600 patients67,035 observations120 daysLIVE
Decision memos written from the data: segments, regimens, diagnostics worth adding at baseline, combinations that keep patients. Rebuilt from the learning layer on every visit.

What does this teach about the drug?

Synthetic demo data
Where it works best
NBL-101 works best in Poor metabolizer

33% of Poor metabolizer patients (n 108) are responding and under the churn line, against 37% overall. Enrich the next cohort and the label language for this segment; the biology is dna · drug exposure.

Evidence: 108 patients · retained response 33% vs 37%
Who needs a different regimen
Adherence < 60% patients need a different regimen

43% carry toxicity ≥ 50 today (n 180). Dose-timing and meal-timing interventions are the cheapest test: in the data, taking the dose after a meal and low-residue meals on dosing days are the primitives with the largest GI effect. A dose-timing sub-study in this segment is a small, fast trial.

Evidence: toxicity ≥ 50 in 43% of the segment
Combination candidate
NBL-101 + GI-04 keeps 32% of the patients it is given to

GI Support Protocol is the protocol that turns an accepted recommendation into a retained patient most often (n 130). Bundling it into the companion from the first dosing day for every patient in the Poor metabolizer segment is the strongest strategy candidate in the data.

Evidence: GI-04 32% (n 130) · SC-12 23% (n 48) · NP-02 12% (n 66)
The drug, in one paragraph
What this says about the drug, in one paragraph

NBL-101 keeps 37% of patients responding and under the churn line. The ones it loses leave through side effects that are predictable from biology and daily behaviour, not from the molecule alone, and the same data shows which protocol brings each biology back. The strategic asset is therefore not the drug by itself but the drug plus its steering layer: diagnostics at baseline, a companion that carries the right content, and a model that keeps learning from every patient.

Evidence: 600 patients · 368 accepted actions · 0 learned rules

Rules behind these memos: Learning. Emerging patterns: Watch.