Different methods for handling incomplete longitudinal binary outcome due to missing at random dropout

Satty, Ali; Mwambi, Henry; Molenberghs, Geert

doi:10.1016/j.stamet.2014.10.002

Cited by 8 publications

(8 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To deal with missing data, multiple imputation based on fully conditional specification was performed to allow data to be missing at random (MAR). Although GEE models are typically taken to allow data to be missing completely at random, Satty et al [ 24 ] show that GEE with multiple imputation can perform well under the assumption of MAR, conditional on the important predictors of missingness being included in the model. Here, age, sex and informant were included as predictors of missingness.…”

Section: Methodsmentioning

confidence: 99%

Identifying disordered eating behaviours in adolescents: how do parent and adolescent reports differ by sex and age?

Bartholdy

Allen

Hodsoll

et al. 2017

Eur Child Adolesc Psychiatry

View full text Add to dashboard Cite

This study investigated the prevalence of disordered eating cognitions and behaviours across mid-adolescence in a large European sample, and explored the extent to which prevalence ratings were affected by informant (parent/adolescent), or the sex or age of the adolescent. The Development and Well-Being Assessment was completed by parent–adolescent dyads at age 14 (n = 2225) and again at age 16 (n = 1607) to explore the prevalence of 7 eating disorder symptoms (binge eating, purging, fear of weight gain, distress over shape/weight, avoidance of fattening foods, food restriction, and exercise for weight loss). Informant agreement was assessed using kappa coefficients. Generalised estimating equations were performed to explore the impact of age, sex and informant on symptom prevalence. Slight to fair agreement was observed between parent and adolescent reports (kappa estimates between 0.045 and 0.318); however, this was largely driven by agreement on the absence of behaviours. Disordered eating behaviours were more consistently endorsed amongst girls compared to boys (odds ratios: 2.96–5.90) and by adolescents compared to their parents (odds ratios: 2.71–9.05). Our data are consistent with previous findings in epidemiological studies. The findings suggest that sex-related differences in the prevalence of disordered eating behaviour are established by mid-adolescence. The greater prevalence rates obtained from adolescent compared to parent reports may be due to the secretive nature of the behaviours and/or lack of awareness by parents. If adolescent reports are overlooked, the disordered behaviour may have a greater opportunity to become more entrenched.

show abstract

Section: Methodsmentioning

confidence: 99%

Identifying disordered eating behaviours in adolescents: how do parent and adolescent reports differ by sex and age?

Bartholdy

Allen

Hodsoll

et al. 2017

Eur Child Adolesc Psychiatry

View full text Add to dashboard Cite

show abstract

“…A common source of missing data in MET data is the selection of preferred phenotypes (Im et al, 1989). Targeted selection can be based on the genotypes' performance or their breeding values (Im et al, 1989), estimated either in longitudinal studies (Satty et al, 2015) or in MET across time. Alternatively, covariates or correlated data can be used for such a selection (Appel et al, 1996(Appel et al, , 1998.…”

Section: Introductionmentioning

confidence: 99%

Effect of missing values in multi‐environmental trials on variance component estimates

Hartung

Piepho

2021

Crop Science

View full text Add to dashboard Cite

A common task in the analysis of multi-environmental trials (MET) by linear mixed models (LMM) is the estimation of variance components (VCs). Most often, MET data are imbalanced, e.g., due to selection. The imbalance mechanism can be missing completely at random (MCAR), missing at random (MAR), or missing not at random (MNAR). If the missing-data pattern in MET was caused by selection, it is usually MAR. In this case, likelihood-based methods are the preferred methods for analysis as they can account for a MAR data pattern. Likelihood-based methods used to estimate VCs in LMM have the This article is protected by copyright. All rights reserved. 2property that all VC estimates are constrained to be non-negative and thus the estimators are generally biased. Therefore, there are two potential causes of bias in MET analysis: a data pattern not being MCAR and the bias of likelihood-based VC estimators. The current study tries to dissect and quantify both possible sources of bias. A simulation study with MET data typical for cultivar evaluation trials was conducted, in which the missing data pattern and the size of VCs were varied. The results showed that for the simulated MET, bias in VC estimates was similar under MCAR and MAR. Thus, the bias is solely due to the likelihoodbased estimation. Bias increase when increasing the ratio of genotype variance to error variance is small. Bias was similar for MAR and MCAR data patterns. Thus, it may be concluded that selection does not increase bias in VC estimation.

show abstract

“…Incomplete records were generally excluded, and the issue of missing value imputation appears to have been only addressed in one study [ 26 ]. Indeed, imputation methods for correlated data might be less popular than standard approaches; in general, likelihood approaches (such as GLMM) have shown to be robust to the “missing completely at random” assumption, while non-likelihood marginal models (such as GEE models) have not [ 55 ].…”

Section: Discussionmentioning

confidence: 99%

A Critical Review of Statistical Methods for Twin Studies Relating Exposure to Early Life Health Conditions

Fasola

Montalbano

Cilluffo

et al. 2021

IJERPH

View full text Add to dashboard Cite

When investigating disease etiology, twin data provide a unique opportunity to control for confounding and disentangling the role of the human genome and exposome. However, using appropriate statistical methods is fundamental for exploiting such potential. We aimed to critically review the statistical approaches used in twin studies relating exposure to early life health conditions. We searched PubMed, Scopus, Web of Science, and Embase (2011–2021). We identified 32 studies and nine classes of methods. Five were conditional approaches (within-pair analyses): additive-common-erratic (ACE) models (11 studies), generalized linear mixed models (GLMMs, five studies), generalized linear models (GLMs) with fixed pair effects (four studies), within-pair difference analyses (three studies), and paired-sample tests (two studies). Four were marginal approaches (unpaired analyses): generalized estimating equations (GEE) models (five studies), GLMs with cluster-robust standard errors (six studies), GLMs (one study), and independent-sample tests (one study). ACE models are suitable for assessing heritability but require adaptations for binary outcomes and repeated measurements. Conditional models can adjust by design for shared confounders, and GLMMs are suitable for repeated measurements. Marginal models may lead to invalid inference. By highlighting the strengths and limitations of commonly applied statistical methods, this review may be helpful for researchers using twin designs.

show abstract

Different methods for handling incomplete longitudinal binary outcome due to missing at random dropout

Cited by 8 publications

References 27 publications

Identifying disordered eating behaviours in adolescents: how do parent and adolescent reports differ by sex and age?

Identifying disordered eating behaviours in adolescents: how do parent and adolescent reports differ by sex and age?

Effect of missing values in multi‐environmental trials on variance component estimates

A Critical Review of Statistical Methods for Twin Studies Relating Exposure to Early Life Health Conditions

Contact Info

Product

Resources

About