An exploratory study to understand how rurality status and demographic characteristics are associated with enrollment, engagement, and retention in a digital health intervention targeting the Appalachian region

Purpose

Digital health studies exploring group disparities across research phases are limited. As a secondary aim of a larger digital health trial, this study explored how rurality and other sociodemographics were associated with enrollment, retention, and engagement in a randomized controlled sugar-sweetened beverage (SSB) reduction trial.

Methods

Participants from a primarily Appalachian sample were randomized into iSIPsmarter (experimental) or static Patient Education (control) websites. Enrollment, retention (6 months), and iSIPsmarter engagement (completion of metered program Core content and SSB and weight diaries) were collected from July 2021 to August 2023. Regression models assessed subgroup associations using Rural Urban Continuum Codes (RUCC), sex, race, age, income, education, and other sociodemographic predictors.

Findings

Of the 509 eligible participants, 249 (49%) enrolled, and 218 (88%) were retained. Participants were predominantly White (89%), college-educated (59%) females (83%) with household incomes <$55,000/year (52%). Rurality varied: RUCC 1-2 (medium-large metro) = 15%, RUCC 3 (small metro) = 45%, and RUCC 4-9 (nonmetro) = 41%. On average, iSIPsmarter participants (n = 127) completed 4.89/6 (SD = 1.69) Cores and 76% (SD = 29%) and 57% (SD = 31%) of SSB and weight diaries. Rurality was a nonsignificant predictor, but higher education and health literacy increased enrollment likelihood by 37% (95% CI = 1.12-1.67) and 23% (95% CI = 1.03-1.47), respectively. Greater education (OR = 1.51, 95% CI = 1.00-2.29), age (OR = 1.04, 95% CI = 1.01-1.07), and income (OR = 1.13, 95% CI = 1.00-1.28) significantly predicted retention. Older age significantly (P<.05) predicted the completion of Cores and diaries.

Conclusions

Results suggested rurality was not significantly associated with enrollment, retention, or engagement, though this conclusion warrants caution. Future digital health studies targeting similar populations should consider additional sociodemographic differences.