SystemCHANGE-HIV: A Pilot Study Exploring the Effect of a Systems-based Intervention to Increase Physical Activity, Sleep Behavior, and Mental Wellness

Sample: The study sample will include 40 HIV+ men and women. Inclusion criteria include: a documented HIV diagnosis; adult (18 years old or greater); and English speaking.

Design: The SystemCHANGE-HIV study will be a 15-week, randomized, two-group experimental design including a 10-week behavioral intervention. In the experimental group, one half of the participants (n=20) will receive the intervention; and in the control group, the other half of the participants will undergo usual care. Both the experimental and control group will complete the same psychometric instrument packet, three times, to give a comparison of self-management before and after intervention and at a four-week follow up visit. This packet includes information on medical history, demographics, HIV self-management,physical activity, social support, social capital, mental wellness, and quality of life.Additionally, all participants will wear an actigraph for seven consecutive days, at each data collection point, to assess sleep.

Analysis:Preliminary data analysis will include descriptive statistics and scatter plots to examine the distribution of each outcome at each time point, as well as each covariate of interest. Although our randomization scheme will ensure balance of baseline variables between the two arms (other than that due to chance alone), we will estimate the standardized magnitude of each baseline difference as well as the correlation between each baseline variable and each outcome of interest at 10 weeks or at 15 weeks to objectively determine baseline variables that will be needed to estimate the unbiased effect of the intervention on each outcome. The primary analysis will be based on intent to treat approach and will investigate the effect of the intervention on continuous outcomes at 10 weeks or 15 weeks. In particular, an Analysis of Covariance model will be used that adjusts for treatment arm and the baseline level of the outcome variable. Such a model will allow us to estimate the effect of the intervention on each outcome, adjusting for the baseline outcome.In addition, linear mixed models will be used to determine if the effect of the intervention on the outcome at 10 weeks differs from the effect of the intervention on the outcome at 15 weeks. If needed, all analyses will be repeated using an as-treated approach. All analyses will be performed using a two-sided significance of 0.05.