Skip to content
ParamergeParamerge
Stephen Lieberman
Through Paramerge
Contact

PAN Lab example

Sepsis Watch deep-learning detection system

The nurse who gets the alert can't give the order: an authority-split detector

A deep-learning detector scores every emergency department (ED) patient every five minutes. Modeled on Sepsis Watch. The alert goes to a nurse - who cannot order treatment. Only the physician can, so the alert becomes care only if the nurse moves the physician across a professional hierarchy. So watch the edge the model diagram omits: the peer edge between two people with unequal authority, where the system's real work - and its hidden cost - actually happens.

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 Sepsis-Watch-class alert with an authority split network: 5 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 nurse-to-physician pathway is drawn at full strength because it is the mechanism: the operator who receives the alert is not the operator empowered to act on it, so the correction runs through peer persuasion rather than authority. Capacity is drawn low against a heavy workload because the independent ethnography found the system worked on hidden repair work - nurses mediating the professional hierarchy and doing the emotional labor of communicating a risk score upward - labor that was structurally necessary, largely invisible to the deployment's formal description, and undervalued. The machine write runs at a substantial level because scoring is continuous, every five minutes over eighty-six variables, not occasional.

  • assumed

    Both deployments front a nurse tier ahead of the acting clinician, and the earlier drafts drew them identically. The record separates them: the regional program pools alerts into one screening queue and resources it round the clock, so it reads capacity-over-demand with a published external evaluation; this one runs a single emergency department at a five-minute cadence on labor nobody costed, so it reads demand-over-capacity with the authority split as its fault line. The difference is staffing and routing, both documented, not a decorative node.

  • baseline

    This models the authority-split alerting pattern documented in the Sepsis Watch case file - not a reconstruction of the actual model. Its defining feature is that the operator who receives the alert (the RRT nurse) is not the operator empowered to act on it (the physician), so the correction runs along a peer edge between two operator classes with different authority.

  • assumed

    The nurse-to-physician peer pathway is drawn active, at a substantial level, but is where the benefit's real cost sits: an independent ethnography found the system worked because nurses performed hidden repair work - mediating the professional hierarchy and doing the emotional labor of communicating a risk score upward - labor that was structurally necessary, largely invisible to the formal description, and undervalued. The same alert produces care in one nurse-physician pairing and nothing in another, because it is a gap in who can act, not in what the model knows.

  • assumed

    The independence absence is drawn on the independent model check, empty at baseline: the implementation study of record was developer-led, and it took an outside ethnography to surface the labor the formal account omitted - exactly the kind of thing developer-produced evidence does not see about itself. Per-alert precision was not published; the nurse-fronted design is the deployment's own answer to the dense five-minute alert stream.

  • assumed

    No patient or sepsis outcome is modeled here. This Lab reads institutional propagation only, and the patients scored every five minutes are boundary-only. The authority split and the repair work live in the case file, and are never computed from anything in this diagram.

What this example does not show

  • No patient or sepsis outcome is modeled. The Lab reads institutional propagation only; the patients scored every five minutes are boundary-only, and the authority split and the repair work live in the case file, never computed on this diagram.
  • The successful-integration finding is from a developer-led implementation study; per-alert precision was not published, and the hidden-repair-work finding comes from a separate independent ethnography of the deployment rather than from a randomized comparison.

Sources and evidence

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

  • Sepsis Watch is a deep-learning sepsis-detection system scoring every emergency-department patient every five minutes over 86 variables, deployed at an academic hospital under a registered clinical trial, with alerts fronted by rapid-response-team nurses who track treatment-bundle completion on three- and six-hour timers. Its structural fault line is an authority split: the operator who receives the alert (the nurse) is not the operator empowered to act on it (the physician who holds treatment authority), so the correction runs through a peer-persuasion edge. An independent ethnography found the system worked because nurses performed hidden repair work — mediating the professional hierarchy and doing the emotional labor of communicating a risk score upward — labor that was structurally necessary, largely invisible to the deployment's formal description, and undervalued.

    empirical
    • Peer-reviewed Sendak, M.P., et al. (2020). Real-World Integration of a Sepsis Deep Learning Technology Into Routine Clinical Care: Implementation Study. JMIR Medical Informatics, 8(7), e15182. https://doi.org/10.2196/15182 https://medinform.jmir.org/2020/7/e15182/
    • Advocacy Elish, M.C., & Watkins, E.A. (2020). Repairing Innovation: A Study of Integrating AI in Clinical Care. Data & Society Research Institute. https://datasociety.net/library/repairing-innovation/

  • 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