Online Program

Return to main conference page
Saturday, October 20
Sat, Oct 20, 10:00 AM - 11:30 AM
Celebrating Our Technical Contributions

Analyzing Studies with Baseline (BL) and Post-BL Measurements (304882)

*yuanshu zou, Procter & Gamble 
Dionne Swift, Procter & Gamble 
Julie Grender, Procter & Gamble 
Greg Carr, Procter & Gamble 

Keywords: Baseline, Post Baseline, correlation, missing data

For studies with baseline and post-baseline measurements, our interest is commonly on differences in change over time between treatment groups, and baseline values are important to consider when trying to assess those change over time. In particular, what is the role of the baseline measurements, and how should they be handled in the analysis? And if we have several post-baseline visits, is it acceptable to analyze the response at each visit separately instead of performing a repeated measures analysis that uses the data at different visits together. We outlined different options from the literature and introduce factors that can impact our choice of methods. Through detailed clarification of methods and simulation studies, we provided analysis recommendations under different scenarios to help researchers analyzing data from ‘pre-post designs’ understand the impact of certain factors, such as the level of correlation (between the baseline and post-baseline measurements) and amount and type of missing data.