Working‐correlation‐structure identification in generalized estimating equations

Hin, Lin-Yee; Wang, You‐Gan

doi:10.1002/sim.3489

Cited by 147 publications

(153 citation statements)

References 16 publications

Supporting

Mentioning

148

Contrasting

Unclassified

Order By: Relevance

“…For example, Pan (2001a) proposed a modification of Akaike's information criterion (AIC), called the "quasi-likelihood under the independence model criterion (QIC)." In addition, Hin and Wang (2009) proposed a correlation information criterion (CIC) that modifies the QIC and substantially improves its performance.…”

mentioning

confidence: 99%

Modifications of QIC and CIC for Selecting a Working Correlation Structure in the Generalized Estimating Equation Method

2011

View full text Add to dashboard Cite

The generalized estimating equation (GEE) method is a popular method for analyzing longitudinal data. An inappropriate specification of the working correlation structure reduces the efficiency of the GEE estimation. Pan (2001a) and Hin and Wang (2009) proposed a quasi-likelihood under the independence model criterion (QIC) and a correlation information criterion (CIC) for selecting a proper working correlation structure, respectively. In this study, we proposed modifications to the QIC and CIC using the variance estimators of the GEE with improved small-sample properties. In a simulation study, the performance of the modified QIC and CIC was better than that of the original QIC and CIC. The modified methods were illustrated using the data for an air pollution study.

show abstract

mentioning

confidence: 99%

Modifications of QIC and CIC for Selecting a Working Correlation Structure in the Generalized Estimating Equation Method

2011

View full text Add to dashboard Cite

show abstract

“…In the present study, the models were fit with the working correlation matrix structures: autoregressive, uniform, independent and unstructured. The choice was made using the correlation information criterion (CIC) proposed by Hin and Wang (2009), considering that the correlation matrix with the lower CIC is the most appropriate.…”

Section: Discussionmentioning

confidence: 99%

“…All statistical analyses were performed with the aid of R statistical software (R DEVELOPMENT CORE TEAM, 2011), and the gee (CAREY et al, 2012) and geepack packages (HØJSGAARD et al, 2006;YAN, 2002;YAN;FINE, 2004) were used to fit the GEE models.…”

Section: Discussionmentioning

confidence: 99%

Risk factors for diarrhea in a cohort of dairy heifer calves reared in individual outdoor hutches in Piraí, Rio de Janeiro, Brazil

Fagundes

Vidal

Alves³

et al. 2016

SCA

View full text Add to dashboard Cite

This study aimed to evaluate the association between diarrhea and risk factors potentially related to diarrhea incidence, such as Cryptosporidium, Giardia, Eimeria and nematode infection, animal age, failure to transfer passive immunity, type of fertilization (artificial insemination - AI, in vitro production of embryos - IVP and natural breeding), mother (heifer/cow) and birth weight of heifer calves reared in individual outdoor hutches on a dairy farm. The database used in the present study refers to a weekly survey of a cohort of 118 heifer calves from birth to 100 days of age. A logistic generalized additive model (GAM) was fit to evaluate the functional relationship between the risk of diarrhea occurrence and birth weight. For analysis purposes, the database was divided into calves up to 30 days old and calves 31 to 100 days old. Bivariate and multivariate analyses were performed using logistic regression models via generalized estimating equations. All variables were included in the models and were removed oneby-one according to Wald statistics to obtain the final model. The relative risks (RR) and their respective 90% confidence intervals were estimated. The final model for calves up to 30 days old showed that IVP [RR=0.33 (90% CI 0.14; 0.75)] is a protective factor and that Cryptosporidium spp. infection [RR=1.95 (90% CI 1.22; 3.12)] is a risk factor. Although not statistically significant [RR=1.46 (90% CI 0.98; 2.18)], failure to transfer passive immunity increases the risk of diarrhea. The final model for calves 31 - 100 days old showed that IVP [RR=0.50 (90% IC 0.28; 0.89)] and birth weight ? 40 kg [RR=0.65 (90% CI 0.47; 0.89)] are protective factors. In this age group, Cryptosporidium spp. infection has no statistical significance [RR=1.25 (90% CI 0.95; 1.64)]. These results indicate that some of the diarrhea episodes observed in animals up to 30 days old could be explained by Cryptosporidium infection; therefore, better monitoring of these animals is recommended. Greater care in colostrum administration is suggested because animals with hypoglobulinemia are more sensitive to diarrhea occurrence. Although attention should be given to all animals with a low birth weight, special care should be given to those 31 to 100 days old.

show abstract

“…GEEs provide consistent parameter and standard error estimates even if the correlation structure among responses is not specified correctly. However, Hin and Wang (2009) and Wang and Carey (2003) showed that the misspecification of working correlation structure may cause inaccurate results of the parameter estimates especially in small sample sizes. 9,10 Besides the GEEs, generalized linear mixed model (GLMM) provides a flexible approach to handle longitudinal data from exponential family distributions (normal, binomial or Poisson) by using link functions.…”

mentioning

confidence: 99%

“…However, Hin and Wang (2009) and Wang and Carey (2003) showed that the misspecification of working correlation structure may cause inaccurate results of the parameter estimates especially in small sample sizes. 9,10 Besides the GEEs, generalized linear mixed model (GLMM) provides a flexible approach to handle longitudinal data from exponential family distributions (normal, binomial or Poisson) by using link functions. GLMM is an extension of generalized linear models including the random effects, which represents the influence of each subject on his/her own repeated responses.…”

mentioning

confidence: 99%

Generalized Estimating Equations Approach to Multi-Way Contingency Tables with Longitudinal Data in Case-Control Studies

Bahçecitapar¹,

Altunay²

2017

Turkiye Klinikleri J Biostat

View full text Add to dashboard Cite

Objective: Log-linear analysis is a classical statistical method for the analysis of association between variables in multi-way contingency tables. Generalized Estimating Equations (GEEs) approach is popular especially for analyzing longitudinal data, since it enables to take into account the correlation of repeated measures over time within subjects by defining a so-called "working correlation structure". GEEs provide consistent regression parameter estimates even if working correlation structure is misspecified. In this paper, we suggest GEEs approach to the analysis of a multi-way contingency table with longitudinal data, which consists of more than one contingency tables obtained over time in case-control studies and examine the method of GEEs by considering four different working correlation structures for correlations between longitudinal count data in the table. Material and Methods: Log-linear analysis and GEEs method for longitudinal data in the multi-way contingency table with time factor are performed by SAS 9.4 statistical software program. A real genetic association case-control study with longitudinal data was illustrated to compare both methods. Results: Using either the classical log-linear analysis or GEEs method with an independent (IND) working correlation structure generates similar results for parameter estimates. It is found that linear model fitting longitudinal data in the multi-way contingency table is observed to be same for both log-linear analysis and GEEs approach performed under all correlation structures. Conclusion: This study indicates that GEEs approach provides more efficient and unbiased regression parameter estimates for the multi-way contingency table designed by responses measured repeatedly over time in the case-control study.

show abstract

Working‐correlation‐structure identification in generalized estimating equations

Cited by 147 publications

References 16 publications

Modifications of QIC and CIC for Selecting a Working Correlation Structure in the Generalized Estimating Equation Method

Modifications of QIC and CIC for Selecting a Working Correlation Structure in the Generalized Estimating Equation Method

Risk factors for diarrhea in a cohort of dairy heifer calves reared in individual outdoor hutches in Piraí, Rio de Janeiro, Brazil

Generalized Estimating Equations Approach to Multi-Way Contingency Tables with Longitudinal Data in Case-Control Studies

Contact Info

Product

Resources

About