Skip to content
ParamergeParamerge
Stephen Lieberman
Through Paramerge
Contact

PAN Lab example

Advance Alert Monitor (AAM) deterioration model

The alert that never reaches the bedside: a screened deterioration model

A deterioration model scores inpatients hourly and alerts about twelve hours ahead. Modeled on AAM. But the alert never reaches the bedside directly: it goes to a dedicated regional tier of critical-care nurses who screen every alert and only then escalate. So watch two things the mortality number cannot show you: that the measured benefit was priced against that whole staffed screening tier - not the model alone - and what happens if the tier is thinned to save money.

Stylized model of a documented deploymentClinical decision support & deterioration alerting

Open this example in PAN Lab v0.1 to apply pressures and levers and watch what the system does.

What this models

This example runs on the AAM-class deterioration alert with a screening tier network: 6 components and 10 pathways between them. Every context in the Lab is a stylized model, never a reconstruction of any actual deployment, and each assumption behind it carries a provenance label.

Evidence base: 5 assumed · 1 published baseline. In the Lab, the shaded evidence band behind each headline readout draws its width from the least-established class below.

  • assumed

    The regional alert queue is drawn as a worklist because the operating-model source describes alerts from twenty-one hospitals being routed to one regional tier of virtual quality nurse consultants who screen every alert around the clock, work up the chart, and only then escalate to the on-site team. The queue is the routing layer that lets a single tier serve a region; without it the two-tier topology on this board reads as two desks rather than as one pooled screening function.

  • assumed

    This deployment is drawn well-resourced - staffing above workload - which is the unusual finding here and the opposite of its sibling deterioration org. The evaluation of record prices its benefit against the whole two-tier staffing topology, not against the model, so a dedicated round-the-clock screening tier is the standing organizational cost the benefit depends on. The independent-validation check is drawn faint rather than empty for the same reason: a peer-reviewed New England Journal of Medicine evaluation of this program exists, which is rare in this catalogue and is the honest difference between this org and the vendor-self-reported ones.

  • baseline

    This models the two-tier screened-alert topology documented in the AAM case file - not a reconstruction of the actual model. Its defining feature is that the alert goes to a dedicated regional screening tier (virtual nurse consultants), not the bedside, and the measured mortality benefit is priced against the whole model-plus-screening-tier topology, not the model alone.

  • assumed

    The screening tier is drawn active (substantial on both the alert-in and the escalation-out pathways) because it is a genuine, staffed, 24/7 function that carried the benefit by absorbing the model's false-alarm load before the bedside saw it. Its capacity is the real constraint: understaff it relative to alert volume and the two-tier topology collapses into an unscreened flood, so the benefit is losable by thinning the tier without touching the model.

  • assumed

    The independence absence is drawn on the independent model check, empty at baseline: the evaluation of record was run inside the deploying system, so the strongest evidence is developer-affiliated, and no validation by a party with no stake is in the record. Per-alert precision was not published; the existence of a dedicated screening tier is itself the system's answer to the raw alert stream's noise.

  • assumed

    No patient or clinical outcome is modeled here. This Lab reads institutional propagation only, and the patients being scored are boundary-only. The mortality association and the standing staffing cost live in the case file, and are never computed from anything in this diagram.

What this example does not show

  • No patient or clinical outcome is modeled. The Lab reads institutional propagation only; the patients being scored are boundary-only, and the mortality association and the standing staffing cost live in the case file, never computed on this diagram.
  • The mortality benefit is an observational association from a developer-affiliated evaluation run inside the deploying system, not a randomized effect; per-alert precision was not published, so the screening tier's false-alarm absorption is a documented design choice rather than a measured rate.

Sources and evidence

What this example rests on, claim by claim. Every entry resolves to the same ledger the Evidence Registry publishes.

  • The Advance Alert Monitor is an in-hospital deterioration model running around the clock across 21 hospitals of an integrated health system, scoring inpatients hourly and firing roughly twelve hours before predicted deterioration; a 2020 New England Journal of Medicine evaluation associated its alert-driven rapid-response workflow with lower mortality. Its defining feature is where the alert goes: not to the bedside, but to a dedicated regional tier of critical-care virtual quality nurse consultants who screen every alert around the clock, work up the chart, and only then escalate to the on-site rapid-response team — so the measured benefit is priced against the whole two-tier staffing topology, not the model alone.

    empirical
    • Academic Escobar, G.J., Liu, V.X., Schuler, A., Lawson, B., Greene, J.D., & Kipnis, P. (2020). Automated Identification of Adults at Risk for In-Hospital Clinical Deterioration. New England Journal of Medicine, 383(20), 1951-1960. https://doi.org/10.1056/NEJMsa2001090 https://www.nejm.org/doi/full/10.1056/NEJMsa2001090
    • Academic The Kaiser Permanente Northern California Advance Alert Monitor Program: An Automated Early Warning System for Adults at Risk for In-Hospital Clinical Deterioration (2022). Joint Commission Journal on Quality and Patient Safety. https://www.jointcommissionjournal.com/article/S1553-7250(22)00110-6/fulltext

Where this connects

Institutional pressures in this domain

  • Workload surge — Demand outruns staffing; per-case attention shrinks and review becomes triage.
  • Reviewer bottleneck — One fixed-capacity checking stage sits between AI output and consequence; everything queues behind it.
  • Vendor opacity — The deploying institution cannot inspect the model, data, or update pipeline it is accountable for.
  • Data & policy drift — The world, the intake process, and the rules change under a system trained on how things used to be — two mechanisms with different remedies: the statistical properties of what the system processes move (concept drift), or the mixture of inputs arriving in deployment differs from the mixture it was trained on (covariate shift).
  • Deadline pressure — Statutory or managerial timeliness rules reward fast approval of machine output over slow disagreement.

All of them in context on the Clinical decision support & deterioration alerting domain page.

Levers available here and the patterns behind them

Documented case histories