scholarly journals Assessment of a Modified Sandwich Estimator for Generalized Estimating Equations with Application to Opioid Poisoning in MIMIC-IV ICU Patients

Stats ◽  
2021 ◽  
Vol 4 (3) ◽  
pp. 650-664
Author(s):  
Paul Rogers ◽  
Julie Stoner
Keyword(s):  
Repeated Measures ◽  
Generalized Estimating Equations ◽  
Rare Event ◽  
Estimating Equations ◽  
Small Sample ◽  
Repeated Measurements ◽  
Regression Coefficients ◽  
Binary Outcomes ◽  
Sandwich Estimator ◽  
Generalized Estimating

Longitudinal data is encountered frequently in many healthcare research areas to include the critical care environment. Repeated measures from the same subject are expected to correlate with each other. Models with binary outcomes are commonly used in this setting. Regression models for correlated binary outcomes are frequently fit using generalized estimating equations (GEE). The Liang and Zeger sandwich estimator is often used in GEE to produce unbiased standard error estimation for regression coefficients in large sample settings, even when the covariance structure is misspecified. The sandwich estimator performs optimally in balanced designs when the number of participants is large with few repeated measurements. The sandwich estimator’s asymptotic properties do not hold in small sample and rare-event settings. Under these conditions, the sandwich estimator underestimates the variances and is biased downwards. Here, the performance of a modified sandwich estimator is compared to the traditional Liang-Zeger estimator and alternative forms proposed by authors Morel, Pan, and Mancl-DeRouen. Each estimator’s performance was assessed with 95% coverage probabilities for the regression coefficients using simulated data under various combinations of sample sizes and outcome prevalence values with independence and autoregressive correlation structures. This research was motivated by investigations involving rare-event outcomes in intensive care unit settings.

scholarly journals Simultaneous inference for multiple marginal generalized estimating equation models

2019 ◽  
Vol 29 (6) ◽  
pp. 1746-1762 ◽  
Author(s):  
Robin Ristl ◽  
Ludwig Hothorn ◽  
Christian Ritz ◽  
Martin Posch
Keyword(s):  
Estimating Equations ◽  
Estimating Equation ◽  
Small Sample ◽  
Regression Coefficients ◽  
Generalized Estimating Equation ◽  
Type I ◽  
Simultaneous Inference ◽  
Multiple Endpoints ◽  
Repeated Observations ◽  
Generalized Estimating

Motivated by small-sample studies in ophthalmology and dermatology, we study the problem of simultaneous inference for multiple endpoints in the presence of repeated observations. We propose a framework in which a generalized estimating equation model is fit for each endpoint marginally, taking into account dependencies within the same subject. The asymptotic joint normality of the stacked vector of marginal estimating equations is used to derive Wald-type simultaneous confidence intervals and hypothesis tests for multiple linear contrasts of regression coefficients of the multiple marginal models. The small sample performance of this approach is improved by a bias adjustment to the estimate of the joint covariance matrix of the regression coefficients from multiple models. As a further small sample improvement a multivariate t-distribution with appropriate degrees of freedom is specified as reference distribution. In addition, a generalized score test based on the stacked estimating equations is derived. Simulation results show strong control of the family-wise type I error rate for these methods even with small sample sizes and increased power compared to a Bonferroni-Holm multiplicity adjustment. Thus, the proposed methods are suitable to efficiently use the information from repeated observations of multiple endpoints in small-sample studies.

scholarly journals Performance of the Beta-Binomial Model for Clustered Binary Responses: Comparison with Generalized Estimating Equations

2021 ◽  
Vol 19 (1) ◽  
pp. 2-25
Author(s):  
Seongah Im
Keyword(s):  
Monte Carlo ◽  
Monte Carlo Simulations ◽  
Generalized Estimating Equations ◽  
Estimating Equations ◽  
Small Sample ◽  
Binomial Model ◽  
Sample Sizes ◽  
Binary Responses ◽  
Large Sample ◽  
Generalized Estimating

This study examined performance of the beta-binomial model in comparison with GEE using clustered binary responses resulting in non-normal outcomes. Monte Carlo simulations were performed under varying intracluster correlations and sample sizes. The results showed that the beta-binomial model performed better for small sample, while GEE performed well under large sample.

Progression of Diabetic Complications in Subgroups of People with Long Term Diabetes Type 1 According to Clinical Course

2020 ◽  
Author(s):  
Christian Gerdes ◽  
Christoph Werner ◽  
Christof Kloos ◽  
Thomas Lehmann ◽  
Gunter Wolf ◽  
...  
Keyword(s):  
Type 1 Diabetes ◽  
Generalized Estimating Equations ◽  
Microvascular Complications ◽  
Microvascular Disease ◽  
Estimating Equations ◽  
Estimating Equation ◽  
Individual Risk ◽  
Binary Outcomes ◽  
Generalized Estimating

Abstract Aims Prevention and prediction of microvascular complications are important aims of medical care in people with type 1 diabetes. Since the course of the disease is heterogenous, we tried to identify subgroups with specific risk profiles for microvascular complications. Methods Retrospective analysis of a cohort of 285 people (22637 consultations) with >10 years of type 1 diabetes. Persons were grouped into slow (<15 years), fast (>15 years) and non progressors according to the average onset of microvascular complications. Generalized estimating equations for binary outcomes were applied and pseudo coefficients of determination were calculated. Results Progression to microvascular disease was associated with age (OR: 1.034 [1.001–1.068]; p=0.04), diabetes duration (OR: 1.057 [1.021–1.094]; p=0.002), HbA1c (OR: 1.035 [1.011–1.060]; p=0.005), BMI (OR: 0.928 [0.866–0.994]; p=0.034) and the social strata index (OR: 0.910 [0.830–0.998]; p=0.046). Generalized estimating equations predicted 31.02% and exclusion of HbA1c marginally reduced the value to 28.88%. The proportion of patients with LADA was higher in fast than slow progressors [13 (26.5%) vs. 14 (11.9%); p=0.019]. A generalized estimating equation comparing slow to fast progressors revealed no significant markers. Conclusion In our analysis, we were able to confirm known risk factors for microvascular disease in people with type 1 diabetes. Overall, prediction of individual risk was difficult, the effect of individual markers minor and we could not find differences regarding slow or fast progression. We therefore emphasis the need for additional markers to predict individual risk for microvascular disease.

Analyzing Cross-Sectionally Clustered Data Using Generalized Estimating Equations

2021 ◽  
pp. 107699862110174
Author(s):  
Francis L. Huang
Keyword(s):  
Least Squares ◽  
Generalized Estimating Equations ◽  
Multilevel Models ◽  
Clustered Data ◽  
Estimating Equations ◽  
Worked Examples ◽  
Ordinary Least Squares ◽  
Binary Outcomes ◽  
Similarities And Differences ◽  
Generalized Estimating

The presence of clustered data is common in the sociobehavioral sciences. One approach that specifically deals with clustered data but has seen little use in education is the generalized estimating equations (GEEs) approach. We provide a background on GEEs, discuss why it is appropriate for the analysis of clustered data, and provide worked examples using both continuous and binary outcomes. Comparisons are made between GEEs, multilevel models, and ordinary least squares results to highlight similarities and differences between the approaches. Detailed walkthroughs are provided using both R and SPSS Version 26.

Sign in / Sign up

Export Citation Format

Share Document