On the Move: Optimizing Participation in Group Exercise

This single-blinded cluster randomized trial compares the effectiveness and sustainability of On the Move and a standard group exercise program in 400 community-dwelling older adults. We will also examine the acceptability and risks associated with exercise participation. We will accomplish the objectives by pursuing the following specific aims:

1. Compare the effects of the On the Move group exercise program to a standard program on self-reported function and disability and walking ability. The On the Move program will produce greater gains in self-reported function and disability (Late Life Function and Disability Index/LLFDI) and walking ability (6-minute walk test/6MWT and gait speed) when delivered by an exercise leader.
2. When feasible to be delivered by staff activity personnel, explore the effectiveness of On the Move compared to a standard program; and sustainability compared to delivery by exercise leaders (when feasible). Explore barriers to identifying and training staff activity personnel. On the Move delivered by staff activity personnel (when feasible) will produce gains in above outcomes that are greater than the standard program; and comparable to when delivered by an exercise leader.
3. Compare the acceptability and the risks of the On the Move and standard exercise programs delivered by (a) exercise leaders and (b) staff activity personnel (when feasible). On the Move will result in greater satisfaction and higher attendance rates than the standard program. Attendance rates and satisfaction will be similar for exercise leader and staff activity personnel led programs when feasible to recruit staff personnel. Adverse event (falls, soft tissue injuries, muscle soreness, etc) rates during exercise will be similar between the two groups and the two facilitators when feasible to recruit staff personnel.
4. Explore potential baseline predictors of benefit and risks of participation in On the Move program to facilitate informed patient decision making. We will be able to identify combinations of baseline physical, psychosocial and demographic factors associated with each of the treatment response and adverse events outcomes.

This single-blinded cluster randomized trial targets community-dwelling older adults residing in Independent Living Facilities and Senior High Rises or living in the community and attending senior community centers. The trial compares type of exercise program (12 week standard group exercise program to the 12 week On the Move group exercise program) and the mode of delivery (exercise leader to staff activity personnel). Primary outcomes include function, disability and walking ability. We will also examine acceptability and risks associated with exercise participation.

Study population We will recruit participants from the Independent Living Facilities (ILFs), Senior Housing sites and from senior community centers.

Interventions:

Exercise program comparison. The On the Move exercise program will be compared to a standard group exercise program. Both exercise programs will be group-based and led by a person. Both programs will be delivered by trained exercise leaders or trained staff activity personnel (only when such a person is available). When a suitable staff activity person is not available, the program will be delivered by one of our exercise leaders. The frequency and duration of the programs are identical (50 minutes, 2 times a week for 12 weeks). The main difference between the standard and the On the Move group exercise programs is the program content which is described below.

The On the Move exercise program is based on principles of motor learning that enhance "skill" or smooth and automatic movement control. The program contains a warm-up (5 minutes), stepping patterns (15 minutes), walking patterns (15 minutes), strengthening exercises (10 minutes), and cool-down exercises (5 minutes). The warm-up and cool down contain gentle range of motion exercises and stretches for the lower extremities and trunk. The stepping and walking patterns are goal-oriented progressively more difficult patterns which promote the timing and coordination of stepping, integrated with the phases of the gait cycle. The strengthening exercises are conducted in sitting and standing and target the lower extremity muscles. The majority of the program will be conducted in standing (40 minutes) with only a small portion conducted in sitting (10 minutes).

The standard group exercise program is based on exercise programs that are currently being conducted at the facilities (i.e. standard of care). The operationally defined program contains a warm-up (5 minutes), upper and lower extremity strength and flexibility exercises (30 minutes), static balance exercises (10 minutes) and a cool-down (5 minutes). The majority of the program will be conducted in sitting (40 minutes) with only a small portion (10 minutes) conducted in standing.

Sustainability of the program: exercise leader and staff activity personnel comparison (Aim 2). The sustainability of the program will be evaluated by assessing the effectiveness of On the Move compared to a standard program when delivered by staff activity personnel and by comparing outcomes obtained by the On the Move program delivered by exercise leaders and staff activity personnel.

Outcomes:

Our main outcomes, function, disability, and walking ability are highly associated with independence and are extremely important to the older adult. Our primary measure of function and disability is the Late Life Function and Disability Instrument (LLFDI) and our main measures of walking ability are Six Minute Walk Test (6MWT) and gait speed. We will also examine confidence in walking (Gait Efficacy Scale), walking under challenging conditions (challenging gait tasks and figure of 8 walk), and gait variability as additional measures of walking ability. All measures will be collected at baseline and immediately following the 12 week intervention by research personnel who are blinded to the intervention group assignment. All testing will be conducted at the ILFs and Senior Housing sites.

Data analysis:

Aim 1:

First, we will perform a multivariate Hotelling t-test to simultaneously compare the baseline to follow-up change in the three primary outcomes between the arms to protect the type I error rate from multiplicity. If significant, subsequent analyses will be performed without further multiplicity adjustment. If not, subsequent comparisons will be performed with a conservative Bonferroni correction at the α=0.05/4=0.0125 level. This protected test approach has been recommended in the statistical literature and used in other exercise intervention trials with multiple outcomes.

Second, we will fit a series of linear mixed models using the SAS® MIXED procedure with baseline to follow-up change in each of the continuous outcomes (LLFDI function/disability, walking ability, other measures of mobility performance) as the dependent variable; intervention arm (standard/On the Move), as the fixed effect of primary interest; baseline value of outcome as an additional fixed effect covariate; and a facility random effect to account for greater similarity of participants from the same facility compared to different facilities and resulting non-independence of observations within facility (ie. clustering). We will construct appropriate means contrasts to estimate difference in gains in the two interventions when delivered by exercise leaders (Aim 1) whose statistical significance of the estimates will serve as formal tests of hypotheses. We will consider adding other baseline measures found to be significantly different between the arms or deemed important as additional fixed effect covariates in the model to assess robustness of the findings. We note that waiting of participants randomized to staff activity personnel classes constitute a group of control participants, if smaller, without any intervention. Therefore, we will perform another sensitivity analysis employing a similar analytic strategy but with three intervention arms (On the Move/Standard/Wait List Control).

Aim 3(a):

We will fit a series of generalized estimating equations (GEE) models(68) using the SAS® GENMOD procedure with each of the dichotomous adverse events, adherence (21+ sessions or ≥90%) and satisfaction outcomes as the dependent variable; a binomial distribution for the outcome and a logit canonical link function; intervention arm, as the effect of primary interest; baseline value of outcome and any other measures found to be different between arms or deemed important as additional fixed effects covariates; and an exchangeable working correlation structure to account for clustering due to facility. We will appropriately construct contrasts to test hypotheses of differential proportions with adverse events based on intervention when delivered by exercise leaders.

Aims 2, 3(b) and 4 Exploratory Analyses

Aim 2:

We propose an analytic strategy with an exploratory philosophy for Aim 2 sustainability aim, because of the uncertainty surrounding our ability to recruit and train a staff activity person to lead a class safely and per protocol.

We will fit a series of linear mixed models using the SAS® MIXED procedure with baseline to follow-up change in each of the continuous outcomes (LLFDI function/disability, walking ability, other measures of mobility performance) as the dependent variable; intervention arm (standard/On the Move), delivery mode (by exercise leader/staff activity personnel) and their interaction as fixed effects of interest; baseline value of outcome and any other measures found to be different between arms or deemed important as additional fixed effects covariates; and a facility random effect to account for greater similarity of participants from the same facility compared to different facilities and resulting non-independence of observations within facility (ie. clustering). We will construct appropriate means contrasts to estimate difference in gains in the two interventions when delivered by staff exercise personnel (Aim 2 effectiveness hypothesis); and difference in gains attributable to On the Move intervention when delivered by exercise leaders and staff activity personnel (Aim 2 sustainability hypothesis). Statistical significance of the estimates will serve as formal tests hypotheses. We note that participants randomized to staff activity personnel had to wait 12 weeks since their baseline assessment and randomization to start the intervention, during which they may have changed; and that they underwent a second baseline assessment immediately prior to starting the exercise intervention. Therefore, we will repeat the above analysis using their second baseline assessment instead of the first one to assess the sensitivity of results.

Aim 3(b):

We will fit a series of generalized estimating equations (GEE) models using the SAS® GENMOD procedure with each of the dichotomous adverse events, adherence (21+ sessions or ≥90%) and satisfaction outcomes as the dependent variable; a binomial distribution for the outcome and a logit canonical link function; intervention arm, delivery model and their interaction as effects of interest; baseline value of outcome and any other measures found to be different between arms or deemed important as additional fixed effects covariates; and an exchangeable working correlation structure to account for clustering due to facility. We will appropriately construct contrasts to test hypotheses of differential proportions with adverse events based on intervention when delivered by staff activity personnel and delivery mode.

Aim 4:

We will perform the exploratory analyses to identify combinations of baseline predictors of treatment response and risks of participating in On the Move program.