Dataset Information

Regression trees for predicting mortality in patients with cardiovascular disease: what improvement is achieved by using ensemble-based methods?

ABSTRACT: In biomedical research, the logistic regression model is the most commonly used method for predicting the probability of a binary outcome. While many clinical researchers have expressed an enthusiasm for regression trees, this method may have limited accuracy for predicting health outcomes. We aimed to evaluate the improvement that is achieved by using ensemble-based methods, including bootstrap aggregation (bagging) of regression trees, random forests, and boosted regression trees. We analyzed 30-day mortality in two large cohorts of patients hospitalized with either acute myocardial infarction (N = 16,230) or congestive heart failure (N = 15,848) in two distinct eras (1999-2001 and 2004-2005). We found that both the in-sample and out-of-sample prediction of ensemble methods offered substantial improvement in predicting cardiovascular mortality compared to conventional regression trees. However, conventional logistic regression models that incorporated restricted cubic smoothing splines had even better performance. We conclude that ensemble methods from the data mining and machine learning literature increase the predictive performance of regression trees, but may not lead to clear advantages over conventional logistic regression models for predicting short-term mortality in population-based samples of subjects with cardiovascular disease.

SUBMITTER: Austin PC

PROVIDER: S-EPMC3470596 | biostudies-other | 2012 Sep

REPOSITORIES: biostudies-other

ACCESS DATA

Similar Datasets

Project description:BackgroundCardiovascular disease is an important driver of the increased mortality associated with chronic kidney disease (CKD). Higher left ventricular mass (LVM) predicts increased risk of adverse cardiovascular outcomes and total mortality, but previous reviews have shown no clear association between intervention-induced LVM change and all-cause or cardiovascular mortality in CKD.MethodsThe primary objective of this meta-analysis was to investigate whether treatment-induced reductions in LVM over periods ≥12 months were associated with all-cause mortality in patients with CKD. Cardiovascular mortality was investigated as a secondary outcome. Measures of association in the form of relative risks (RRs) with associated variability and precision (95% confidence intervals [CIs]) were extracted directly from each study, when reported, or were calculated based on the published data, if possible, and pooled RR estimates were determined.ResultsThe meta-analysis included 42 trials with duration ≥12 months: 6 of erythropoietin stimulating agents treating to higher vs. lower hemoglobin targets, 10 of renin-angiotensin-aldosterone system inhibitors vs. placebo or another blood pressure lowering agent, 14 of modified hemodialysis regimens, and 12 of other types of interventions. All-cause mortality was reported in 121/2584 (4.86%) subjects in intervention groups and 168/2606 (6.45%) subjects in control groups. The pooled RR estimate of the 27 trials ≥12 months with ≥1 event in ≥1 group was 0.72 (95% CI 0.57 to 0.90, p = 0.005), with little heterogeneity across studies. Directionalities of the associations in intervention subgroups were the same. Sensitivity analyses of ≥6 months (34 trials), ≥9 months (29 trials), and >12 months (10 trials), and including studies with no events in either group, demonstrated similar risk reductions to the primary analysis. The point estimate for cardiovascular mortality was similar to all-cause mortality, but not statistically significant: RR 0.67, 95% CI 0.39 to 1.16.ConclusionsThese results suggest that LVM regression may be a useful surrogate marker for benefits of interventions intended to reduce mortality risk in patients with CKD.

Dataset Information

Regression trees for predicting mortality in patients with cardiovascular disease: what improvement is achieved by using ensemble-based methods?

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets