Developing a digital ecological momentary assessment tool for 'real time' evaluation in implementation science:...

Social prescribing is increasingly adopted as a strategy to address psychosocial determinants of health, yet evaluating complex community-based interventions remains methodologically challenging. Digital ecological momentary assessment (EMA) offers potential for capturing real-time, multi-modal data in naturalistic settings. This study aimed to develop and assess the feasibility of a novel digital EMA approach integrating wearable-derived physiological data with repeated self-report wellbeing measures within a digital social prescribing (DSP) context. Despite increasing interest in EMA and wearable technologies, their integration within social prescribing evaluations remains underexplored.

A mixed-methods feasibility study was conducted alongside a four-week online chair-based yoga programme delivered through a social prescribing service in East London, UK. Participants wore smartwatches to collect physiological indicators (stress, sleep, heart rate) and completed twice-daily wellbeing assessments using an adapted Short Warwick-Edinburgh Mental Wellbeing Scale. Quantitative data were analysed using exploratory linear mixed-effects models to examine data behaviour and integration within the intensive longitudinal dataset. Participant workshops explored feasibility and acceptability.

Thirteen participants were recruited, with eleven included in quantitative analyses. The study indicated that integrating wearable and self-report EMA data within routine-style DSP delivery is feasible, although challenges were identified regarding device synchronisation, questionnaire adherence, and missing data. Exploratory modelling illustrated substantial within- and between-person variability in wellbeing trajectories and provided an indication of the feasibility of analysing intensive longitudinal EMA data, with no consistent associations between same-day yoga participation and physiological stress markers. Qualitative findings suggested that participants found the approach acceptable and highlighted factors influencing engagement, including flexibility, motivation, and perceived burden.

Integrated wearable and self-report EMA methodologies can be deployed in DSP contexts, but their implementation is associated with important methodological and technical challenges. Findings highlight the need for validated EMA measures, improved data infrastructure, and careful management of participant burden. While not designed to assess intervention effectiveness, this study provides one of the first applied demonstrations of integrating wearable-derived physiological data with EMA self-report measures in a DSP context, offering a methodological foundation for real-time evaluation of complex community-based health interventions.