Modelling opioid-related deaths during the overdose crisis

The results of the model suggest that, under some scenarios, the number of opioid-related deaths through to December 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 December 2023

Time		Observed number of deaths based on March 2023 surveillance data release	Number of deaths estimated by the modelFootnote a	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 stays the same	Scenario 3: Health interventions prevent more deaths; level of fentanyl increases	Scenario 4: Health interventions prevent more deaths; level of fentanyl stays the same
2016	Jan - Mar	668	668	N/A	N/A	N/A	N/A
	Apr - June	636	565	N/A	N/A	N/A	N/A
	July - Sept	654	814	N/A	N/A	N/A	N/A
	Oct - Dec	862	958	N/A	N/A	N/A	N/A
2017	Jan - Mar	921	1068	N/A	N/A	N/A	N/A
	Apr - June	1021	1161	N/A	N/A	N/A	N/A
	July - Sept	1056	1167	N/A	N/A	N/A	N/A
	Oct - Dec	921	1122	N/A	N/A	N/A	N/A
2018	Jan - Mar	999	1185	N/A	N/A	N/A	N/A
	Apr - June	971	1160	N/A	N/A	N/A	N/A
	July - Sept	1100	1131	N/A	N/A	N/A	N/A
	Oct - Dec	1111	1088	N/A	N/A	N/A	N/A
2019	Jan - Mar	1043	1089	N/A	N/A	N/A	N/A
	Apr - June	1054	1022	N/A	N/A	N/A	N/A
	July - Sept	742	1021	N/A	N/A	N/A	N/A
	Oct - Dec	863	893	N/A	N/A	N/A	N/A
2020	Jan - Mar	1041	921	N/A	N/A	N/A	N/A
	Apr - June	1693	1249	N/A	N/A	N/A	N/A
	July - Sept	1793	1744	N/A	N/A	N/A	N/A
	Oct - Dec	1879	1870	N/A	N/A	N/A	N/A
2021	Jan - Mar	1912	2126	N/A	N/A	N/A	N/A
	Apr - June	1912	1927	N/A	N/A	N/A	N/A
	July - Sept	1977	1810	N/A	N/A	N/A	N/A
	Oct - Dec	2199	1872	N/A	N/A	N/A	N/A
2022	Jan - Mar	1934	2076	N/A	N/A	N/A	N/A
	Apr - June	1737	1903	N/A	N/A	N/A	N/A
	July - Sept	1678	1732	N/A	N/A	N/A	N/A
	Oct - Dec	1834Footnote b	N/A	1922	1858	1761	1712
2023	Jan - Mar	N/A	N/A	2311	2095	1765	1603
	Apr - June	N/A	N/A	2170	1969	1661	1509
	July - Sept	N/A	N/A	2054	1864	1575	1432
	Oct - Dec	N/A	N/A	2102	1908	1612	1466
N/A: not applicable
Footnote a Number of deaths estimated by the model (solid black line): represents the model output for the period from January 2016 to September 2022, which was calibrated to the observed data shown by the grey bars. Comparing the number of observed deaths from the past to the simulated number of deaths from the model shows us that the model is well set up. Return to footnote a referrer Footnote b Based on data released in June 2023. Return to footnote b referrer

Scenario 1 (solid purple line): If health interventions prevent the same proportion of opioid-related deaths as they did in September 2022 and the level of fentanyl and its analogues in the illegal drug supply increases after September 2022, then deaths may increase through to December 2023.

Scenario 2 (square hashed green line): If health interventions prevent the same proportion of opioid-related deaths as they did in September 2022, and the level of fentanyl and its analogues in the illegal drug supply remains the same as September 2022, then deaths through to December 2023 may be similar to levels observed in 2021 and early 2022.

Scenario 3 (dotted hashed red line): If health interventions prevent more opioid-related deaths after September 2022, and the level of fentanyl and its analogues in the illegal drug supply increases after September 2022, then deaths through to December 2023 may decrease.

Scenario 4 (rectangular hashed blue line): If health interventions prevent more opioid-related deaths after September 2022, and the level of fentanyl and its analogues in the illegal drug supply remains the same as September 2022, then deaths through to December 2023 may further decrease but not to levels seen before the onset of the COVID-19 pandemic.

Notes

PHAC developed a simulation model of opioid-related deaths using a method called system dynamics modelling.

The simulations presented in the model reflect a combination of:

• model structure
• assumptions about what may happen to the opioid drug supply and the proportion of opioid-related deaths that are prevented through health interventions, such as naloxone or supervised consumption sites

In the model, people using opioids medically or non-medically can die from an opioid-related overdose at different rates depending on the type of opioid use. Mortality rates for non-medical opioid use (the use of opioids without a prescription or in ways other than medically prescribed) are also affected by:

• the time of year
• the proportion of fentanyl and its analogues in the illegal drug supply
• the proportion of deaths prevented by health interventions designed to reduce opioid-related deaths

Values used in the model were identified from:

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

The 4 projected scenarios each had the following assumptions:

Scenario 1: Health interventions prevent 20% of opioid-related deaths and fentanyl and its analogues make up 80% of the opioid drug supply.

Scenario 2: Health interventions prevent 20% of opioid-related deaths and fentanyl and its analogues make up 71% of the opioid drug supply.

Scenario 3: Health interventions prevent 40% of opioid-related deaths and fentanyl and its analogues make up 80% of the opioid drug supply.

Scenario 4: Health interventions prevent 40% of opioid-related deaths and fentanyl and its analogues make up 71% of the opioid drug supply.

Ongoing considerations

As opioid use and related harms have changed significantly over the past few years, this model will be published twice a year. Changes between the previous model released in December 2022 and the current model released in June 2023 include:

• the measure of toxicity in the drug supply has been updated to represent the proportion of fentanyl and its analogues in the opioid drug supply
• other parameter values have been updated as more peer-reviewed literature and data has become available
• the model has been updated with the surveillance data on apparent opioid-related deaths through September 2022

PHAC will update the model as we learn more about the opioids overdose crisis.

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 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

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 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.

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 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.

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 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.

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 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.