BR-03 · Behavioural Recommendation
Behavioural recommendationSynthetic demo dataTarget state: Adherence decline · onset 7 d · prior: symptom −10%, persistence +10%, tolerance +7%Simulate for P-01454
For P-01454
Eligible at 75% confidence. Among the 284 most similar trajectories: 2 accepted, 100% improved, mean effect −14%.
What it is
A 14-day behavioural plan: fixed sleep window, a 20-minute walk after the main meal, a 12-hour overnight eating pause and 5 minutes of slow breathing before each dose.
How it works
Targets the behavioural spiral (poor sleep → higher GI and stress → lower adherence). Improves HRV and sleep quality first; GI follows in patients whose pattern is stress- or sleep-led.
What the patient does, day by day
| Every night | ✓ | Sleep window 22:30–06:30 Screens off 45 min before. |
| Every day | ✓ | 20-minute walk after the main meal Tracked by steps. |
| Every day | ✓ | 12-hour overnight eating pause Last food 20:00, first 08:00. |
| Dosing days | ✓ | Breathing before dosing 5 minutes, guided in the companion. |
Delivery
- Owner
- Companion
- Channel
- Daily nudges; wearable-verified
- Duration
- 14 days
- Touchpoints
- Evening sleep reminder; post-meal walk reminder
- Cost
- ≈ €25 per patient
Monitoring
- · Sleep quality
- · HRV
- · Steps
- · Stress check-in
Stop rules
- · GI ≥ 60 for 3 days → switch to GI-04
Contraindications
- · Mobility limitation for the walk (replace with seated activity)
Evidence in NBL-101
10 accepted · 50% improved · mean effect −10%By biology segment
| FUT2 non-secretor | n 5 | 60% |
| Low microbial diversity | n 4 | 25% |
By trajectory cluster
| EC-01 | n 4 | 50% |
Works best when
- · Sleep quality < 40 with stress > 6
- · HRV worsening
- · GI < 60
Works poorly when
- · GI ≥ 60 (biology-led, not behaviour-led)
- · Toxicity ≥ 55
What the model has learned about it
No learned rule yet for BR-03. Run learning after campaign results.
Campaigns using it
None yet. 232 patients look eligible right now.