If about 10% of participants followed up, wouldn’t there be a possibility of non-response/self-selection bias in the pre-post study? If I suffered cognitive deficit, I’d be more/less (I’d think less) likely to follow up.