Dear Dr. Shi,
I hope you are doing well. My colleagues and I are planning a longitudinal mediation analysis and are currently exploring whether CMAverse would be appropriate for our study.
Briefly, we have five waves of data and aim to estimate interventional (stochastic) direct and indirect effects of rainfall variability (SPEI-6) on child height-for-age Z-scores (HAZ), with household food insecurity (HFIAS) as the mediator. The mediator could be discretized if needed. Our setting includes time-varying confounding, including mediator–outcome confounders that are themselves affected by prior exposure, as well as some missing data.
While reviewing the CMAverse website and documentation, I became a little confused. The package website indicates that CMAverse implements the estimation approach of VanderWeele and Tchetgen Tchetgen's work on mediation analysis with time-varying exposures and mediators. However, after reading the documentation and vignettes, all examples and the documented workflow appear to focus on single-time-point mediation analyses. Based on this, my impression is that the current version of CMAverse does not yet implement the general longitudinal mediational g-formula for repeated exposures and mediators described in that paper, and therefore may not be suitable for our planned analysis.
Could you kindly let me know whether this interpretation is correct? If longitudinal mediation of the type described above is supported, I would be very grateful if you could point me to the relevant documentation or examples. If not, it would also be helpful to know whether such functionality is planned for a future release or whether you would recommend another software implementation for this setting.
Thank you very much for your time and for developing CMAverse. I look forward to your advice.
Best,
Epi17
Dear Dr. Shi,
I hope you are doing well. My colleagues and I are planning a longitudinal mediation analysis and are currently exploring whether CMAverse would be appropriate for our study.
Briefly, we have five waves of data and aim to estimate interventional (stochastic) direct and indirect effects of rainfall variability (SPEI-6) on child height-for-age Z-scores (HAZ), with household food insecurity (HFIAS) as the mediator. The mediator could be discretized if needed. Our setting includes time-varying confounding, including mediator–outcome confounders that are themselves affected by prior exposure, as well as some missing data.
While reviewing the CMAverse website and documentation, I became a little confused. The package website indicates that CMAverse implements the estimation approach of VanderWeele and Tchetgen Tchetgen's work on mediation analysis with time-varying exposures and mediators. However, after reading the documentation and vignettes, all examples and the documented workflow appear to focus on single-time-point mediation analyses. Based on this, my impression is that the current version of CMAverse does not yet implement the general longitudinal mediational g-formula for repeated exposures and mediators described in that paper, and therefore may not be suitable for our planned analysis.
Could you kindly let me know whether this interpretation is correct? If longitudinal mediation of the type described above is supported, I would be very grateful if you could point me to the relevant documentation or examples. If not, it would also be helpful to know whether such functionality is planned for a future release or whether you would recommend another software implementation for this setting.
Thank you very much for your time and for developing CMAverse. I look forward to your advice.
Best,
Epi17