Multiple imputation compared with restricted pseudo‐likelihood and generalized estimating equations for analysis of binary repeated measures in clinical studies

Lipkovich, Ilya; Duan, Yuyan; Ahmed, Saeeduddin

doi:10.1002/pst.188

Cited by 19 publications

(17 citation statements)

References 30 publications

(45 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The MI specification of this study is described as follows: due to the large portion of missing values, the MI specification included 50 imputations, which is reported to produce consistent parameter estimates in direction and magnitude (Horton & Lipsit, 2001) and is often used in many studies with large datasets (e.g., (Kamakura & Wedel, 2000) and (Lipkovich, Duan, & Ahmed, 2005). All study variables were entered in the imputation process to maximize all of the covariate information.…”

Section: -2: Checking Missing Patternsmentioning

confidence: 99%

Racial/Ethnic Inequality Among Older Workers: Focusing On Whites, Blacks, and Latinos Within the Cumulative Advantage/Disadvantage Framework

Choi

Tang

Copeland

2016

Journal of Social Service Research

View full text Add to dashboard Cite

As the number of older workers increases, scholars from variable fields have made an effort to understand what makes the current older generation want to remain in the labor market; however, our knowledge of older minority workers is fragmentary. The experience of older minority workers, defined as Blacks and Latinos in this study, may differ from that of their White counterparts because of persistent racial/ethnic differences among older workers, especially in four domains: 1) financial resources for retirement, 2) health, 3) career patterns, and 4) education. Existing literature has little to offer for older minority workers because of limited methodological approaches and a dearth of appropriate theories. Thus, this study aims to identify factors that explain older Americans' labor market participation after retirement age and whether there are racial/ethnic differences among those factors. It also presents fragmentary knowledge from existing studies within the framework of the cumulative advantage/disadvantage (CAD) theory.Using the 1,114 pre-retirement aged workers from the 2004 and 2008 datasets of the Health and Retirement Study (HRS), probit analysis was run to estimate the analytic model. RACIAL/ETHNIC INEQUALITY AMONG OLDER WORKERS: FOCUSING ON WHITES, BLACKS, AND LATINOS WITHIN THE CUMULATIVE ADVANTAGE/DISADVANTAGE FRAMEWORKEunhee Choi, M.S.W., Ph.D.University of Pittsburgh, 2013 v Racially separate analyses were performed to treat race as a major indicator for the outcome (workers' labor market status after age 65) and to systematically compare factors by race.The results showed that, after adjusting for covariates, factors influencing older individuals' employment after age 65 were conditioned by race. Whereas health and the meaning of work most influence Whites, financial resources (particularly home ownership) most affect Blacks, while health alone most concerns Latinos. Overall, this study suggests that access to the labor market-as opposed to personal preference, financial necessity, or job dissatisfactionmost influences the labor market participation of minorities.These findings of this study highlight the role of race, suggesting that one universal model cannot fully explain older individuals' labor market participation among different race/ethnic groups. Furthermore, this study's inclusion of Latinos expanded the scope of the existing literature, finding that both Black and Latino workers experience a lifetime accumulation of disadvantages. However, their employment experiences differ significantly: while class differences potentially exist among Black workers, Latinos can be more disadvantaged as a whole, experiencing substantial barriers related to education level and language in the workplace.This study helps to correct theoretical limitations in studies of older workers by using the more revealing CAD framework to interpret results. Racial/ethnic inequality among older workers should be understood not as sudden occurrences in old age, but as a byproduct of the interplay betwe...

show abstract

Section: -2: Checking Missing Patternsmentioning

confidence: 99%

Racial/Ethnic Inequality Among Older Workers: Focusing On Whites, Blacks, and Latinos Within the Cumulative Advantage/Disadvantage Framework

Choi

Tang

Copeland

2016

Journal of Social Service Research

View full text Add to dashboard Cite

show abstract

“…In future research we intend to apply our suggested missing imputation technique to repeated models using generalized estimation equations. 44,45 Our study demonstrated a successful application of statistical theory in imputing missing data to psychophysiology research. Psychiatrists and clinicians should find the suggested method accessible and useful for attaining meaningful conclusions when some data are missing.…”

Section: Discussionmentioning

confidence: 84%

“…Our model (Equation 1) based on the literature in psychophysiology did not account for the repeated measure structure of the heart rate data with a possible inflation of standard errors, in other words, deflation of P ‐values. In future research we intend to apply our suggested missing imputation technique to repeated models using generalized estimation equations 44,45 …”

Section: Discussionmentioning

confidence: 99%

Heart rate reactivity and current post‐traumatic stress disorder when data are missing

Jeon–Slaughter¹,

Tucker²,

Pfefferbaum³

et al. 2011

Psychiatry Clin Neurosci

View full text Add to dashboard Cite

Aims: This study demonstrates that auxiliary and exclusion criteria variables increase the effectiveness of missing imputation in correcting underestimation of physiologic reactivity in relation to post‐traumatic stress disorder (PTSD) caused by deleting cases with missing physiologic data. Methods: This study used data from survivors of the 1995 Oklahoma City bombing and imputed missing heart rate data using auxiliary and exclusion criteria variables. Logistic regression was used to examine heart rate reactivity in relation to current PTSD. Results: Of 113 survivors who participated in the bombing study's 7‐year follow‐up interview, 42 (37%) had missing data on heart rate reactivity due to exclusion criteria (medical illness or use of cardiovascular or psychotropic medications) or non‐participation. Logistic regression results based on imputed heart rate data using exclusion criteria and auxiliary (the presence of any current PTSD arousal symptoms) variables showed that survivors with current bombing‐related PTSD had significantly higher heart rates at baseline and recovered more slowly back to baseline heart rate during resting periods than survivors without current PTSD, while results based on complete cases failed to show significant correlations between current PTSD and heart rates at any assessment points. Conclusions: Suggested methods yielded an otherwise undetectable link between physiology and current PTSD.

show abstract

“…Likelihood‐based methods, such as mixed models for repeated measures (MMRM), do not generate explicit predictions of unobserved outcomes but model them implicitly through estimated within‐subject correlations between time points. The MMRM approach is commonly used for continuous outcomes, whereas generalized linear mixed models (GLMM) can be used for binary endpoints . For continuous outcomes, MMRM is easy to implement and typically yields efficient and consistent estimates.…”

Section: Choice Of Estimands In Long‐term Studiesmentioning

confidence: 99%

“…The MMRM approach is commonly used for continuous outcomes, 18 whereas generalized linear mixed models (GLMM) can be used for binary endpoints. 19 For continuous outcomes, MMRM is easy to implement and typically yields efficient and consistent estimates.…”

Section: Analysis Considerationsmentioning

confidence: 99%

Points to consider for analyzing efficacy outcomes in long‐term extension clinical trials

Ratitch¹,

Lipkovich

Erickson

et al. 2018

Pharmaceutical Statistics

Self Cite

View full text Add to dashboard Cite

This article focuses on 2 objectives in the analysis of efficacy in long-term extension studies of chronic diseases: (1) defining and discussing estimands of interest in such studies and (2) evaluating the performance of several multiple imputation methods that may be useful in estimating some of these estimands. Specifically, 4 estimands are defined and their clinical utility and inferential ramifications discussed. The performance of several multiple imputation methods and approaches were evaluated using simulated data. Results suggested that when interest is in a binary outcome derived from an underlying continuous measurement, it is preferable to impute the underlying continuous value that is subsequently dichotomized rather than to directly impute the binary outcome. Results also demonstrated that multivariate Gaussian models with Markov chain Monte Carlo imputation and sequential regression have minimal bias and the anticipated confidence interval coverage, even in settings with ordinal data where departures from normality are a concern. These approaches are further illustrated using a long-term extension study in psoriasis.

show abstract

Multiple imputation compared with restricted pseudo‐likelihood and generalized estimating equations for analysis of binary repeated measures in clinical studies

Cited by 19 publications

References 30 publications

Racial/Ethnic Inequality Among Older Workers: Focusing On Whites, Blacks, and Latinos Within the Cumulative Advantage/Disadvantage Framework

Racial/Ethnic Inequality Among Older Workers: Focusing On Whites, Blacks, and Latinos Within the Cumulative Advantage/Disadvantage Framework

Heart rate reactivity and current post‐traumatic stress disorder when data are missing

Points to consider for analyzing efficacy outcomes in long‐term extension clinical trials

Contact Info

Product

Resources

About