Methodology
Ontario Health publishes no bulk file for ER wait times, so this project scrapes the monthly per-site tables from the agency's own data endpoints and rebuilds them as one clean time series. Every rule below is applied exactly as described.
The cleaning rules
01
Scrape the site's own endpoints
The monthly tables come from Ontario Health's public data endpoints (the same ones that feed its wait-time tool), called once per hospital per measure with polite pacing. Nothing is estimated or modelled.
02
One stable key per hospital
Hospital names change month to month (accents appear, sites rename). Every row is keyed on the site's numeric ID, and the latest published name is used as the display name, so renames never split one hospital into two.
03
Quarantine, don't guess
Months with a missing average are excluded from that hospital's series and logged in the quarantine list, never interpolated. A gap in the chart means the data was not published, not that the wait was zero.
04
Geocode with confidence tiers
Hospital coordinates come from Nominatim address matching, trying the full site name first, then the corporation name. Each match carries a confidence tier, and ungeocoded sites are excluded from the nearest-ER view rather than placed by guess.
05
Rank on the published average
Rankings use Ontario Health's published monthly average in hours. Averages, not medians: a few very long stays pull the average up, which is stated wherever rankings appear.
What each measure means
First assessment (PIA)
Average hours from check-in, counted from triage or registration whichever came first, to first assessment by a doctor, nurse practitioner, or dentist. This is the headline number: how long you wait before a doctor first sees you.
Low-urgency stay (L)
Average hours low-urgency patients spent in the emergency department without being admitted. For minor cases, this is how long the whole ER visit took.
High-urgency stay (NAH)
Average hours high-urgency patients spent in the emergency department without being admitted. For serious cases that did not need a hospital bed, this is how long the ER visit took.
Admitted stay (A)
Average hours admitted patients spent in the emergency department before moving to a hospital bed. This measures the wait after the decision to admit, and it is where hallway medicine shows up in the numbers.
What this does not do
- Monthly averages only. No open source carries visit-level timestamps, so the worst hour to arrive is honestly unanswerable from public data.
- Averages, not medians. A small number of very long stays moves the average more than the median.
- As of the December 2025 data refresh, 43 additional sites were added to reporting. Ontario Health warns that data exported after that should not be compared to earlier exports, so year-over-year movers near that boundary carry extra uncertainty.
- Nearest-ER distances are straight-line from geocoded addresses, not drive times.
Honest unknowns
- Why a specific hospital got faster or slower. The data shows the change, not the cause (staffing, volume, policy).
- Whether ungeocoded sites would change any nearest-ER result. Sites without coordinates are excluded from that view.
- How small hospitals that are not required to report would rank. They are absent from the source, not from the province.
National context (CIHI)
CIHI's National Ambulatory Care Reporting System counted more than 16.1 million unscheduled ED visits in Canada in 2024-2025, up from almost 15.5 million the year before. One in 10 patients waited more than 14 hours, a 28% increase from 2018-2019. For admitted patients, 9 out of 10 ED visits were completed within 48.5 hours; 1 in 10 admitted patients spent more than two days in the emergency department, up 33% from 2018-2019.
Source: CIHI, NACRS Emergency Department Visits and Lengths of Stay, February 2026.
Quarantine log
153 month-site rows were excluded for missing values and logged here instead of being interpolated.
Source: Ontario Health monthly per-site tables, 2025-08 to 2026-08 (monthly).