Skip to content
Product · Document 01

The Outbreak Dashboard

A working instrument for hospital infection-control teams and clinical epidemiologists. Four compositing signals — clinical encounters, pharmacy co-purchase, wastewater, and traveler-clinic reports — refreshed every six hours across a single annotated map of influenza A and B activity.

Prepared for
Infection-control committees, hospital preparedness leads, state and provincial public-health epidemiologists
Refresh cadence
Every 6 hours, 04:00 / 10:00 / 16:00 / 22:00 UTC
Last revised
14 February 2025
Coverage
192 countries · 1,840+ acute-care facilities · 412 wastewater sites
§1 · Architecture

The four signals the dashboard composites

Each layer is a separate feed with its own collection methodology, latency, and coverage footprint. The map you see is a weighted composite, not a sum. Layer weights are tunable per account, and every signal carries a published provenance footnote on the panel itself.

  1. 01

    Clinical encounter reporting

    De-identified ICD-10 coded visits from a hospital-data network of 1,840+ acute-care facilities covering 71% of U.S. admissions, supplemented by ambulatory feeds from 312 partner clinics. Suspected and lab-confirmed influenza A and B encounters are extracted from chief-complaint and result fields.

    Cadence
    Every 6 hours
    Coverage
    U.S. national, metro-level
    Volume
    3.2M encounters / day at peak
  2. 02

    Pharmacy OTC and Rx co-purchase

    Aggregated, store-level point-of-sale data from 18,400 retail pharmacies across 50 states, tracking symptomatic-relief SKUs (analgesics, antipyretics, oral electrolyte solutions) co-purchased with oseltamivir or baloxavir prescriptions. The co-purchase index detects new community outbreaks up to 9 days before pediatric ED visits rise.

    Cadence
    Every 6 hours
    Coverage
    U.S., ZIP-3
    Latency
    ~18 hours from sale
  3. 03

    Wastewater surveillance panel

    Quantitative PCR for influenza A and B RNA at 412 wastewater treatment plants across 38 states — the largest privately operated influenza A/B signal in North America. Subtype-specific primer panels are updated weekly against GISAID reference sequences to preserve sensitivity as the virus drifts.

    Cadence
    Daily, 24-hr composite samples
    Coverage
    38 U.S. states, 412 sites
    Method
    qPCR, subtype-specific
  4. 04

    Traveler-clinic reports and air-travel overlay

    Case reports from 1,950 traveler-medicine clinics in 192 countries, fused with OAG air-travel capacity data to model inter-city import risk. The traveler layer surfaces activity in regions where hospital-network coverage is thin, and pre-positions surge alerts near hub airports.

    Cadence
    Daily, 02:00 UTC
    Coverage
    192 countries
    Overlay
    OAG scheduled capacity
§2 · Annotated view

U.S. influenza A activity · week 47 of the 2023/24 season

A single annotated screenshot from an enterprise instance, reading left-to-right and top-to-bottom. Callouts name the visible widgets and the clinical meaning of each.

Annotated view of the FluTrack U.S. influenza A activity dashboard for week 47 of the 2023/24 season, showing the choropleth map, trend chart, and alert panel with numbered callouts.
Figure 2.1 — Single enterprise instance, view dated 24 November 2023, 10:00 UTC refresh. Click any state to drill to metro-level encounter, wastewater, and pharmacy layers.
  1. A

    Composite activity choropleth

    State-by-state influenza A weighted-activity index, percentiles against the prior three seasons. The current "now" state is marked in the accent hue on the map's legend.

  2. B

    Refresh marker ribbon

    Vertical guide lines at each six-hour refresh tick. The most recent ribbon is filled; the four prior ticks are unfilled for at-a-glance currency checks during morning briefings.

  3. C

    Wastewater concentration panel

    Subtype-specific PMMoV-normalized RNA copies per milliliter, overlaid on encounter counts. The panel shows the wastewater-to-clinical correlation study from The Lancet Digital Health (2023).

  4. D

    Triggered alert threshold list

    Each row is a configured threshold the account has crossed, with the value, the geofence, and the contact-routed escalation path. Alerts are auditable in the activity log.

§3 · Operations

How it updates, what triggers an alert, what you can pull out, and how it integrates

The mechanics a procurement evaluator or infection-control lead needs before booking a demo. Numbers below are not aspirational — they describe the shipped product.

  1. 01

    Refresh cadence

    The composite map rebuilds four times per day at 04:00, 10:00, 16:00, and 22:00 UTC. Clinical and pharmacy layers settle within 90 minutes of their tick; wastewater updates alongside its 24-hour composite sample; traveler reports settle at 02:00 UTC. A status ribbon on every panel reports the last successful refresh and any degraded feeds.

  • 02

    Alert threshold logic

    Thresholds are configured per account against a geofence (state, HSA, county, ZIP-3) and a metric (weighted-activity percentile, encounter growth rate, wastewater PMMoV-normalized copies, pharmacy co-purchase index). Crossing a threshold fires one or more channels — email, secure webhook, HL7 FHIR observation, or in-app banner — and writes an entry to the immutable activity log.

  • 03

    Export formats

    Every panel exports to CSV, Parquet, and PNG at any refresh tick, with retained provenance columns for feed version, model version, and ingestion timestamp. PDF briefing exports are pre-formatted for hospital preparedness committee packs and include the watermark "Not for clinical diagnosis — surveillance only" by default.

  • 04

    API and integration

    A REST API and a streaming webhook expose the same metrics shown in the UI. Bulk pulls are paginated; rate limits are documented per endpoint. Pre-built connectors are maintained for EPIC, Cerner, and a FHIR R4 observation resource. SOC 2 Type II and HIPAA-compliant infrastructure; a BAA is offered on enterprise contracts.

  • §4 · Validation

    Across five consecutive flu seasons, weighted correlation between the FluTrack wastewater signal and aggregated clinical encounter counts remained r = 0.87 (95% CI 0.83–0.90), with the wastewater layer leading clinical encounters by a median of 7.2 days.

    The Lancet Digital Health (2023), "Wastewater-to-clinical signal correlation for seasonal influenza in U.S. metropolitan catchments," §3 Results.

    §5 · Datasheet

    At-a-glance specifications

    Read top-to-bottom as a datasheet. Where a value is sourced from a verification footnote, the citation is inline.

    Hospital-data network
    1,840acute-care facilities
    71% of U.S. admissions
    Wastewater sites
    412across 38 states
    Largest privately operated panel
    Refresh interval
    6hours
    04:00 / 10:00 / 16:00 / 22:00 UTC
    Peak daily encounters
    3.2M/ day
    De-identified, peak season
    Median lead time
    7.2days vs. CDC FluView
    Validated across 5 seasons
    Alert latency
    < 90min from refresh
    Webhook, email, FHIR
    Country coverage
    192countries
    Traveler-clinic + OAG overlay
    Compliance
    SOC 2 II+ HIPAA
    BAA on enterprise contracts

    A 20-minute demo walks through your account's geofences and the alert thresholds that match your facility's preparedness plan.

    Request a demo