Modelling opioid-related deaths during the COVID-19 outbreak

On this page

• Overview
• Modelling and the opioid overdose crisis
• Projections
• Considerations
• Model development
• Canada's approach to the opioid overdose crisis

Overview

The Public Health Agency of Canada (PHAC) developed a simulation model of opioid-related deaths that provides information on the number of these deaths that might occur during the COVID-19 outbreak in 2020 through to June 2023.

The opioid overdose crisis continues to have significant impacts on people living in Canada, their families, and communities, and remains one of the most serious public health crises in Canada's recent history.

Data from several jurisdictions across Canada and PHAC showed a substantial increase in opioid-related harms and deaths at the beginning of the COVID-19 pandemic and deaths have remained above pre-pandemic levels since this time. Initial increases in opioid-related harms may have been related to changes in the illegal drug supply, less access to supports and services for people who use drugs, and increased use of substances as a way to cope with stress. However, these factors may have changed throughout the pandemic period, such as increased availability or access to prevention, harm reduction, or treatment services over time.

PHAC publishes data on opioid-related harms quarterly. It helps to provide a national picture, inform decision makers, and guide response efforts. These data are based on complex coroner and medical examiner investigations. At the national level, there is an average 6-month delay between a death occurring and it being included in the data.

Modelling and the opioid overdose crisis

Models use mathematical equations to estimate how many cases of a disease or health event may occur in the coming weeks or months. They help researchers simulate real-world possibilities in a virtual environment.

Models can't predict what will happen, but they can help governments and Canadians understand what might happen in certain scenarios. This can help us plan and act to achieve the best possible outcome.

PHAC developed a simulation model that provides the number of opioid-related deaths that might occur.

Simulations were based on several scenarios, specifically:

• the level of toxic fentanyl in the drug supply
• the proportion of opioid-related deaths prevented by health interventions, like supervised consumption sites and naloxone use

Please refer to the Model development section for more information.

Projections

Projections to June 2023

The results of the model suggest that, under some scenarios, the number of opioid-related deaths through to June 2023 may remain high or may decrease, but not to levels seen before the onset of the COVID-19 pandemic.

Figure 1 text description: Observed and projected opioid-related deaths, Canada, January 2016 to June 2023

• Scenario 1: Health interventions prevent the same proportion of deaths, level of fentanyl increases
• Scenario 2: Health interventions prevent the same proportion of deaths, level of fentanyl remains the same
• Scenario 3: Health interventions prevent more deaths, level of fentanyl increases
• Scenario 4: Health interventions prevent more deaths, level of fentanyl remains the same
• Solid black line: represents the model output for the period from January 2016 to March 2022, which was calibrated to observed data shown by the grey bars. Comparing the real data from the past to the simulated deaths from the model shows us that the model is well set up

• Scenario 1 (purple line): If health interventions prevent 40% of opioid-related deaths from April 2022 through June 2023 and the level of fentanyl in the drug supply increases after March 2022, then deaths may increase through to June 2023.
• Scenario 2 (green line): If health interventions prevent 40% of opioid-related deaths from April 2022 through June 2023, and the level of fentanyl in the drug supply remains the same as in March 2022, then deaths through June 2023 may decrease to levels seen at the onset of the COVID-19 pandemic.
• Scenario 3 (red line): If health interventions prevent 60% of opioid-related deaths from April 2022 through June 2023, and the level of fentanyl in the drug supply increases after March 2022, then deaths through to June 2023 may decrease but not to levels seen before the onset of the COVID-19 pandemic.
• Scenario 4 (blue line): If health interventions prevent 60% of opioid-related deaths from April 2022 through June 2023, and the level of fentanyl in the drug supply remains the same as in March 2022, then deaths through June 2023 may further decrease but not to levels seen before the onset of the COVID-19 pandemic.

Notes

PHAC's model simulates opioid-related deaths in a way that closely matches observed opioid-related deaths from 2016 through March 2022.

As opioid use and related harms have changed significantly over the past years, this model will be published twice a year. Changes between the previous model released on June 23, 2022 and the new model released on December 14, 2022:

• The model has been updated with the surveillance data on apparent opioid-related deaths through March 2022.
• The model assumes higher values for the proportion of deaths prevented by health interventions beginning in 2022 as availability and accessibility of services may have increased over the course of the COVID-19 pandemic.
• Comparisons between modelling results and the most recent surveillance data released in December 2022 should be made with caution. The model uses data from the September 2022 release, which may vary slightly from the data released in December.

Considerations

The simulations presented here reflect a combination of:

• model structure, and
• assumptions of what may happen to the illegal drug supply and the proportion of opioid-related deaths that are prevented through health interventions (e.g. naloxone or supervised consumption sites) to prevent opioid-related deaths during the COVID-19 period.

When interpreting model results, it is important to recognize that:

• if the model structure, parameter values or assumptions are incorrect, the simulations may not be an accurate reflection of what may happen
• this model considers Canada as a whole, and does not account for regional variation in the parameter values in the model
  • for example, changes to the drug supply may have started earlier in some jurisdictions, and some provinces and territories may have taken steps during the COVID-19 pandemic to adapt health interventions to prevent opioid-related deaths
• as with all models, if the projections result in the introduction of measures by officials that reduce the number of opioid-related deaths, the projected numbers will be less accurate.

PHAC will update its model as it learns more about the effect of the COVID-19 outbreak on the opioid overdose crisis.

Model development

PHAC developed a system dynamics model of opioid-related deaths.

Model parameter values were identified from:

• Statistics Canada
• Canadian Institute for Health Information
• Public Health Agency of Canada
• Health Canada
• provinces and territories, and
• peer-reviewed literature

In the model, people using opioids can die from an opioid overdose at different rates depending on the type of opioid use. Mortality rates for non-medical opioid use are also affected by:

• seasonality
• level of fentanyl in the drug supply
• proportion of deaths prevented by health interventions designed to reduce opioid-related deaths

Projections for 2022 (released June 2022)

The results of the model suggest that, under some scenarios, the number of opioid-related deaths may remain high or may even increase through to December 2022.

Figure 1 text description: Observed and projected opioid-related deaths, Canada, January 2016 to December 2022

• Scenario 1: Health interventions prevent the same proportion of deaths, level of fentanyl increases
• Scenario 2: Health interventions prevent the same proportion of deaths, level of fentanyl remains the same
• Scenario 3: Health interventions prevent more deaths, level of fentanyl increases
• Scenario 4: Health interventions prevent more deaths, level of fentanyl remains the same
• Solid black line: represents the model output for the period from January 2016 to September 2021, which was calibrated to observed data shown by the grey bars. Comparing the real data from the past to the simulated deaths from the model shows us that the model is well set up

• Scenario 1 (purple line): If health interventions prevent 30% of opioid-related deaths from October 2021 through December 2022 and the level of fentanyl in the drug supply increases after December 2021, then a further increase in deaths may occur through December 2022.
• Scenario 2 (green line): If health interventions prevent 30% of opioid-related deaths from October 2021 through December 2022, and the level of fentanyl in the drug supply remains the same as in December 2021, then deaths through December 2022 may remain high.
• Scenario 3 (red line): If health interventions improve after September 2021 to prevent 50% of opioid-related deaths, and the level of fentanyl in the drug supply increases after December 2021, then deaths through December 2022 may remain relatively consistent with levels observed from April 2020 through September 2021.
• Scenario 4 (blue line): If health interventions improve after September 2021 to prevent 50% of opioid-related deaths, and the level of fentanyl in the drug supply remains the same as in December 2021, then deaths through December 2022 may decrease but not to below levels seen at the peak of the opioid overdose crisis before the onset of COVID-19.

Notes

PHAC's model simulates opioid -related deaths in a way that closely matches observed opioid-related deaths from 2016 through September 2021.

As opioid use and related harms have changed significantly over the past years, this model will be published twice a year. Changes between the previous model released on December 15, 2021 and the new model released on June 23, 2022:

• The model has been updated with the surveillance data on apparent opioid-related deaths through September 2021.
• Comparisons between modelling results and the most recent surveillance data released in June 2022 should be made with caution. The model uses data from the March 2022 release, which may vary slightly from the data released in June.

Projections for 2022 (released December 2021)

The results of the model suggest that, under some scenarios, the number of opioid-related deaths may remain high or may even increase through to June 2022.

Figure 1 text description: Observed and projected opioid-related deaths, Canada, January 2016 to June 2022

• Scenario 1: Health interventions prevent the same proportion of deaths, level of fentanyl increases
• Scenario 2: Health interventions prevent the same proportion of deaths, level of fentanyl remains the same
• Scenario 3: Health interventions prevent more deaths, level of fentanyl increases
• Scenario 4: Health interventions prevent more deaths, level of fentanyl remains the same
• Solid black line: represents the model output for the period from January 2016 to March 2021, which was calibrated to observed data shown by the grey bars. Comparing the real data from the past to the simulated deaths from the model shows us that the model is well set up

• Scenario 1 (purple line): If health interventions prevent 30% of opioid-related deaths from October 2020 through June 2022 and the level of fentanyl in the drug supply increases after March 2021, then a further increase in deaths may occur through June 2022.
• Scenario 2 (green line): If health interventions prevent 30% of opioid-related deaths from October 2020 through June 2022, and the level of fentanyl in the drug supply remains the same as in March 2021, then deaths through June 2022 may remain high.
• Scenario 3 (red line): If health interventions improve after March 2021 to prevent 50% of opioid-related deaths, and the level of fentanyl in the drug supply increases after March 2021, then deaths through June 2022 may be lower than previous levels seen during the COVID-19 period from June 2020 through March 2021.
• Scenario 4 (blue line): If health interventions improve after March 2021 to prevent 50% of opioid-related deaths, and the level of fentanyl in the drug supply remains the same as in March 2021, then deaths through June 2022 may decrease but not to below levels seen at the peak of the opioid overdose crisis before the onset of COVID-19.

Notes

PHAC's model simulates opioid -related deaths in a way that closely matches observed opioid-related deaths from 2016 through March 2021.

Changes between the previous model released on June 23, 2021 and the new model released on December 15, 2021:

• The model has been updated with the surveillance data on apparent opioid-related deaths through March 2021.
• Comparisons between modelling results and the most recent surveillance data released in December 2021 should be made with caution. The model uses data from the September 2021 release, which may vary slightly from the data released in December.

Projections for 2021 (released June 2021)

The model's projections to December 2021 suggest that, under some scenarios, the number of opioid-related deaths may remain high or may even increase through the remainder of 2021.

Figure 1 text description: Observed and projected opioid-related deaths, Canada, January 2016 to December 2021

• Scenario 1: health interventions prevent the same proportion of deaths, level of fentanyl increases
• Scenario 2: health interventions prevent the same proportion of deaths, level of fentanyl remains the same
• Scenario 3: health interventions prevent more deaths, level of fentanyl increases
• Scenario 4: health interventions prevent more deaths, level of fentanyl remains the same
• Solid black line: represents the model output for the period from January 2016 to September 2020, which was calibrated to observed data shown by the solid grey bars. Comparing the real data from the past to the simulated deaths from the model shows us that the model is well set up (i.e. properly calibrated).

• Scenario 1 (purple line): If health interventions prevent 20% of opioid-related deaths, as was estimated from June through September 2020 during the COVID-19 outbreak, and the level of fentanyl in the drug supply increases, then even further increases in deaths may occur through December 2021.
• Scenario 2 (green line): If health interventions prevent 20% of opioid-related deaths (starting in April 2020), and the level of fentanyl in the drug supply remains the same, then quarterly deaths are projected to meet or surpass the levels seen at the height of the opioid overdose crisis in 2018.
• Scenario 3 (red line): If health interventions improve after September 2020 to prevent 40% of opioid-related deaths, and the level of fentanyl in the drug supply increases, then deaths from October 2020 through December 2021 may remain similar to levels seen during the COVID-19 period from April through September 2020.
• Scenario 4 (blue line): If health interventions improve after September 2020 to prevent 40% of opioid overdose deaths, and the level of fentanyl in the drug supply remains the same, then deaths through December 2021 may decrease but not to levels seen before the onset of COVID-19.

Notes

PHAC's model simulates opioid-related deaths in a way that closely matches observed opioid-related deaths from 2016 through September 2020.

Changes between the model released on December 16, 2020 and the new model released on June 23, 2021:

• The June model no longer uses data from the Public Health Agency of Canada's Opioid-and Stimulant-related Harms Surveillance for information on the level of fentanyl in the drug supply. Health Canada's Drug Analysis Service (DAS) data are used for this. DAS for 2019. The original October model used data from DAS for 2018 and applied these data to 2019.
• The June model incorporates data from Health Canada's Drug Analysis Service (DAS) for 2020. The December model used data from DAS for 2019 and applied these data to 2020.
• The June model has been updated with the surveillance data on apparent opioid-related deaths from April through September 2020.
• Comparisons between modelling results and the most recent surveillance data released in June 2021 should be made with caution. The model uses data from the March 2021 release, which may vary slightly from the data released in June.

Projections for 2021 (released December 2020)

The model's projections for April 2020 to June 2021 suggest that, under some scenarios, the opioid overdose crisis may resurge or surpass levels seen at the height of the opioid overdose crisis in late 2018.

Figure 1 text description: Observed and projected opioids overdose deaths, Canada, 2016 to 2021

• Scenario 1: Public health measures prevent 60% of deaths
• Scenario 2: Public health measures prevent 40% of deaths
• Scenario 3: Public health measures prevent 20% of deaths

Within each scenario, there can be varying levels of fentanyl.

• Dotted line: reflects when fentanyl in the illegal drug supply accounts for 80% of overdose deaths
• Dashed line: reflects when fentanyl in the illegal drug supply accounts for 85% of overdose deaths
• Solid line: reflects when fentanyl in the illegal drug supply accounts for 90% of overdose deaths
• Solid black line: represents the model output for the period from January 2016 to December 2019, which was calibrated to observed data shown by the grey bars. Comparing the real data from the past to the simulated deaths from the model shows us that the model is well set up (i.e. properly calibrated).

• Scenario 1 (purple lines): If public health measures continue to prevent 60% of opioid overdose deaths, as was estimated in October to March 2020 before the onset of the COVID-19 outbreak, then deaths for April through June 2021 will remain similar to those observed at the end of 2019.
• Scenario 2 (green lines): If public health measures prevent 40% of opioid overdose deaths (starting in April 2020), then quarterly deaths are projected to meet or surpass the levels seen at the height of the opioid overdose crisis in 2018.
• Scenario 3 (red lines): If public health measures prevent only 20% of opioid overdose deaths (starting in April 2020), then quarterly deaths are projected to be substantially higher than observed in any previous quarter.
• The level of fentanyl in the illegal drug supply also has an effect on the number of deaths, as shown by the dotted, dashed and full lines.

Notes

PHAC's model simulates opioid overdose deaths in a way that closely matches observed opioid overdose deaths from 2016 through early 2020.

Changes between the model released on October 26, 2020 and the new model released on December 16, 2020:

• The December model incorporates data from Health Canada's Drug Analysis Service (DAS) for 2019. The original October model used data from DAS for 2018 and applied these data to 2019.
• The December model has been updated with the statistics on apparent opioid-related deaths from January to March 2020.
• Comparisons between modelling results and the most recent surveillance data released in December 2020 should be made with caution due to changes in the way data are reported by some jurisdictions for the December 2020 release.

Canada's approach to the opioid overdose crisis

• Federal Actions on Opioids to date
• Responding to Canada's opioid overdose crisis