Healthy Family Program on population blood pressure: a multicenter, parallel group, cluster randomized, controlled trial in rural China (Healthy Family Program)

The Healthy Family Program trial aims to evaluate the effectiveness of a community-based, health instructor led, multifaceted family intervention, as compared with usual care, on blood pressure (BP) management among Chinese rural residents, with or without hypertension. It is designed as a two-arm, parallel, cluster randomized trial of 80 villages for 12 months. This statistical analysis plan pre-specifies the method of analysis for every outcome and key variables conducted in the trial. The primary outcome is change in systolic BP from baseline to 6 months in all participants, reported as the absolute difference between intervention and control groups. The primary outcome will be modelled using a linear mixed effect model based on a participant-level analysis. The model will include random effects at the village and family level to account for the clustering of participants within village and within family, respectively. Revision SummaryThis version contains this updated information. O_LIWe modified the primary analysis to account for potential correlation of outcomes within families. A mixed effect model was extended with the same set of fixed effects, village random effect, and family random effect (nested within village). However, since the number of participants within a family is relatively small, the estimation of family random effect could be numerically unstable. When the variance of the family random effect could not be estimated, a mixed effect model with village random effect will be conducted instead. C_LIO_LIAdditional sensitivity analysis is planned at the family level, where family-level mean change in systolic blood pressure will be analyzed using a linear mixed effect model. C_LIO_LIAdditional subgroup analyses is included by stratifying participants based on family size and township socio-economic level. C_LIO_LIIn the subgroup analysis, the categorization of the Mini-EAT dietary scale was modified from the original cutoffs of <60 and [&ge;]60 points to new thresholds based on the sample medium (i.e., below the median vs. equal to or above the median). C_LI