The Clinical Referral to Activity Study

The main aim of this effort is to test a physical activity intervention, for adult clinical care patients at risk of developing cardiovascular disease. The intervention includes social-cognitive theory-based sessions, remote coaching, a body worn physical activity tracker (PAT), and is delivered online over one year. The investigators hypothesis that this intervention will be more successful at increasing physical activity (defined as objectively measured step counts and % of individuals meeting the moderate-vigorous physical activity goal) as an active control group who receives a body worn PAT and information on the CDC activity recommendations. The proposed intervention will be aligned with efforts by the US Centers for Disease Control to increase population physical activity levels and the American College of Sports Medicine (ACSM) to increase physical activity prescription in primary care.

The main aim of this effort is to test a physical activity intervention, for adult clinical care patients at risk of developing cardiovascular disease. The intervention includes 13 weekly and 6 monthly social-cognitive theory-based sessions, remote coaching, a body worn physical activity tracker (PAT), and is delivered online over one year. A total of n=54 Clinical care patients aged 40-70 years of age with low physical activity (PA) levels (\<150 minutes/week) at risk for cardiovascular disease will be recruited. Study outcomes will be measured at baseline, 6 months after intervention start and 12 months after intervention start. For the primary outcomes, the investigators hypothesize that participants randomized to ActiveGOALSv2 + PAT will have significantly larger (a) increases in step counts/day and % meeting moderate-vigorous physical activity goal of 150 minutes/week (to 12 months) compared to the active PAT control group. Descriptive analyses and graphic displays will be used to identify outliers, missing data, and pattern of attrition. To ensure rigor, the primary analytic strategy will be a linear or generalized mixed-effect models approach in which treatment group, time, and time by group interaction are treated as fixed effects, and subject is treated as a random effect to account for individual subject variability (two sided hypothesis test; .05 level). Mixed models are applicable to longitudinal datasets that contain missing observations, (assuming data is missing at random). Regression modeling will be conducted to adjust for important covariates. The investigators will perform similar analyses to examine changes in secondary outcomes. Descriptive statistics on patient experience, beliefs/attitudes related to maintenance of physical activity changes, and program cost will also be reported. Power and sample size: For a 2 sided test with alpha level set at .05, and assuming 20% attrition at 12 months, the investigators plan to recruit 54 participants to identify clinically relevant mean difference between randomized groups at 12-months of 2000 steps/day (with a power of .80), and 35% meeting MVPA goal (with a power of .80), given reference mean (sd) values were calculated from baseline waist worn accelerometer data from the current K12 of: 5250 (2200) steps/ day and the investigators assume 5% meeting goal at baseline in the control group. Descriptive statistics will be used to describe the total eligible population, those referred, and those referred who participated in the study; in order to determine program reach. Exploratory Analysis of Important Pre-Existing Factors: Factor analysis will be used to identify importance of factors related to physical activity levels/ activity goal achievement over the 12 month follow-up (Mplus) from participant EHR data collected prior to intervention start. Factor structures of how individual factors contribute to activity outcomes will be explored at each time point. The best factor structure will be determined with Eigenvalues and fit indices (RMSEA, CFI, and TLI). Factor structures at the 2 time points will be compared at three levels to establish measurement invariance: 1) Configural equivalence, number of factors/ pattern of factor-indicator relationships are identical over time; 2) Metric equivalence, factor loadings are equal over time; and 3) Scalar equivalence, means and values are equivalent over time. Finally, the investigators will use descriptive statistics to report presence/absence of data on physical activity levels, cardiometabolic, and patient-centered outcomes in the EHR. The investigators will apply the same regression approaches utilized in Aim 1a to determine differences between randomized groups for EHR reported outcomes.