012 through Q2/2017), there was a significant 2% reduction per quarter in the adjusted odds of pneumonia for MI-Collaborative hospitals (n=33) relative to the STS-nonMI hospitals. There was a significant 3% per quarter reduction in the adjusted odds of pneumonia for the MI-CollaborativeOnly (n=15) hospitals relative to the STS-nonMI hospitals. Over the course of the overall study period, the STS-nonMI hospitals had a 1.96% reduction in risk-adjusted pneumonia (pre- vs. intervention periods), which was less than the MI-Collaborative (3.23%, P=0.011). Over the same time period, the MI-CollaborativePlus (n=18) reduced adjusted pneumonia rates by 10.29%, P=0.001. CONCLUSIONS:Participation in a physician-led collaborative was associated with significant reductions in pneumonia relative to a national quality reporting program. Interventions including collaborative learning may yield superior outcomes relative to solely using physician feedback reporting. CLINICAL TRIAL REGISTRATION:URL: https://www.clinicaltrials.gov . Unique identifier: NCT02068716.

Dataset Information

Predicting Postoperative Length of Stay for Isolated Coronary Artery Bypass Graft Patients Using Machine Learning

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets