A Multiple Imputation approach for combining data across multiple trials that use different outcome measures.
Ahnalee Brinks, University of Miami
Keywords: individual participant data, harmonization, meta-analysis, calibration
We discuss a missing data approach for combining data from multiple clinical trials when not all the studies use the same outcome measure. We treat those measures that were not administered as missing data and use multiple imputation to impute the missing outcomes. We apply our method to a data synthesis of five adolescent depression clinical trials where 4 studies used the Children’s Depression Rating Scale and one study used the Hamilton Depression Rating Scale and no studies used both scales.
Important Dates & Deadlines
- October 9 - 11, 2013