Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 461 - Design and Analytic Approaches to Address Unmeasured Confounding
Type: Contributed
Date/Time: Thursday, August 6, 2020 : 10:00 AM to 2:00 PM
Sponsor: Section on Statistics in Epidemiology
Abstract #313164
Title: Is Anemia Prevalence a Good Proxy for Malaria Prevalence for Children? a Community-Level Perspective via Matched Logistic Regression
Author(s): Emily Diana* and Shuxiao Chen and Sheng Gao and Siyu Heng and Hongming Pu and Dylan Small and Hua Wang
Companies: University of Pennsylvania and University of Pennsylvania and University of Pennsylvania and University of Pennsylvania and University of Pennsylvania and University of Pennsylvania and University of Pennsylvania, Statistics Department of Wharton
Keywords: anemia; malaria; matching; logistic regression; causal inference
Abstract:

Anemia prevalence has been recommended as an easy to collect proxy for malaria prevalence in African countries. However, close inspection of supporting research reveals that the validity of this proxy has been established only from an individual-level point of view, overlooking the complicated interactions between individuals and their environments. We consider the relationship between anemia and malaria at the community-level, taking into consideration confounders. In particular, we study the causal effect of malaria prevalence on anemia prevalence at the community level and propose a novel two-step matching procedure followed by a difference-in-difference-type mixed-effect logistic regression to unbiasedly estimate the causal effect. We show that after controlling for bias introduced by confounders, the change of children's malaria prevalence may not be a statistically significant causal factor for the change of children's anemia prevalence. Our method can be applied generally to study the causal relationship between treatment and response at the community-level using purely observational data and enables one to rigorously detect unmeasured confounders.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2020 program