Can Multiple Studies Combine Their Results When Their Confounder Adjustment Sets Are Different?

Ghassan Hamra, PhD

In research, directed acyclic graphs typically guide decisions on choosing confounders that need to be adjusted. These graphs also suggest how to get unconfounded, or unbiased, effect estimates, which are based on different statistical models and do not always provide the same actual numeric values. Therefore, Ghassan Hamra, PhD, of Johns Hopkins University Bloomberg School of Public Health, and his team from Wake Forest University and University of California Davis joined together to better understand when these values do and do not correspond.

Their study, “Combining Effect Estimates Across Cohorts and Sufficient Adjustment Sets for Collaborative Research: A Simulation Study” was recently published in the February issue of Epidemiology. To conduct this research, all study data were simulated, and Hamra built all of the models and analyzed all of the information. Ultimately, the team found the estimators showed similar numeric values of the estimates the majority of the time. The exception is when using a model known as logistic regression, which did not give similar estimates. Logistic regression provides an odds ratio, which is a non-collapsible quantity, or one that cannot be reliably combined.

“This study has a significant impact on the ECHO Program, as it allows researchers to understand and combine information across ECHO groups,” said Hamra. In the future, researchers can make additional models to explore other forms of bias that make combining information across studies difficult.

Read the research summary here.

Jessie Buckley: Estimating Effects of Exposure Mixtures on Child Health: Novel Methods for Solution-oriented ECHO Research

return to discovery home

Estimating Effects of Exposure Mixtures on Child Health: Novel Methods for Solution-oriented ECHO Research

Speaker:

Jessie Buckley, PhD, MPH

Johns Hopkins Bloomberg School of Public Health

ECHO Data Analysis Center (DAC)

 

 

Speaker Bio:  Jessie Buckley is an environmental and pediatric/perinatal epidemiologist, working on research to inform environmental policies targeted at improving children’s health. Her work looks at developmental origins of health and disease framework and focuses on determining effects of early life exposure to endocrine disrupting chemicals on child physical growth and development. Using molecular epidemiology and advanced statistical approaches, she has conducted several studies evaluating the role of environmental chemical exposures in the development of childhood obesity. She has also researched the utility of biomarkers of exposure to several classes of environmental chemicals that have widespread human exposure, including phthalates.

Topic:  Interest in understanding the combined effects of multiple exposures (i.e., mixtures) on children’s health is rapidly increasing, with a related proliferation of methods for estimating these effects. In this talk, Jessie will demonstrate two useful approaches for estimating mixtures effects – Bayesian kernel machine regression and quantile g-computation – with an application to endocrine disrupting chemical mixtures and childhood bone health. In addition, Jessie will introduce a novel statistical framework to advance solution-oriented mixtures research in ECHO by more directly informing practices, programs, and policies to improve children’s health.

Date: Wednesday, June 10, 1 to 2pm