- Human-centred iterative design with a community advisory board produced a feasible, usable, acceptable adaptive mHealth adherence intervention, though further optimisation is required.
- Usability and acceptability ratings were high across prototypes; enrolment was 63% and participants completed mean 67.2% of automated check-ins.
- Field trial suggested medium effect on self-reported adherence; only percentage of time taking medications over past month increased significantly.
JMIR Form Res. 2026 Aug 5;10:e87124. doi: 10.2196/87124.
ABSTRACT
BACKGROUND: Adolescents and young adults with chronic health conditions often struggle to adhere to their daily oral medications. Transdiagnostic mobile health (mHealth) interventions have the potential to promote medication adherence by reaching youth at a large scale.
OBJECTIVE: This study aimed at designing an adaptive medication adherence mHealth intervention (Adaptive Cell Phone Support), guided by iterative feasibility, usability, and acceptability feedback. A secondary objective was to explore changes in self-reported medication adherence during a field trial.
METHODS: Using human-centered design methods, researchers collaborated with a community advisory board of young adult patients to conduct 3 cycles of iterative design and usability testing. Adolescents and young adults aged 15-20 years (N=22) were recruited from a large pediatric hospital to user-test the intervention. Data collection included self-report questionnaires, think-aloud usability testing, semistructured interviews, and a 3-week field trial. Quantitative measures included the mHealth App Usability Questionnaire Ease of Use and Usefulness subscales, the Theoretical Framework of Acceptability Questionnaire, and visual analogue scales assessing medication adherence, as well as enrollment and engagement metrics. Qualitative data were analyzed using rapid assessment methods to identify actionable design insights, while quantitative data were analyzed using descriptive statistics and paired-samples t tests with a Holm-Bonferroni correction.
RESULTS: Enrollment was 63% and participants completed a mean of 67.2% (SD 23.5) of automated check-ins. Usability and acceptability ratings were relatively high across prototypes (eg, mHealth App Usability Questionnaire Ease of Use was mean 6.40, SD 0.64 for the initial prototype and mean 6.31, SD 0.61 for the third prototype, on a 7-point scale; Theoretical Framework of Acceptability was mean 4.33, SD 0.52 for the initial prototype and mean 4.50, SD 0.53 for the third prototype, on a 5-point scale). Qualitative data emphasized that the intervention was simple, easy to use, convenient, appropriate, and helpful for staying accountable for medication adherence, while also highlighting areas for improvement. Uncontrolled, 2-tailed, pre-post t tests estimated medium-sized improvements in self-reported medication adherence. However, only the percentage of time taking medications over the past month significantly increased (t21=3.26, d=0.70, 95% CI 0.22-1.16; P=.004).
CONCLUSIONS: Integrated qualitative and quantitative results still suggest that more refinement is needed to optimize the intervention. Partnering with community members early in the development of an intervention may improve the ultimate feasibility, usability, and acceptability of digital health tools. Human-centered design offers a rapid, practical, and creative framework for identifying what works and what needs to be improved early in the lifecycle of a new intervention.
PMID:42555969 | DOI:10.2196/87124
Share Evidence Blueprint
Save to Google Notes
Search Google Scholar
Save as PDF
⭐ My Revision List
