Skip to content
OutbreakWatch

Methodology

How OutbreakWatch classifies activity

Every colored area on the map uses one of the methods below, and every map layer links here. We never turn raw case counts into a personal risk score, and we show Insufficient Data instead of guessing.

Activity level colors

  • Low — Low reported activity
  • Moderate — Moderate reported activity
  • Elevated — Elevated reported activity
  • High — High reported activity
  • Insufficient Data — Insufficient or unavailable data — this does not mean there are no infections

Some layers are not levels at all: trend layers (U.S. counties) show direction of change; count layers show reported numbers in purple shades; notice layers show where an official notice exists. Each has its own legend so they are never confused with severity.

How recent is the data?

Official data is published weekly at best, so nothing on OutbreakWatch is real-time. Next to each “data through” date we show one of these labels, counted from the end of the latest reported week:

  • Current — the latest week ended 14 days ago or less.
  • Delayed — 15 to 56 days ago. The publisher reports late or irregularly; treat the level with caution.
  • Historical — more than 56 days ago, or totals for an earlier year. It describes the past, not what is happening now.
  • Year to date — a running total for this year that the publisher updates on its own schedule.

Official notices (CDC travel notices, WHO Disease Outbreak News) carry their publication date instead. “Significant activity changes” are only created from Current data.

Plain-language summaries

Summary sentences such as “Low levels” or “Emergency visits for influenza are increasing in the Harris County area” are written by OutbreakWatch from the levels and trends above. They always name the area the data covers (a country, a whole state, or a CDC Health Service Area) and never add numbers or levels that the data does not contain.

CDC-published

CDC hospital admission levels (NHSN)

CDC assigns each state a weekly level (Very Low → Very High) from new laboratory-confirmed hospital admissions per 100,000 people, using virus-specific thresholds. OutbreakWatch displays CDC's level; it does not compute its own.

How it works

  1. CDC's National Healthcare Safety Network collects weekly admissions from U.S. acute-care and critical-access hospitals.
  2. CDC divides admissions by population (per 100,000) and applies its published, virus-specific thresholds for COVID-19, influenza and RSV. OutbreakWatch does not change them; the exact cut-offs are on CDC's data page linked under Source.
  3. OutbreakWatch maps CDC's five levels onto the four-color legend: Very Low and Low → Green, Moderate → Yellow, High → Orange, Very High → Red. CDC's original label is always shown next to the color.
  4. Trend compares this week's admission rate with the previous week (±10% or more = increasing/decreasing).

Limitations

  • Hospital admissions measure severe illness, not every infection; they lag infections by 1–2 weeks.
  • Values reflect hospitals that reported that week. CDC does not backfill late reports.

CDC-published

CDC acute respiratory illness levels

CDC assigns each state a weekly level for all acute respiratory illness seen in emergency departments (colds, flu, COVID-19, RSV and others), comparing it with that state's own past seasons. OutbreakWatch displays CDC's level unchanged.

How it works

  1. CDC measures the share of emergency department visits for acute respiratory illness in each state.
  2. CDC compares it with the state's historical baseline and publishes Very Low to Very High, using thresholds CDC documents with the data. OutbreakWatch does not change them.
  3. OutbreakWatch maps Very Low and Low → Green, Moderate → Yellow, High → Orange, Very High → Red, and shows CDC's label.

Limitations

  • Covers all respiratory illness together, so it cannot tell which virus is driving activity.

CDC-published

CDC wastewater viral activity levels

CDC rates each wastewater monitoring site's virus level (Very Low → Very High) by comparing it with that site's own baseline. OutbreakWatch counts how many sites in a state are in each CDC category; it does not create a state level.

How it works

  1. Sewage samples are tested for virus genetic material at participating treatment plants.
  2. CDC compares each site's result with that site's own baseline and assigns a category using CDC's published wastewater thresholds; OutbreakWatch does not change them.
  3. OutbreakWatch shows the number of sites per category for the latest week.

Limitations

  • Not every community has monitoring; sites cover different population sizes.
  • Wastewater trends often lead hospital data by days to weeks.

CDC-published

CDC emergency department trends (NSSP)

CDC reports, for each county's Health Service Area, the share of emergency department visits diagnosed as COVID-19, influenza or RSV and whether that share is increasing, decreasing or unchanged. OutbreakWatch shows CDC's trend; CDC does not publish county severity levels, so none is shown.

How it works

  1. Hospitals send emergency department visit records to CDC's National Syndromic Surveillance Program.
  2. CDC groups counties into Health Service Areas (HSAs) and calculates the percentage of visits for each virus.
  3. CDC labels each week's trend Increasing / Decreasing / No Change, or marks the data as too sparse.

Limitations

  • Several counties share one HSA, so neighboring counties often show identical values — the data is not county-specific.
  • Not every hospital participates; coverage varies by state.

OutbreakWatch method

OutbreakWatch history-relative method (OWM-HIST-1)

Countries test and report very differently, so OutbreakWatch never compares one country's raw numbers with another's. Instead each country is compared with its own recent history: how unusual is the last 4 weeks compared with the same country's past seasons?

How it works

  1. Current value = the 4 weeks ending with the most recent reported week, combined (influenza: positive tests ÷ tests processed; COVID-19 & mpox: average weekly reported cases per million). Weeks with no report are left out — never counted as zero.
  2. Baseline = the same 4-week value calculated for every week of the country's own history (influenza: up to 3 prior years; COVID-19: up to 2 prior years), excluding the most recent 8 weeks.
  3. Percentile = share of baseline weeks below the current value. Below 50th percentile → Low (green); 50th–74th → Moderate (yellow); 75th–89th → Elevated (orange); 90th and above → High (red).
  4. Influenza only: if fewer than 5% of tests are positive the level cannot be higher than Low, so small numbers of positives in quiet countries are not exaggerated.
  5. Insufficient Data (gray) when: the latest report is older than the disease's freshness limit (influenza and RSV: 4 weeks; COVID-19 and mpox: 8 weeks), fewer than 2 of the last 4 weeks were reported, too few tests (influenza and RSV: fewer than 30 specimens), or too few baseline weeks exist (40 weeks; mpox: 20 weeks, because WHO's mpox series is shorter).
  6. Trend compares the current 4-week value with the previous 4 weeks: +20% or more = increasing, −20% or more = decreasing.

Limitations

  • This is an OutbreakWatch classification, not an official WHO or national level.
  • It describes reported activity relative to a country's own past — not an individual's risk of infection.
  • Changes in a country's testing or reporting practices can shift its level without a real change in spread.

OutbreakWatch method

Reported case counts (no severity level)

Shows cases exactly as the authority reported them, shaded by count. No severity level is assigned because the source does not provide one and counts are not comparable across places with different testing.

How it works

  1. Counts are copied from the source without modification.
  2. Shading uses the source's own ranges where provided, otherwise logarithmic bins.

Limitations

  • Counts depend on testing and reporting; larger counts do not by themselves mean higher personal risk.

OutbreakWatch method

Official notices and outbreak reports

Highlights places named in an official notice (CDC Travel Health Notice or WHO Disease Outbreak News). The highlight means a notice exists — it is not an activity level, and places without a notice are not necessarily free of disease.

How it works

  1. Notices are matched to countries by the names the publisher lists.
  2. WHO outbreak reports are shown for 120 days after publication.

Limitations

  • Notices cover only events the publishing authority chose to report.

OutbreakWatch method

CDC wastewater detections

Counts how many wastewater sites in each state reported a detection of the target in the last 4 weeks of CDC's National Wastewater Surveillance System data. A detection shows genetic material was found — for H5 it may come from animal sources such as milk or birds, not people.

How it works

  1. Site samples with a 'detected' result in the last 28 days are counted per state.
  2. The number of sites tested is shown alongside.

Limitations

  • Many communities are not covered by wastewater monitoring.
  • No detection does not prove absence.

OutbreakWatch method

Combined view

For each place, shows the highest level among the disease layers that publish a level for that geography (U.S. states: COVID-19, influenza, RSV hospital levels; countries: influenza and COVID-19 OWM-HIST-1). It does not add diseases together.

How it works

  1. Take every level-type layer for the geography.
  2. Show the highest non-gray level; if every layer is gray, show Insufficient Data.
  3. The details panel lists which disease set the color.

Limitations

  • A place can look green while a disease without a level layer (e.g. measles) is active — check the notices layer too.
Methodology · OutbreakWatch