Sample size calculation

With a small to moderate within cluster correlation (ICC = 0.10) and moderate individual autocorrelation (0.50), this analysis will have adequate power (power = 0.945) to detect an OR of 1.5 in the proportion of participants achieving viral suppression with at least 90 participants per site. Our effect size estimate is conservative given that the iCARE UG3 pilot trial findings (OR = 14.0) was not randomized and not controlled. Findings from our prior randomized trial of text messaging on medication adherence (a related outcome) among YLH in the United States suggested an effect size of OR = 2.12 (95% CI 1.01–4.45).

Primary study data will be analyzed as soon as possible after the end of data collection, with study findings disseminated in peer-review public health journals.