Chart 1 — rolling mean triage-severity by patient group

The chart your Chief Medical Officer should look at first. Where the silent H3 divergence begins, in one visual.

What you are looking at

• X-axis: Day 1 through Day 90 of the audit period
• Y-axis: 7-day rolling mean of the AI's severity score (0-100 scale)
• Each line: one patient group (H1 = young low-comorbidity, H2 = middle-aged moderate-comorbidity, H3 = elderly multimorbid, H4 = pediatric, H5 = obstetric)
• Baseline window: the first 30 days establishes what "normal" looks like per patient group

What the eye sees

Four patient groups stay roughly parallel across the 90 days. One patient group — H3 — drops. The drop begins around Day 44, and it does not recover. The line separates from the pack.

This is the pattern a physician-in-the-room instinctively recognizes: something in the H3 stream changed, and the AI is scoring it as less severe than baseline. The 30-percentage-point-lane-shift follow-on is the operational consequence.

What the sensor fires on

The sensor is not looking at the eye's pattern. It is running four quantitative thresholds against each 7-day window per patient group:

• Mean-shift: is the 7-day mean > 1.5 standard deviations from baseline mean?
• Variance-ratio: is the 7-day variance > 1.5x baseline variance?
• KL-divergence: is the shape divergence between baseline and current distribution > 0.2?
• Group-differential: is the max cross-group shift > 12%?

Day 44 fires H3 KL-divergence 2.108 (10x the threshold). The pattern the eye sees is the pattern the sensor quantifies.

What regulators care about

The chart also tells the counterfactual story: had this monitoring been in place operationally, the drift window would have been ~7-14 days rather than 45+.

The next chart

The rolling-mean chart shows when. The baseline-vs-drifted chart shows how the H3 distribution shape shifted.

Chart 2 — baseline vs recent distribution →