Adjusting for confounding by cluster using generalized linear mixed models

Brumback, Babette; Dailey, Amy B.; Brumback, Lyndia C.; Livingston, Melvin D.; He, Zhulin

doi:10.1016/j.spl.2010.07.006

Cited by 53 publications

(69 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Interestingly, the estimates for the adjusted OR describing the relationship between area-level SEP and repeat mammography were similar regardless of method used. Because there are potential complications and limitations with both of these methods, methodological advancements in the area of multilevel modeling with complex survey data are needed (21,23). A limitation of ordinary logistic regression with survey data is its inability to capture an unmeasured neighborhood effect; that is, were we interested in an overall effect of neighborhood, we would need to employ generalized linear mixed models.…”

Section: Discussionmentioning

confidence: 99%

Area-Level Socioeconomic Position and Repeat Mammography Screening Use: Results from the 2005 National Health Interview Survey

Dailey

Brumback

Livingston

et al. 2011

Cancer Epidemiology, Biomarkers &Amp; Prevention

Self Cite

View full text Add to dashboard Cite

Background: Despite a considerable number of studies describing the relationship between area-level socioeconomic conditions and mammography screening, definitive conclusions have yet to be drawn. The aim of this study was to examine the relationship between area-level socioeconomic position (SEP) and repeat mammography screening, using nationwide U.S. census SEP data linked to a nationally representative sample of women who participated in the 2005 National Health Interview Survey (NHIS).Methods: An area-level SEP index using 2000 U.S. census tract data was constructed and categorized into quartiles, including information on unemployment, poverty, housing values, annual family income, education, and occupation. Repeat mammography utilization (dichotomous variable) was defined as having three mammograms over the course of 6 years (24-month interval), which must have included a recent mammogram (in past 2 years). Results were obtained by ordinary multivariable logistic regression for survey data. Women ages 46 to 79 years (n ¼ 7,352) were included in the analysis.Results: In a model adjusted for sociodemographics, health care factors, and known correlates of mammography screening, women living in more disadvantaged areas had lower odds of engaging in repeat mammography than women living in the most advantaged areas [OR comparing quartile 4 (most disadvantaged) to quartile 1 (most advantaged) ¼ 0.63; 95% confidence interval, 0.50-0.80].Conclusion: The results of this nationwide study support the hypothesis that area-level SEP is independently associated with mammography utilization.Impact: These findings underscore the importance of addressing area-level social inequalities, if uptake of mammography screening guidelines is to be realized across all social strata. Cancer Epidemiol Biomarkers Prev; 20(11); 2331-44. Ó2011 AACR.

show abstract

Section: Discussionmentioning

confidence: 99%

Area-Level Socioeconomic Position and Repeat Mammography Screening Use: Results from the 2005 National Health Interview Survey

Dailey

Brumback

Livingston

et al. 2011

Cancer Epidemiology, Biomarkers &Amp; Prevention

Self Cite

View full text Add to dashboard Cite

show abstract

“…How big a problem is this? Brumback et al (2010Brumback et al ( :1651 found that, in running simulations, "it was difficult to find an example in which the problem is severe" (see also Goetgeluk and Vansteelandt 2008). In a later paper, however, Brumback et al (2013) did identify one such example, but only with properties unlikely to be found in real life data (Allison 2014)-x i and i very highly correlated, and few observations per level-2 entity.…”

Section: Generalising the Re Model: Binary And Count Dependent Variablesmentioning

confidence: 99%

Fixed and random effects models: making an informed choice

2018

View full text Add to dashboard Cite

This paper assesses the options available to researchers analysing multilevel (including longitudinal) data, with the aim of supporting good methodological decision-making. Given the confusion in the literature about the key properties of fixed and random effects (FE and RE) models, we present these models' capabilities and limitations. We also discuss the within-between RE model, sometimes misleadingly labelled a 'hybrid' model, showing that it is the most general of the three, with all the strengths of the other two. As such, and because it allows for important extensions-notably random slopes-we argue it should be used (as a starting point at least) in all multilevel analyses. We develop the argument through simulations, evaluating how these models cope with some likely mis-specifications. These simulations reveal that (1) failing to include random slopes can generate anticonservative standard errors, and (2) assuming random intercepts are Normally distributed, when they are not, introduces only modest biases. These results strengthen the case for the use of, and need for, these models.

show abstract

“…The approach taken in the current manuscript (e.g., Equation 6) has been shown to produce consistent and efficient estimates of the within-cluster effect when the fitted model is correctly specified (33,34). Although the approach taken in the current manuscript is both practical and useful (particularly for the type of examples reviewed in this manuscript) it also has been shown to be a special case of a more general approach (35).…”

mentioning

confidence: 93%

“…That two parameterizations of a MLM, which differ only by centering decision(s), are statistically equivalent does not necessarily mean that both models do an equally good job of representing specific research questions and / or theory-driven hypotheses – a point that will be made clear in the Familias Unidas example. Prior to proceeding to the Familias Unidas example it is worth noting that centering (as described above) within a MLM can also be conceptualized within a broader framework as one way to adjust the estimated within-cluster effect in the presence of confounding by cluster (33–35). 3 …”

Section: Mlms: a Brief Introductionmentioning

confidence: 99%

“… 3 The cited studies (33–35) provide a thorough account of when the estimated within-cluster effect is likely to be accurately adjusted in the presence of confounding by cluster under various models (e.g., conditional linear/logistic regression, general(ized) linear mixed models) and link functions (e.g., identity, logit, log-log). The approach taken in the current manuscript (e.g., Equation 6) has been shown to produce consistent and efficient estimates of the within-cluster effect when the fitted model is correctly specified (33,34).…”

mentioning

confidence: 99%

See 1 more Smart Citation

Multilevel Modeling in Psychosomatic Medicine Research

Myers

Brincks

Ames

et al. 2012

Psychosomatic Medicine

View full text Add to dashboard Cite

The primary purpose of this manuscript is to provide an overview of multilevel modeling for Psychosomatic Medicine readers and contributors. The manuscript begins with a general introduction to multilevel modeling. Multilevel regression modeling at two-levels is emphasized because of its prevalence in psychosomatic medicine research. Simulated datasets based on some core ideas from the Familias Unidas effectiveness study are used to illustrate key concepts including: communication of model specification, parameter interpretation, sample size and power, and missing data. Input and key output files from Mplus and SAS are provided. A cluster randomized trial with repeated measures (i.e., three-level regression model) is then briefly presented with simulated data based on some core ideas from a cognitive behavioral stress management intervention in prostate cancer.

show abstract

Adjusting for confounding by cluster using generalized linear mixed models

Cited by 53 publications

References 9 publications

Area-Level Socioeconomic Position and Repeat Mammography Screening Use: Results from the 2005 National Health Interview Survey

Area-Level Socioeconomic Position and Repeat Mammography Screening Use: Results from the 2005 National Health Interview Survey

Fixed and random effects models: making an informed choice

Multilevel Modeling in Psychosomatic Medicine Research

Contact Info

Product

Resources

About