A comparison of personalized and fixed interval signal-contingent ecological momentary assessment to capture dietary...

Ecological momentary assessment (EMA) may address limitations of traditional dietary assessment methods such as high burden and memory bias, improving measurement in young adults with irregular eating patterns. This study assessed the feasibility (adherence, agreement, acceptability) of delivering EMA surveys at times personalized to young adults' (18-30 years) eating patterns compared to fixed interval EMA.

Twenty-four participants (13 female; mean age 26.0, SD 2.1) were randomized to start with either the personalized or fixed EMA schedule in a double-blinded crossover trial. Personalized schedules were based on time-stamped food images collected before randomization. Adherence was response rate and agreement was the proportion of EMA responses that matched with simultaneously collected food images and 24-h recalls.

Adherence was 65.7% (SD 14.8%) and 66.3% (SD 18.2%), and agreement was 52.0% (SD 18.5%) and 47.7% (SD 29.9%) for the fixed and personalized EMA schedules. Beverages were most frequently omitted when compared against food images and 24-h recalls. Both groups found survey length and recording duration acceptable, but a greater proportion in the fixed interval phase reported receiving too many EMA per day.

Personalizing EMA delivery to young adults' eating patterns did not improve adherence as, on average, there was a similar number of daily surveys whether personalised or our predefined six prompts and variability in timing of eating occasions made personalizing challenging. Future studies could explore advanced personalization, such as wearable sensor-triggered event-contingent EMA combined with images or digital entries to enhance accuracy while reducing burden.