|
Activity Number:
|
403
- SPAAC Poster Competition
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, July 30, 2019 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
|
Abstract #304341
|
|
|
Title:
|
Use of Quadratic Inference Function for Estimation of Marginal Intervention Effects in Cluster Randomized Trials
|
|
Author(s):
|
Hengshi Yu* and Fan Li and Elizabeth L Turner
|
|
Companies:
|
University of Michigan, Ann Arbor and Duke University and Duke University
|
|
Keywords:
|
Cluster randomized trial;
Generalized estimating equation;
Quadratic Inference Function;
Marginal intervention effect
|
|
Abstract:
|
Cluster randomized controlled trials (c-RCTs) are trials that randomize clusters, but measure outcomes on individuals. Marginal intervention effects are commonly of interest for their population-averaged interpretation. Such effects are typically estimated using the generalized estimating equation (GEE) approach for the correlated nature of outcomes measured on individuals from the same cluster. An alternative approach is the quadratic inference function (QIF) approach. Evidence from the literature suggests that QIF can provide an efficiency improvement over GEE in the longitudinal data setting. There is limited evidence of their potential benefits for the correlated data in c-RCTs. In this paper, we apply QIF and GEE in the estimation of the marginal intervention effects in c-RCT with continuous outcomes at one follow-up time point. We concentrate on c-RCTs with a large number of clusters and approximately the same cluster sizes. We compare QIF and GEE through simulation studies and provide three novel theorems about equivalence of point estimation from QIF and GEE. We demonstrate the two methods using data from a real c-RCT in Kenya.
|
Authors who are presenting talks have a * after their name.
