Supporting maintenance of sugar-sweetened beverage reduction using automated versus live telephone support: Findings from a randomized control trial

Abstract

Background: Although reducing sugar-sweetened beverage (SSB) intake is an important behavioral strategy to improve health, no known SSB-focused behavioral trial has examined maintenance of SSB behaviors after an initial reduction. Guided by the RE-AIM framework, this study examines 6-18 month and 0-18 month individual-level maintenance outcomes from an SSB reduction trial conducted in a medically-underserved, rural Appalachia region of Virginia. Reach and implementation indicators are also reported. Methods: Following completion of a 6-month, multi-component, behavioral RCT to reduce SSB intake (SIPsmartER condition vs. comparison condition), participants were further randomized to one of three 12-month maintenance conditions. Each condition included monthly telephone calls, but varied in mode and content: 1) interactive voice response (IVR) behavior support, 2) human-delivered behavior support, or 3) IVR control condition. Assessments included the Beverage Intake Questionnaire (BEVQ-15), weight, BMI, and quality of life. Call completion rates and costs were tracked. Analysis included descriptive statistics and multilevel mixed-effects linear regression models using intent-to-treat procedures. Results: Of 301 subjects enrolled in the 6-month RCT, 242 (80%) were randomized into the maintenance phase and 235 (78%) included in the analyses. SIPsmartER participants maintained significant 0-18 month decreases in SSB. For SSB, weight, BMI and quality of life, there were no significant 6-18 month changes among SIPsmartER participants, indicating post-program maintenance. The IVR-behavior participants reported greater reductions in SSB kcals/day during the 6-18 month maintenance phase, compared to the IVR control participants (- 98 SSB kcals/day, 95% CI = - 196, - 0.55, p < 0.05); yet the human-delivered behavior condition was not significantly different from either the IVR-behavior condition (27 SSB kcals/day, 95% CI = - 69, 125) or IVR control condition (- 70 SSB kcals/day, 95% CI = - 209, 64). Call completion rates were similar across maintenance conditions (4.2-4.6 out of 11 calls); however, loss to follow-up was greatest in the IVR control condition. Approximated costs of IVR and human-delivered calls were remarkably similar (i.e., $3.15/participant/month or $38/participant total for the 12-month maintenance phase), yet implications for scalability and sustainability differ. Conclusion: Overall, SIPsmartER participants maintained improvements in SSB behaviors. Using IVR to support SSB behaviors is effective and may offer advantages as a scalable maintenance strategy for real-world systems in rural regions to address excessive SSB consumption. Trial registry: Clinicaltrials.gov; NCT02193009; Registered 11 July 2014. Retrospectively registered.