ME-01 · Monitoring Escalation
Monitoring escalationSynthetic demo dataTarget state: Any deteriorating state · onset 3 d · prior: symptom −5%, persistence +12%, tolerance +5%Simulate for P-01339
For P-01339
Eligible at 71% confidence. Among the 284 most similar trajectories: 67 accepted, 100% improved, mean effect −3%.
What it is
Monitoring escalation: the study nurse is looped in, the patient gets a daily symptom check-in and the wearable thresholds are tightened, so a deterioration is caught before it becomes a clinical event.
How it works
Does not change the biology; changes the detection latency. Every response counts as "improved" because the endpoint is early detection, not GI reduction — which is why its improved share is 100 % and its effect size small.
What the patient does, day by day
| Day 1 | 🏥 | Nurse introduction call Explains the tighter monitoring and what triggers a call. |
| Every day | ✓ | Daily symptom check-in GI, fatigue, temperature. |
| Every day | ✓ | Wearable thresholds tightened HR +6 bpm, sleep < 5.5 h, steps < 1 500 trigger a review. |
Delivery
- Owner
- Study nurse
- Channel
- Nurse dashboard alerts + companion check-ins
- Duration
- 14 days
- Touchpoints
- Daily check-in; nurse call at any threshold
- Cost
- ≈ €90 per patient (nurse time)
Monitoring
- · All key trackers, daily
- · Nurse review of alerts
Stop rules
- · Any threshold → nurse call; two thresholds → investigator
Contraindications
- none recorded
Evidence in NBL-101
70 accepted · 100% improved · mean effect −3%By biology segment
| FUT2 non-secretor | n 46 | 100% |
| IL6 high-expression variant | n 34 | 100% |
| Poor metabolizer | n 10 | 100% |
| Low butyrate capacity | n 51 | 100% |
| Low microbial diversity | n 55 | 100% |
| Baseline state S14 | n 44 | 100% |
| Cluster EC-03 (high response / high intolerance) | n 67 | 100% |
| Adherence < 60% | n 66 | 100% |
By trajectory cluster
| EC-03 | n 67 | 100% |
Works best when
- · Discontinuation risk > 50 % with unclear cause
- · Evidence for a treatment protocol is thin
Works poorly when
- · Used instead of a treatment protocol when one is clearly indicated
What the model has learned about it
No learned rule yet for ME-01. Run learning after campaign results.
Campaigns using it
None yet. 232 patients look eligible right now.