Author + information
Background: Mortality due to ST-elevation myocardial infarction (STEMI) is higher in women compared with men. We aimed to develop and validate prediction models for in-hospital mortality in women with STEMI. In particular, we compared traditional logistic regression (LR) with a machine learning approach, the random forest (RF) model.
Methods: We used the 2011—2013 National Inpatient Sample (NIS) data and identified women admitted with STEMI. The main outcome was all-cause in-hospital mortality. Based on the 2012 dataset, patients were divided randomly into 80% development and 20% validation cohorts. Models were trained on the development set, internally validated using the validation set, and externally validated using the 2011 and 2013 data. A multivariate LR, a full and a reduced RF models were developed and compared.
Results: In the multivariate LR, 11 variables were included in the model based on backward elimination. The full RF model contained 32 variables, and the reduced model contained 17 variables selected based on individual variable importance. In the internal validation cohort, the C-index was comparable for the 3 models (Figure). The models showed good stability in the external validation cohorts with C-index for the LR, full, and reduced RF models of 0.84, 0.85, and 0.81 for year 2011, and 0.82, 0.81, and 0.81 for year 2013, respectively.
Conclusions: RF was comparable to LR in predicting in-hospital mortality in women with STEMI, and can be a useful and accurate tool to use in clinical practice.
Poster Hall, Hall C
Saturday, March 18, 2017, 9:45 a.m.-10:30 a.m.
Session Title: Cardiac Arrest, Diabetes, and Other High Risk Features of Patients With Acute Coronary Syndrome
Abstract Category: 2. Acute and Stable Ischemic Heart Disease: Clinical
Presentation Number: 1204-325
- 2017 American College of Cardiology Foundation