Print  Close

What's the Weight? Controlled Outcome Differences in Complex Surveys for Racial Disparities Research

Abstract Number:

3655 

Submission Type:

Contributed Abstract 

Contributed Abstract Type:

Paper 

Participants:

Stephen Salerno (1), Emily Roberts (2), Belinda Needham (3), Tyler McCormick (4), Bhramar Mukherjee (3), Xu Shi (3)

Institutions:

(1) Fred Hutchinson Cancer Center, Seattle, WA, (2) University of Iowa, Iowa City, IA, (3) University of Michigan, Ann Arbor, MI, (4) University of Washington, Seattle, WA

Co-Author(s):

Emily Roberts  
University of Iowa
Belinda Needham  
University of Michigan
Tyler McCormick  
University of Washington
Bhramar Mukherjee  
University of Michigan
Xu Shi  
University of Michigan

First Author:

Stephen Salerno  
Fred Hutchinson Cancer Center

Presenting Author:

Stephen Salerno  
University of Michigan

Abstract Text:

A basic descriptive question is whether differences between groups exist based on levels of a discrete covariate (e.g., racial disparities in health outcomes). When this covariate is correlated with other factors related to the outcome, direct comparisons may yield misleading patterns without appropriate adjustment. Propensity scores are often employed as tools to achieve covariate balance in observational data, but their use in certain complex survey settings remains an open challenge. We focus on a specific problem – when sample selection depends on group membership. We propose novel estimands to study the average controlled difference in outcomes between groups. In simulation, our methods outperform traditional approaches, with less bias, lower mean squared error, and nominal coverage rates. Our motivation uses data from the National Health and Nutrition Examination Survey to study the interplay of self-reported race and social determinants of health on telomere length. We find that estimated differences in telomere length between Black and White individuals are attenuated compared to methods that do not appropriately account for simultaneous confounding and selection biases.

Keywords:

Confounding Bias|Average Controlled Difference|Complex Surveys|NHANES|Racial Disparities|Telomere Length

Sponsors:

Section on Statistics in Epidemiology

Tracks:

Miscellaneous

Can this be considered for alternate subtype?

No

Are you interested in volunteering to serve as a session chair?

Yes

I have read and understand that JSM participants must abide by the Participant Guidelines.

Yes

I understand that JSM participants must register and pay the appropriate registration fee by June 1, 2024. The registration fee is non-refundable.

I understand