A randomized control trial (RCT) has nothing to do with a cross sectional study. A RCT implies randomization and the assessment criteria in groups can be accomplished at different times if the protocol study provides that.

Thank you so much for your valuable inputs. The biggest challenge that I have is primary outcome variable. The primary outcome variable is min. acceptable diet among age group of 6-23.9 months during baseline survey. In general, when we do baseline survey for a group of households, we need to follow with them during endline/followup survey in order to understand the program effect using difference in difference model. But the problem starts here. when we do followup survey, those kids age might be out of age group for measuring min. acceptable diet. I have gone through few papers, they used repeated cross-sectional survey but as Bjoern pointed out, it does not mater provided if we randomly assigned the intervention. The second challenge that I have, when I calculated power to detect my program effect at certain level, i realized that I must increase number of clusters as we randomized at cluster level. But, it will be difficult for us to increase the cluster as we constraint to the program intervention areas and hence we need to adjust other power ingredients in order to detect sufficient power to measure the impact of the program intervention. Finally, we decided to change the primary outcome variable from min. acceptable diet to min. child dietary diversity age group between 2-24 months, where we will be following same child during endline survey. At that time, they might be older. However, they will be less than 5 years, which very important for us as FAO definition says that child dietary diversity can be measured less than 5 years. Once again thank you so much for your help and suggestions.