I have been asked to contribute to a statistical analysis plan where the primary endpoint is a baseline adjusted comparison of means across the two treatment arms at 6 months. The study continues on, however, collecting data for a further 18 months with secondary endpoints being 2 further comparisons of means on the same outcome. The investigators want to assess the primary endpoint only, at the 6 month mark as well as a complete analysis on these and other endpoints at end of study.
The treatment effect at 6 months could be estimated using ANCOVA (as a standard linear model) or by using a mixed model approach with a random intercept for subject, adjusting for the baseline and incorporating a treatment x (categorical) time interaction for the complete data at end of study. I hadn’t previously given thought to this but comparing the two approaches on some dummy data, the treatment effect at 6 months can differ slightly between them.
My question is, within the context of an RCT analysis plan which is the more cogent approach. My thought is to specify an ANCOVA for the primary endpoint and then use the end of study mixed model analysis only for the secondary endpoints. Does this make sense?
but at 6 months you have only one time point (with follow-up after this)? and thus no repeated measures? in this case, 6 months is the primary analysis as they will want to report results at this time (im assuming)
if you have multiple timepoints then it’s probably considered ‘old school’ to compare at the single timepoint, and you have an issue with missing data if you use ancova that would need to be considered in the analysis plan.
ICH E9 doesnt land on one approach or another:
The assessment of functional status over time in studying treatment for chronic disease presents other challenges in selection of the primary variable. There are many possible approaches, such as comparisons of the assessments done at the beginning and end of the interval of observation, comparisons of slopes calculated from all assessments throughout the interval, comparisons of the proportions of subjects exceeding or declining beyond a specified threshold, or comparisons based on methods for repeated measures data. To avoid multiplicity concerns arising from post hoc definitions, it is critical to specify in the protocol the precise definition of the primary variable as it will be used in the statistical analysis. In addition, the clinical relevance of the specific primary variable selected and the validity of the associated measurement procedures will generally need to be addressed and justified in the protocol.