RationAI: Feature Set for Personalized Alcohol Reduction Intervention in UK Veterans

Daniel Leightley's dataset contains features for modeling RationAI, a personalized AI-supported messaging framework tested to reduce alcohol consumption. The data originates from a 12-week feasibility study involving 2,871 UK Armed Forces veterans using the DrinksRation mobile app, with 343 participants in a personalized messaging group and 385 in a control group. The dataset was last updated in April 2026.

Use Cases

• Modeling the impact of personalized vs. generic messaging on app engagement based on usage time data.
• Analyzing the relationship between message delivery frequency and alcohol consumption reduction.
• Predicting participant retention and long-term engagement based on early usage patterns.
• Comparing intervention effectiveness between groups at specific weekly intervals where significant differences were observed.

Strengths

• Data is derived from a substantial initial cohort of 2,871 registered participants.
• Includes comparative metrics between intervention and control groups, such as app usage seconds and message counts.
• Results show statistically significant between-group differences in alcohol reduction at multiple time points.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The study reported high attrition, with only 25.4% meeting engagement criteria.

Provenance

Source

Daniel Leightley via figshare

Collection Method

Collected during a 12-week randomized feasibility study using the DrinksRation mobile app.

Time Range

Study period covered 12 weeks; specific dates are not provided.

Freshness

Last updated 2026-04-10 17:45:16; freshness should be verified.

Geography

United Kingdom (UK Armed Forces veterans).