Study procedure
Staff training
All staff will be adequately trained by the research team to ensure subject safety, blinding, data quality, and study- protocol adherence. Initial staff training spans three months using a variety of methods including reading, lectures, demonstration, discussions, role plays, practice rounds, and test runs with real subjects. Subsequent train- ing of research assistants (RAs) and volunteers are con- ducted jointly by staff and the investigators and tailored to the skills needed for a given role. Ongoing training for staff, RAs, and volunteers is provided by the research team through weekly staff meeting, monthly research team meeting, briefing after each treatment fidelity check, and participation in new staff/RA/volunteer training.
Randomization, allocation concealment, and blinding Randomization will allocate subjects 2:1 (cycling: stretch- ing) within each age stratum (66 to 75, 76 to 85, and 85+), and will use randomly permutated blocks of 3 and 6 subjects (Figure 1). We do not expect equal numbers of subjects in each age stratum, but want to ensure that the two randomized groups are balanced across each age stratum. Because the effect of the intervention may vary across age strata similar to varying AD prevalence across age groups, this design will reduce the variability in the comparisons within and between randomized groups.
Allocation is concealed from all investigators except for the statistician. Study staff will not be blinded. Enrolled subjects will be randomized after completing baseline data collection.