Missing data is a common occurrence when collecting patient responses via patient symptom diaries. Many methodologies exist for imputing missing data and are used to estimate statistical outcomes that would have been achieved had no missing data occurred. Among the many imputation methodologies are varying multiple imputation techniques which are becoming more prevalent in clinical research. We performed a simulation study to compare statistical outcomes based on three MI methods to handle missing data versus statistical outcomes based on single imputation methods as previously reported by Hsu et al. at the 2013 ARVO conference.
This poster details:
- Analysis of three multiple imputation methods for handling missing data
- Comparison with previously reported data on single imputation methods
- Results and recommendations on how to handle missing data in clinical trials
- Lot Slade, MS
- Dale Usner, PhD
- Kirk Bateman, MS