Applying multiple imputation to multi-item patient reported outcome measures: advantages and disadvantages of imputing at the item, sub-scale or score level
<p>Missing data are generally unavoidable in clinical trials (RCTs), particularly in patient reported outcome measures (PROMs) and can introduce bias into the study results. Multiple imputation (MI) is considered to be one of the most reliable methods to handle this problem.</p> <p>...
| Main Authors: | , , , , |
|---|---|
| Format: | Conference item |
| Published: |
BioMed Central
2016
|
| Summary: | <p>Missing data are generally unavoidable in clinical trials (RCTs), particularly in patient reported outcome measures (PROMs) and can introduce bias into the study results. Multiple imputation (MI) is considered to be one of the most reliable methods to handle this problem.</p> <p>Traditionally applied to the full PROMs score of multi-item instruments, some recent research suggests that MI at the item level may be preferable under certain scenarios.</p> <p>We present practical guidance on the choice of MI models, and offer advice on improving convergence of complex models.</p> |
|---|