Bootstrap Methods for Developing Predictive Models
The American StatisticianPublished 1 May 2004
Peter C. Austin, Jack V. Tu
Citations620
SJR quartileQ1
SJR score0.78
SNIP1.50
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
Using data on patients admitted to hospital with a heart attack, it is demonstrated that selecting those variables that were identified as independent predictors of mortality in at least 60%% of the bootstrap samples resulted in a parsimonious model with excellent predictive ability.
Abstract
Researchers frequently use automated model selection methods such as backwards elimination to identify variables that are independent predictors of an outcome under consideration. We propose using bootstrap resampling in conjunction with automated variable selection methods to develop parsimonious prediction models. Using data on patients admitted to hospital with a heart attack, we demonstrate that selecting those variables that were identified as independent predictors of mortality in at least 60%% of the bootstrap samples resulted in a parsimonious model with excellent predictive ability.
Keywords
MathematicsEnvironmental Science
An Introduction to the Bootstrap
39,744 Citations1994Bradley Efron, Robert Tibshirani
BiometricsComparing the Areas under Two or More Correlated Receiver Operating Characteristic Curves: A Nonparametric Approach
22,316 Citations1988Elizabeth R. DeLong, David M. DeLong +1 more
A nonparametric approach to the analysis of areas under correlated ROC curves is presented, by using the theory on generalized U-statistics to generate an estimated covariance matrix.
RadiologyThe meaning and use of the area under a receiver operating characteristic (ROC) curve.
21,812 Citations1982J A Hanley, Barbara J. McNeil
A representation and interpretation of the area under a receiver operating characteristic (ROC) curve obtained by the "rating" method, or by mathematical predictions based on patient characteristics, is presented and it is shown that in such a setting the area represents the probability that a randomly chosen diseased subject is (correctly) rated or ranked with greater suspicion than a random chosen non-diseased subject.
Statistics in MedicineA COMPARISON OF GOODNESS-OF-FIT TESTS FOR THE LOGISTIC REGRESSION MODEL
1,805 Citations1997D Hosmer, Trina Hosmer +2 more
An examination of the performance of the tests when the correct model has a quadratic term but a model containing only the linear term has been fit shows that the Pearson chi-square, the unweighted sum-of-squares, the Hosmer-Lemeshow decile of risk, the smoothed residual sum- of-Squares and Stukel's score test, have power exceeding 50 per cent to detect moderate departures from linearity.
CirculationTIMI Risk Score for ST-Elevation Myocardial Infarction: A Convenient, Bedside, Clinical Score for Risk Assessment at Presentation
1,575 Citations2000David A. Morrow, Elliott M. Antman +7 more
The TIMI risk score for STEMI captures the majority of prognostic information offered by a full logistic regression model but is more readily used at the bedside, likely to be clinically useful in the triage and management of fibrinolytic-eligible patients with STEMI.
CirculationPredictors of 30-Day Mortality in the Era of Reperfusion for Acute Myocardial Infarction
1,093 Citations1995Kerry L. Lee, Lynn H. Woodlief +8 more
Through consideration of multiple characteristics, including age, medical history, physiological significance of the infarction, and medical treatment, the prognosis of an individual patient can be accurately estimated.
CirculationPredictors of Outcome in Patients With Acute Coronary Syndromes Without Persistent ST-Segment Elevation
962 Citations2000Eric Boersma, Karen S. Pieper +12 more
The occurrence of adverse events after presentation with acute coronary syndromes is affected by multiple factors and these factors should be considered in the clinical decision-making process.
British Journal of Mathematical and Statistical PsychologyBackward, forward and stepwise automated subset selection algorithms: Frequency of obtaining authentic and noise variables
782 Citations1992Shelley Derksen, H. J. Keselman
The use of automated subset search algorithms is reviewed and issues concerning model selection and selection criteria are discussed and a Monte Carlo study is reported which presents data regarding the frequency with which authentic and noise variables are selected by automated subset algorithms.
Statistics in MedicineA bootstrap resampling procedure for model building: Application to the cox regression model
599 Citations1992Willi Sauerbrei, Martin Schumacher
A bootstrap-model selection procedure is developed, combining the bootstrap method with existing selection techniques such as stepwise methods, for the selection of variables in the framework of a regression model which might influence the outcome variable.
Selection of subsets of regression variables
347 Citations1984Alan J. Miller, J. Miller
Journal of the American College of CardiologyDevelopment and validation of the ontario acute myocardial infarction mortality prediction rules
299 Citations2001Jack V. Tu, Peter C. Austin +4 more
The Ontario AMI mortality prediction rules predict quite accurately 30-day and one-year mortality after an AMI in linked hospital discharge databases of AMI patients from Ontario, Manitoba and California.
The LancetA simple risk index for rapid initial triage of patients with ST-elevation myocardial infarction: an InTIME II substudy
293 Citations2001David A. Morrow, Elliott M. Antman +7 more
A simple risk index based on characteristics easily assessed by any paramedical or clinical personnel captures most of the information from more complex tools, and is likely to be useful in the rapid triage of patients with STEMI outside hospital or on first arrival in the hospital.
CirculationPREDICT: A Simple Risk Score for Clinical Severity and Long-Term Prognosis After Hospitalization for Acute Myocardial Infarction or Unstable Angina
214 Citations1999David R. Jacobs, Candyce H. Kroenke +5 more
The simple PREDICT risk score was a powerful prognosticator of 6-year mortality after hospitalization and declined 32% between 1970 and 1990.
The American StatisticianFrequency of Selecting Noise Variables in Subset Regression Analysis: A Simulation Study
157 Citations1987Virginia F. Flack, Potter C. Chang
CirculationComparing AMI Mortality Among Hospitals in Patients 65 Years of Age and Older
155 Citations1999Harlan M. Krumholz, Jersey Chen +4 more
A simple 7-variable risk model performs as well as more complex models in comparing hospital outcomes for acute myocardial infarction and provides a basis for hospitals to develop a simple approach to compare outcomes.
Statistics in MedicineA comparison of statistical learning methods on the GUSTO database
89 Citations1998Marguerite Ennis, Geoffrey E. Hinton +3 more
A battery of modern, adaptive non-linear learning methods are applied to a large real database of cardiac patient data and it is found that none of the methods could outperform a relatively simple logistic regression model previously developed for this problem.
Communications in Statistics - Simulation and ComputationMethods of variable selection in regression modeling
53 Citations1998Paul A. Murtaugh
Simulation was used to evaluate the performances of several methods of variable selection in regression modeling: stepwise regression based on partial F-tests, stepwise minimization of Mallows’ C p statistic and Schwarz’s Bayes Information Criterion, and regression trees constructed with two kinds of pruning.