Personalizing a weight loss program with cognitive-behavioral phenotypes dataset

Title: Dataset from a quasi-experimental study testing phenotype-tailored weight loss advice in a digital behavioural programme

Description:
This dataset supports a quasi-experimental study examining whether tailoring advice based on cognitive-behavioural phenotypes improves engagement and weight loss outcomes in a UK digital weight management programme. The research tested the hypothesis that participants who received phenotype-based advice would demonstrate greater in-app engagement and weight loss compared to two comparison groups: a historical cohort and non-responders.

Participants in the intervention group completed a 17-item questionnaire that matched them to one of four behavioural phenotypes. They were then sent seven weekly advice documents tailored to their assigned phenotype. App engagement and weight change were passively collected and compared across groups. The dataset includes baseline demographic information (e.g. age, gender, Index of Multiple Deprivation), phenotype assignment, engagement metrics (e.g. number of meals tracked, messages sent, lessons viewed), and self-reported weight change.

Notable findings:
Phenotype group participants showed significantly higher engagement levels than both comparison groups. Weight loss was higher in the phenotype group, though the difference did not reach statistical significance. Sub-analyses showed no single engagement type accounted for the difference; rather, all app activities contributed equally. The Index of Multiple Deprivation did not moderate the relationship between group assignment and outcomes.

How to interpret:
The data are anonymised. Engagement metrics are cumulative counts over the intervention period. Weight change is expressed in kilograms. Variables are labelled and explained within the file. Researchers can reuse the dataset to explore digital engagement patterns, test additional moderators, or replicate analyses in other digital health settings.

Data collection method:
Participants were existing users of a digital weight management programme. Intervention group data were collected in May 2024; historical cohort data come from the same programme delivered one year earlier (May 2023). All data were collected passively via the programme app, except for phenotype assignment (via questionnaire) and self-reported weight.