Apparent opioid toxicity deaths
Latest four quarters are preliminary; broader BC/Quebec definitions apply.
THE REAL DOPE PROJECT · CANADA
Count the harm. Question the story. Protect people.
Deaths are real. They are not a verdict on everyone who uses substances.
A death count cannot tell us a person’s risk per occasion of use, why a policy worked or failed, or how many people survived. Ask what was counted, who was included and what changed.
Understanding the numbers
Count the harm. Keep people in the picture.
Every recorded death represents a person whose life mattered. Every emergency deserves attention. Understanding the numbers helps us ask better questions—and build better ways to protect people.
Different numbers tell different stories.
Deaths, hospital stays, emergency department visits, and ambulance responses measure different things. One person or emergency may appear in several records. We cannot add these totals together to count people, or subtract deaths to work out how many survived.
A total needs context.
Ask what was counted, where, and over what period. Did reporting coverage or definitions change? Are the figures preliminary? A rise or fall deserves investigation; the number alone does not explain what caused it.
Death counts do not tell us someone’s individual risk.
They do not show how many people used a substance, how often, or under what circumstances. They cannot, on their own, tell us the risk each time someone uses.
Survival matters, too.
Death totals cannot tell the whole story of people who survived, the harms they experienced, or the support they need. Prevention, emergency response, and ongoing care all belong in the picture.
Evidence should help us act.
At TRDP, we ask what the findings mean for practical safety: What could prevent harm? What would help someone recognize an emergency? What support is missing? Are people able to use the guidance and services offered?
Numbers can help us assess policies and safety practices. They should never become a judgment on someone’s worth.
Harm reduction for all means using evidence to protect people throughout the lifespan—whether they use substances or not.
Four measures. Four different questions.
Reported January–March 2026, from the jurisdictions available to PHAC. These categories overlap: do not add them or subtract deaths to calculate survivors.
Opioid poisoning hospitalizations
Hospital stays, not unique people.
Opioid poisoning ED visits
Visits; incomplete geographic coverage.
EMS responses to suspected opioid overdoses
Responses; regional definitions and coverage vary.
Deaths over time: keep the definition attached.
Canada, accidental drug poisoning deaths, all drugs: Statistics Canada’s February 19, 2025 publication. This is not PHAC’s opioid toxicity series. Historical published estimates may be revised; we have not merged them with newer totals.
| Year | Deaths | Visual comparison |
|---|---|---|
| 2000 | 795 | |
| 2001 | 811 | |
| 2002 | 881 | |
| 2003 | 758 | |
| 2004 | 802 | |
| 2005 | 930 | |
| 2006 | 1,042 | |
| 2007 | 1,146 | |
| 2008 | 1,154 | |
| 2009 | 1,244 | |
| 2010 | 1,324 | |
| 2011 | 1,453 | |
| 2012 | 1,538 | |
| 2013 | 1,568 | |
| 2014 | 1,769 | |
| 2015 | 2,176 | |
| 2016 | 3,286 | |
| 2017 | 4,558 | |
| 2018 | 4,673 | |
| 2019 | 4,039 | |
| 2020 | 6,412 | |
| 2021 | 7,405 | |
| 2022 | 7,179 | |
| 2023 | 7,162 |
Before 2016: 19,391 recorded accidental drug poisoning deaths across 2000–2015. TRDP calculation: sum of the published male and female annual counts. All drugs; excludes intentional and undetermined poisonings. Not an all-time total and not an opioid-only count.
When did Canada start counting?
Drug poisoning was recorded before 2016. June 6, 2017 marks a national opioid surveillance release, not the first-ever Canadian overdose death record. The earliest dedicated drug-overdose publication has not been established in this review.
- 1979–1999
ICD-9 mortality coding. Drug-poisoning records predate modern opioid surveillance.
- 2000
Canada switched mortality coding to ICD-10. Check comparability before joining older series.
- 2017-06-06
National opioid surveillance release: 2016 data, initially at least 2,458 deaths. That early estimate was provisional, not a current 2016 total.
- 2017-05-01 / 2021-09-01
Ontario introduced an enhanced coroner collection tool, then a new case-management system.
- 2018-04-01
Fentanyl-specific diagnosis codes introduced for hospital and ED records.
- 2021-11 / 2022-05
Winnipeg switched incidents to persons receiving naloxone, revising all years; Saskatchewan expanded its EMS definition, affecting data from 2021 onward.
- 2023-04
Manitoba EMS coverage narrowed to Winnipeg.
- 2024-03-27
PHAC’s release introduced new ED data. ED records existed earlier; this was not the first recording of an overdose.
- 2025 onward / depuis 2025
National death totals include broader unregulated-drug deaths in BC and drug/opioid intoxication deaths in Quebec. Not a strictly unchanged opioid-only series.
Definitions and coverage notes · 2017 · ICD / CIM · 2024
What was happening around the increases?
Health Canada links the rise around 2016 to increasing supply potency, including fentanyl. Its account of the pandemic includes a more toxic supply, disrupted services and more solitary use. Statistics Canada reports sharp increases in 2016 and 2020. Timing alone does not establish how much any policy or reporting change caused.
Read Health Canada’s explanation
Challenge the double standard. Bring the evidence.
TRDP advocacy: if safety is for everyone, emergency care and practical safety knowledge must not depend on approval of someone’s choices. Criticize policies and actions; do not turn people into targets.
- Check the exact statement, date, source and jurisdiction.
- Compare the stated safety promise with verified access, funding and outcomes.
- Ask whether definitions, coverage or population size changed.
- Offer a factual correction and a useful next step, without humiliation.
This review does not establish that a named leader manipulated data or that reporting changes explain the crisis. Death, ED and EMS counts also cannot prove TRDP’s effectiveness.
What could reducing harm save over ten years?
We cannot price the cost of “keeping people uninformed”: no verified Canadian counterfactual measures that effect. Public knowledge has not been shown to remain constant. Education alone does not make substance use safe.
CSUCH estimated opioid-related economic costs at $7.1 billion in 2020, across healthcare, productivity, criminal justice and other costs. This is broader than overdose costs and is not a measure of ignorance.
| Reduction | Gross avoided loss / 10 years |
|---|---|
| 0% | $0 CAD |
| 1% | $710 million CAD |
| 5% | $3.55 billion CAD |
| 10% | $7.1 billion CAD |
Calculation: $7.1 billion × 10 × assumed reduction. Net economic benefit = avoided loss − programme costs. Avoided economic loss is not necessarily cash available to governments. We have no verified effect size or budget for TRDP, so none of these savings can be claimed for it.
Test education alongside naloxone access, care and service continuity, voluntary treatment and recovery support. Measure understanding, actual access and repeat emergencies separately. Reducing or stopping use is a choice; dependence may require medical support to avoid dangerous withdrawal.
CSUCH: economic baseline and limits
Indigenous respect means more than a welcome.
First Nations, Inuit and Métis are distinct peoples, not a single audience. TRDP does not speak for them or claim Indigenous endorsement. Follow Indigenous-led knowledge; respect community authority over stories, cultural imagery and data. Do not presume identity, trauma or substance use. Cultural safety requires accountability to the people receiving care.
FNHA: Indigenous-led harm reduction (BC context)
Sources and review limits
- Statistics Canada · historical accidental drug poisoning deaths
- PHAC · current harms and reporting coverage
- PHAC · technical notes and definition changes
- PHAC · June 6, 2017 national release
- Statistics Canada · cause-of-death coding
- Health Canada · toxic supply, stigma and access barriers
- PHAC · March 27, 2024 ED data release
- FNHA · Indigenous-led harm reduction
Public education. French text is an AI-assisted translation awaiting independent language review. Indigenous community review and a full human accessibility audit remain pending.