Dear Statalisters,
I am happy to announce a new command, cemimix, available on SSC.
It conducts reference-based multiple imputation of two repeatedly measured continuous outcomes. It was designed primarily for the imputation of cost and effectiveness variables in clinical randomised controlled trials.
It follows the method described in Leurent et al. 2020, based on the work by Carpenter et al. 2013. It allows for various missing data assumptions, including 'Missing at random' (MAR), 'Jump to reference' (J2R), 'Copy increments in reference' (CIR), 'Last mean carried forward (LMCF), and 'Baseline mean carried forward' (BMCF).
Installation
Example
Thank you to Suzie Cro for helping with this program.
Suggestions for future updates and bug reporting would be appreciated.
Best wishes,
Baptiste
References:
Leurent B & Cro S (2022). "CEMIMIX: Stata module to perform reference-based multiple imputation of cost-effectiveness data in clinical trials," Statistical Software Components S459096, Boston College Department of Economics.
Leurent, B., Gomes, M., Cro, S., Wiles, N., & Carpenter, J. R. (2020). "Reference‐based multiple imputation for missing data sensitivity analyses in trial‐based cost‐effectiveness analysis." Health economics, 29(2), 171-184.
Carpenter, J. R., Roger, J. H., & Kenward, M. G. (2013). "Analysis of longitudinal trials with protocol deviation: a framework for relevant, accessible assumptions, and inference via multiple imputation." Journal of biopharmaceutical statistics, 23(6), 1352-1371.
I am happy to announce a new command, cemimix, available on SSC.
It conducts reference-based multiple imputation of two repeatedly measured continuous outcomes. It was designed primarily for the imputation of cost and effectiveness variables in clinical randomised controlled trials.
It follows the method described in Leurent et al. 2020, based on the work by Carpenter et al. 2013. It allows for various missing data assumptions, including 'Missing at random' (MAR), 'Jump to reference' (J2R), 'Copy increments in reference' (CIR), 'Last mean carried forward (LMCF), and 'Baseline mean carried forward' (BMCF).
Installation
Code:
. ssc install cemimix, all
Code:
. use 10tt.dta . cemimix, effectv(qol_0 qol_3 qol_6 qol_12 qol_18 qol_24) costv(totalcost) covariates(age sex) emethod(J2R) ref(0) treatv(arm) Performing MVN procedure for group 1 of 2... Performing MVN procedure for group 2 of 2... Performing imputation procedure... Imputed 5 datasets under J2R (effectiveness) and MAR (cost) assumptions
Suggestions for future updates and bug reporting would be appreciated.
Best wishes,
Baptiste
References:
Leurent B & Cro S (2022). "CEMIMIX: Stata module to perform reference-based multiple imputation of cost-effectiveness data in clinical trials," Statistical Software Components S459096, Boston College Department of Economics.
Leurent, B., Gomes, M., Cro, S., Wiles, N., & Carpenter, J. R. (2020). "Reference‐based multiple imputation for missing data sensitivity analyses in trial‐based cost‐effectiveness analysis." Health economics, 29(2), 171-184.
Carpenter, J. R., Roger, J. H., & Kenward, M. G. (2013). "Analysis of longitudinal trials with protocol deviation: a framework for relevant, accessible assumptions, and inference via multiple imputation." Journal of biopharmaceutical statistics, 23(6), 1352-1371.
