An Early-Stage Digital Therapeutic Intervention to Enhance Affective Response During Physical Activity Among Adults With Overweight or Obesity: Benchmark-Driven Formative Testing Study.

No adverse events were reported across any of the four DTx versions during the 14-day study period.

• The a priori safety benchmark was defined as fewer than 1% of participants reporting an adverse event.
• N=36 participants were enrolled and assessed for safety via official Institutional Review Board reporting channels.
• All four DTx versions met the safety threshold with zero adverse events reported.

More than half of participants who received affect goals reported increased PA enjoyment at the end of the study, meeting the plausibility benchmark.

• The a priori plausibility benchmark required that at least 51% of participants experience increased enjoyment in physical activity.
• Participants in affect goal conditions (versions 2, 3, and 4) were assessed for PA enjoyment.
• Plausibility was assessed via interviews.
• The plausibility benchmark was met, with more than half of affect-goal participants reporting increased PA enjoyment.

Between 64% and 72% of participants rated the DTx at or above the standard System Usability Scale cutoff for acceptable usability.

• The a priori usability benchmark was defined as at least 51% of participants reporting adequate usability.
• Usability was assessed using the System Usability Scale (SUS).
• 23 to 26 out of 36 participants rated the DTx at or above the standard SUS acceptable usability cutoff point.
• All four DTx versions met the usability benchmark.

More than 60% of participants reported satisfaction with all DTx components, supporting sustainability and feasibility benchmarks.

• The a priori feasibility benchmark was defined as fewer than 70% of participants reporting dissatisfaction.
• 22 out of 36 participants (>60%) reported satisfaction with all DTx components.
• Sustainability and feasibility were assessed using the Delighted-Terrible Scale.
• Both sustainability and feasibility benchmarks were met across DTx versions.

There was evidence for equity, with plausibility and accessibility comparable across multiple demographic and health-related subgroups.

• The equity benchmark required that equity and accessibility be approximately equal across subgroups.
• Subgroups examined included sex, race, ethnicity, income, age, BMI, mobility, and physical constraint.
• Equity was assessed via interviews and questionnaires.
• The equity benchmark was met, with no substantial disparities in plausibility or accessibility observed across subgroups.

The study used a small sample size with stratified random assignment to test four DTx versions over 14 days in adults with overweight or obesity.

• N=36 participants were enrolled (mean age 46, SD 14 years; 54% female).
• Participants were assigned to one of four DTx versions (n=9 each): (1) intensity PA goals, (2) affect PA goals with type and context recommendations, (3) affect PA goals with savoring exercises, and (4) affect PA goals with type, context, and savoring.
• Participants wore smartwatches to track PA and completed ecological momentary assessment (EMA) questions about how they felt during PA.
• The study ran for 14 days with daily intervention sessions.

The study applied an a priori benchmark-driven formative evaluation framework to minimize confirmation bias in early-stage digital therapeutic development.

• Benchmarks were defined a priori using the DTx Real-World Evidence (RWE) Framework.
• The framework covered seven domains: safety, plausibility, accessibility, usability, sustainability, feasibility, and equity.
• The approach used small sample sizes to facilitate efficient iteration and agile learning.
• Performance was assessed via IRB reporting (safety), interviews (plausibility, accessibility, usability), and questionnaires (SUS, Delighted-Terrible Scale, equity measures).