Process and Outcome Evaluation of the Walk With Ease Program for Fall Prevention

The study will evaluate the effectiveness of incorporating physical therapy exercises (primary prevention strategy) within an established physical activity intervention called Walk with Ease (WWE) to reduce incidence and risk of falls in community-dwelling older adults. A randomized controlled design will be used to systematically test the Standard Implementation (SI) model with generic exercise recommendations against an Enhanced Implementation (EI) model that provides individually-prescribed, physical therapy exercises. Standard Education (SE) resources will also be directly compared against an Enhanced Education (EE) resources based on theory-based, habit formation approaches using a 2 x 2 factorial approach.

The primary hypothesis is that the participants in the EI condition will have lower incidence of falls and larger reductions in fall risk than those in the SI condition. The secondary hypothesis is that participants receiving the EE resources will have greater compliance to the prescribed exercise, higher levels of physical activity, and better fall risk reduction outcomes than those receiving SE resources. The study will include a robust process evaluation (Aim 1) and outcome evaluation (Aim 2), along with a supplemental feasibility study (Aim 3) to identify actionable factors within the implementation context that could positively influence the scalability of WWE to reach rural populations.

The primary outcome based on the frequency of falls will be analyzed with a negative binomial regression model. The primary interest is in the β coefficient that quantifies the program main effect, averaged over behavioral treatments. If dropout is informative, we will impute the missing observations using a mixed model to account for variability between individuals. The primary hypothesis will also be evaluated using a continuous fall risk score developed by Helsel et al. based on longitudinal fall risk data. Assuming normally distributed data, our primary hypothesis will be a contrast evaluating the average change between baseline and 6 weeks in the two enhanced program treatments.

We will recruit 240 participants into the overall trial and project a final sample of 180 completers, based on a 25% dropout rate, providing a final sample size is 45 individuals in each of the four treatment groups. Based on outcomes from a related clinical trial by Li et al.70 we would have 80% power to detect a 14% reduction in fall frequency between conditions and 95% power to detect a 18% reduction in fall frequency between conditions. Using the continuous risk score proposed by Helsel et al. we would have 80% power to detect a 0.7 unit change in Helsel score or an 11% increase in the odds of a fall in one year. We would have 95% power to detect a 0.8 unit change in Helsel score or an 13% increase in the odds of a fall in one year.