Physical activity is a widely recognized modifiable health behavior with potential to mitigate cognitive decline in older adults. LEAP! Rx (Lifestyle Empowerment for Alzheimer’s Prevention) is a 52-week healthy lifestyle program that successfully increased cardiorespiratory fitness in a randomized controlled trial in older adults. The present study aimed to evaluate whether the intervention led to improved cognition in several domains compared to control group participants 12 and 52 weeks after the start of the intervention.

We randomized 219 older adults (Mean age (SD) = 72.00 (4.83)) to the LEAP! Rx program (n=110) or a control group (n=109). We also compared two different recruitment methods, self-referral (typical method) vs. physician referral (novel method). We evaluated cognitive function using the National Institutes of Health (NIH) Toolbox Cognitive Battery. We summarized the findings in the cognitive domains of processing speed and verbal memory. We conducted a repeated measures analysis of covariance (ANCOVA) across three timepoints: baseline, week 12, and week 52 to compare cognitive changes between the intervention and control groups while accounting for age, gender, and education.

For processing speed, the ANCOVA revealed a significant three-way interaction among treatment group (intervention vs. control), referral method (self vs. physician referral) and measurement time (baseline, week 12, week 52), F(4, 195) =  3.57, p<0.05. Bonferroni pairwise post-hoc comparisons showed no significant differences in processing speed at different timepoints, likely due to limited statistical power. Visual inspection suggests greater cognitive benefits of the intervention in the physician referred group compared to the self-referred group. For processing speed and verbal learning, older age significantly predicted poorer performance F(1, 218) = 11.86, p<0.05; F(1, 218) = 13.52, p<0.05. Males performed better at baseline and females performed better by week 52 on average for processing speed, F(5, 213) = 5.12, p<0.05. For verbal learning, women improved more than men F(1, 217) = 25.50, p<0.05.

Findings suggest that effectiveness of the intervention may depend on how participants were recruited to the study. Our physician-referred participant group was significantly more racially and ethnically diverse, and their participation was significantly more likely to be interrupted by the pandemic. While limited statistical power affects clear timepoint differences in processing speed, non-significant trending patterns suggest intervention benefits, especially for physician referred participants. These results have important implications for inclusive recruitment strategies in prevention trials and the potential of tailored exercise interventions to improve or maintain cognitive function in older adults.