Mobile Ecological Momentary Assessment and Intervention and Health Behavior Change among Adults in Rakai, Uganda: a Pilot Randomized Controlled Trial. (Preprint)
BACKGROUND An extraordinary increase in mobile phone ownership has revolutionized opportunities to employ mHealth approaches in lower- and middle-income countries (LMICs). Ecological Momentary Assessment (EMA) and Intervention (EMAI) uses mobile technology to gather data and deliver timely, personalized behavior change interventions in an individual’s natural setting. To our knowledge, there have been no previous trials of EMAI in sub-Saharan Africa. OBJECTIVE To advance the evidence base for mHealth interventions in LMICs, we conducted a pilot randomized trial to assess the feasibility of EMAI and to establish estimates of the potential effect of EMAI on a range of health-related behaviors in Rakai, Uganda. METHODS This prospective, parallel group, randomized pilot trial compared health behaviors between adult participants submitting EMA data and receiving behaviorally-responsive interventional health messaging (EMAI) to those submitting EMA data, alone. Using a fully-automated mobile phone application, participants submitted daily reports on five different health behaviors during a 30-day period prior to randomization (P1). Participants were then block randomized to control arm, continuing EMA reporting through exit, or intervention arm, EMA reporting plus behavioral health messaging receipt. Participants exited after 90 days of follow-up, divided into study Periods 2 (P2: randomization+29 days) and 3 (P3: 30 days post-randomization to exit). We used descriptive statistics to assess EMAI feasibility through completeness of data, and differences in reported behaviors between time periods and study arms. RESULTS The study included 24 participants per arm (48% female, median age 31 years). EMA data collection was feasible, with 85.5% of the combined 4,418 days reporting some behavioral data. There was a decrease in the mean proportion of days when alcohol was consumed in both arms over time (control: P1, 9.6% of days to P2, 4.3% of days, P=.03; intervention P1, 7.2% of days to P3, 2.4% of days, P=.01). Decreases in sex with a non-long-term partner without a condom were also reported in both arms (P1 to P3 control: 1.9% of days to 1.0% of days, P=.02, intervention: 6.6% of days to 1.3% of days P=.03). An increase in fruit and vegetable consumption was found in the intervention (P1 to P3 fruit: 78.6% of days to 87.0% of days, P=<.001; vegetable: 65.6% of days to 76.6% of days, P =.03 ) but not the control arm. Between arms, there was a significant difference in the change in reported vegetable consumption between P1 and P3 (control: 7.99% decrease in the mean proportion of days vegetables consumed, intervention: 11.05% increase, P=.01). CONCLUSIONS Preliminary estimates suggest that EMAI may be a promising strategy to promote behavior change across a range of behaviors. Larger trials examining the effectiveness of EMAI in LMICs are warranted. CLINICALTRIAL ClinicalTrials.gov NCT04375423; https://www.clinicaltrials.gov/ct2/show/NCT04375423.