Global obesity has doubled since 1990, and approximately one-quarter of UK adults are affected. Obesity is now recognized as a complex condition that requires evidence-based and personalized treatment. Many weight loss programs are delivered globally using generic advice, which contributes to highly variable clinical outcomes. Digital interventions are becoming more effective and allow easier measurement of user engagement, a factor associated with weight loss. Personalizing/tailoring advice may further enhance user engagement, but cognitive-behavioral personalization remains underexplored. This study tested whether phenotype-matched guidance within Oviva’s national health services (NHS) program improves app engagement and weight loss. It compared these outcomes with non-responder and historical groups. It also examined whether socioeconomic status influenced these effects.
In this quasi-experimental study, participants were recruited from Oviva, an NHS-funded 12-week digital weight management program. Eligible adults (18-80 years) with hypertension, diabetes, or obesity (body mass index [BMI] >25 kg/m2) received an email invitation if they enrolled during a 3-week period in May 2024. This program provides support from health coaches and dieticians along with access to the Oviva app for behavior tracking, goal setting, coach messaging, and lessons. Any app activity was counted as an engagement. Participants were excluded if they were frail, had unmanaged comorbid conditions, had recently undergone bariatric surgery, or were pregnant.
This study compared three groups such as phenotype group (n = 148, mean age = 48±12 years, female = 86%, initial mean BMI = 39±6 kg/m2), participants who completed a 17-item quiz and received 7 weeks of tailored advice; a historical cohort (n = 241, mean age = 44±10 years, female = 71%, initial mean BMI = 39±8 kg/m2) from the previous year; and nonresponders group (n = 394, mean age = 44±11 years, female = 76%, initial mean BMI = 39±8 kg/m2), participants who did not submit a completed quiz. Outcomes were weekly app engagement and 7-week weight change.
In the phenotype group, the fidelity rate was found to be 73.6% (109/148), remaining 6/148 (4.1%) patients followed another profile, 16/148 (10.8%) opened none, and 17/148 (11.5%) showed mixed results. The phenotype group had significantly higher total mean engagement of 257±232 compared to mean engagement of 135±198 nonresponders (p <0.001) and 159±187 historical group (p <0.001), corresponding to 62–90% greater app activity. Socioeconomic status did not significantly affect engagement. The average 7-week weight loss did not differ significantly between groups: phenotype group versus nonresponders (-2.23±7.97kg vs −0.69±13.23; p = 0.23) and phenotype group versus historical cohort (-2.23±7.97kg vs −1.60±5.39; p = 0.29).
Moreover, opening a greater number of correct documents correlated with higher total engagement with r146= 0.481and P <0.001 but not weight loss with r119 = −0.074 and P = 0.42. Phenotype participants engaged more compared to the nonresponders group with t 232.25 = 5.68 and p <0.001, but showed similar weight loss (t445 = 1.21) with a non-statistically significant p = 0.23. Over 7 weeks, BMI dropped in 0.79±2.90 kg/m2 versus 0.56±1.94 kg/m2, and body mass fell 1.93±7.24% vs 1.27±4.14% in phenotype vs historical groups, whereas nonresponders showed minimal change (0.11±5.01 kg/m2, 0.09±13.01%). Repeated measures ANOVA demonstrated significant week-by-week changes in app engagement with (F4.32,3367.68 = 81.85; p <0.001) and a significant interaction with condition (F8.64,3367.68=5.57; p <0.001) using Greenhouse-Geisser correction.
This study had several limitations, including a quasi-experimental design, email-delivered content, uneven phenotype distribution, self-reported weight, and self-selection of motivated participants, which may introduce bias.
In conclusion, this study highlights that phenotype-tailored advice may enhance app engagement and adherence in weight management programs. Larger, longer-term randomized trials are needed to confirm the impact of phenotype-tailored guidance on engagement and weight loss outcomes.
Reference: Szypula J, Jarvstad A, Jones LA, Tapper K. Personalizing a Weight Loss Program Using Cognitive-Behavioral Phenotypes to Improve Engagement and Weight Loss in Adults with Overweight or Obesity: Quasi-Experimental Study. JMIR Form Res. 2025;9:e72645. doi:10.2196/72645
