Abstract
Abstract
Objectives
Our study investigated the age-adjusted incidence rates of non-fatal overdoses by HIV status and sex, and examined trends over time.
Design
We used data from the Comparative Outcomes and Service Utilization Trends study, a population-based cohort study that includes clinical and administrative health data on virtually all people with HIV (PWH) and a 10% random sample of people without HIV in the province.
Setting
British Columbia, Canada.
Participants
Between April 2012 and March 2020, 11 050 PWH (81.8% male) and 473 952 people without HIV (50.3% male) who were 19 years and older contributed 68 035 and 3 285 824 person years (PY) of follow-up, respectively.
Outcome measures
The primary outcome was age-adjusted incidence rates of non-fatal overdose events stratified by sex and HIV status. Trends over time were also assessed.
Results
Age-adjusted non-fatal overdose incidence rates among males with and without HIV were 36.4 and 3.12 per 1000 PY, respectively (incidence rate ratio (IRR) = 11.7, 95% CI 10.9 to 12.5). For females with and without HIV, the age-adjusted incidence rates were 61.4 and 2.33 per 1000 PY, respectively (IRR=26.3, 95% CI 24.0 to 28.7). Between 2013 and 2019 (calendar years with full-year data), the age-adjusted non-fatal overdose rate increased significantly among males and females without HIV but not among PWH.
Conclusions
We observed a significantly higher non-fatal overdose rate among PWH compared to people without HIV. The rate was highest among females with HIV. These findings underline the need for policies and programmes oriented towards PWH to mitigate overdoses, especially for females.
Keywords: HIV & AIDS, Substance misuse, EPIDEMIOLOGY, PUBLIC HEALTH, Sexual and Gender Minorities
STRENGTHS AND LIMITATIONS OF THIS STUDY.
This population-based cohort study includes data on nearly all known people with HIV and 10% of the general population without HIV in the province of British Columbia, Canada.
The study period covers 8 years, from 2012 until 2020.
Only non-fatal overdoses which were attended to at an emergency department, by a healthcare practitioner in another setting, or led to hospitalisation were captured.
Substance use disorder (SUD) was defined based on SUD-related healthcare utilisation in the 5 years before baseline, and therefore SUD cases not leading to health care use were not captured.
Introduction
The drug toxicity crisis continues to devastate communities worldwide; North America is one of the most impacted regions in the world.1 The crude opioid toxicity death rate in Canada was 21.5 per 100 000 population in 2023, an almost threefold increase compared to 2016.2 Within Canada, overdose rates are highest in the province of British Columbia (BC): between 2015 and 2024, the number of unregulated drug deaths nearly quadrupled, claiming the lives of 2271 British Columbians in 2024—a crude rate of 39.9 per 100 000 population.3 This increase is largely attributable to an increasingly toxic unregulated drug supply, with fentanyl or its analogues involved in more than 80% of all unregulated drug deaths since 2018, and an increasing involvement of benzodiazepines.3 4
People with HIV (PWH) are uniquely affected by the toxic drug crisis. Sociostructural factors, including access to healthcare, income inequality, and racial/ethnic segregation, are associated with both risk of HIV infection and higher prevalence of substance use.5 Injection drug use is a major route of HIV transmission; in 2022, 13.1% of new HIV infections among males and 36.1% among females in Canada were attributed to injection drug use.6 We have previously shown that overdose and drug-related deaths constitute a large proportion of overall deaths among PWH in BC, reducing gains in life expectancy achieved since the advent of modern antiretroviral therapy (ART) and the subsequent public health response.7 8
A substantial proportion of overdoses are non-fatal.9 In a study among attendees of a supervised consumption site, 59% of participants reported ever having experienced a non-fatal overdose.10 The risk of experiencing a subsequent fatal or non-fatal overdose in the year following a non-fatal overdose is substantially elevated.11 While overdose rates in the general population are typically higher among males, females with HIV may face a disproportionate burden of risk possibly due to additional gender-related barriers at the interpersonal, sociocultural, and structural level.12,14 This sex-based difference has not been well examined in population-based studies of non-fatal overdose. Insight into the occurrence of non-fatal overdoses, identifying populations at risk, and understanding whether this risk differs by sex especially among PWH is therefore essential to adequately inform policies and design equitable interventions.
To this end, we assessed rates and trends of non-fatal overdoses among males and females with and without HIV, between April 2012 and March 2020. We hypothesised that the incidence of non-fatal overdoses would be higher among PWH and would increase during this period in both people with and without HIV. Additionally, we hypothesised that females, particularly those with HIV, would experience higher rates of non-fatal overdose compared with males.
Methods
Design and setting
The Comparative Outcomes And Service Utilization Trends (COAST) study is a population-based observational cohort study in BC, Canada.15 The COAST study consists of a linkage between the BC Centre for Excellence in HIV/AIDS’s Drug Treatment Program registry16 and Population Data BC’s data holdings. Population Data BC is a multidisciplinary institution that holds a provincial repository of individual-level administrative health data for all 5.7 million BC residents.17 COAST includes data on all diagnosed PWH aged ≥19 years in BC, between 1992 and 2020, registered in the Drug Treatment Program. The Drug Treatment Program centrally manages the distribution of ART to virtually all medically eligible PWH in the province. Through a linkage to laboratory data, people with a detectable HIV plasma load (and no ART exposure) are also included in the Drug Treatment Program registry. In addition, COAST includes a 10% random sample from the general BC population without HIV, aged ≥19 years from 1992 to 2020, identified through Population Data BC. Further details on the COAST cohort are described elsewhere.15 The following datasets were used in this study: National Ambulatory Care Reporting System (emergency department visits), Discharge Abstract Database (hospital separations), Medical Service Plan payment information file (healthcare practitioner billings), Vital Statistics—Deaths18 and Consolidation File (demographics). Access to data provided by the Data Stewards is subject to approval but can be requested for research projects through the Data Stewards or their designated service providers. Further information regarding these datasets can be obtained by visiting the Population Data BC project webpage at: https://my.popdata.bc.ca/project_listings/18-223. All inferences, opinions and conclusions drawn in this publication are those of the author(s), and do not reflect the opinions or policies of the Data Steward(s).
Participants
In this analysis, we included PWH with an existing linkage to Population Data BC and the 10% random general population sample without HIV with at least 1 day of follow-up between 1 April 2012 and 31 March 2020. Despite having data available from 1992, we restricted our study start date to 1 April 2012, to ensure consistent capture of overdose-related emergency department visits through the National Ambulatory Care Reporting System, which is only available from 2012 onwards in BC. We defined baseline for PWH as the latest of the following dates: first record of HIV-positive status (based on either registration in the Drug Treatment Program registry or first HIV plasma viral load measurement), their 19th birthday, or 1 April 2012. For people without HIV, we used the later of their 19th birthday or 1 April 2012, to define baseline. Individuals were followed until death, end of follow-up (defined as the later of the last date of registration in BC’s universal health insurance programme or the last healthcare contact date) or 31 March 2020, whichever came first.
Non-fatal overdoses
We identified non-fatal overdose events from psychoactive substances, including opioids and stimulants, that resulted in a healthcare practitioner encounter, emergency department visit or hospitalisations. Operational definitions for overdose, opioid overdose and stimulant overdose used International Classification of Diseases diagnostic codes and were modified from previously published definitions (table 1).19 20 We combined multiple overdose-related healthcare encounters on the same date as one non-fatal overdose.
Table 1. Case definitions of non-fatal overdoses.
| Data source | Definition | |
|---|---|---|
| Any non-fatal overdose | NACRS | Any of the following CED-DxS38 ICD-10-CA codes in the discharge diagnosis field: T40.1, T40.5, T40.6, T40.7, T40.9, T42.4, T43.9, or T50.9. |
| DAD* | Any of the diagnosis fields ‘M’ = most responsible diagnosis, ‘1’ = pre-admission comorbidity, ‘2’ = postadmission comorbidity, ‘W’, ‘X’ or ‘Y’ = service transfer diagnosis containing any of the following ICD-10-CA codes: T40*, T42.3, T42.4, T42.6, T42.7, T43.6, T43.8 or T43.9. | |
| MSP | First/main diagnosis field containing any of the following ICD-9/9-CM codes: 965, 965.0, 965.00, 965.01, 965.02, 965.09, 967*, 969*, 970*, 977*, E850.0, E850.1, E850.2, E852.5, E852.8, E852.9, E853.2, E854.1, E854.2, E854.3 or E854.8. | |
| Opioid overdose | NACRS | Discharge diagnosis field containing ICD-10-CA code T401 or T406. |
| DAD†‡ | Any of the diagnosis fields ‘M’, ‘1’, ‘2’, ‘W’, ‘X’ or ‘Y’ containing any of the following ICD-10-CA codes: T40.0, T40.1, T40.2*, T40.3, T40.4* or T40.6. | |
| MSP | First/main diagnosis field containing any of the following ICD-9/9-CM codes: 965.0, 965.00, 965.01, 965.02, 965.09, E850.0, E850.1 or E850.2. | |
| Stimulant overdose | NACRS | Discharge diagnosis field containing ICD-10-CA code T405. |
| DAD | Any of the diagnosis fields ‘M’, ‘1’, ‘2’, ‘W’, ‘X’ or ‘Y’ ICD-10-CA code T40.5 or T43.6. | |
| MSP | First/main diagnosis field containing any of the following ICD-9/9-CM codes: 969.7, 969.70, 969.79, 969.72, 969.73 or E854.2. |
The asterisk (*) after an ICD code indicates: any code starting with that value.
Adapted from the definition as published by Public Health Agency Canada (PHAC)20 to exclude alcohol poisoning.
Adapted from the definition as published by PHAC to include T40.2* instead of only T40.20, T40.21, T40.22, T40.23 and T40.28 and to include T40.4* instead of only T40.40, T40.41 and T40.48.
Adapted from the definition as published by MacDougall et al19 to include diagnosis fields ‘M’, ‘1’, ‘2’, ‘W’, ‘X’, ‘Y’ instead of only the first diagnosis field in the DAD.
CED-DxS, Canadian Emergency Department Diagnosis Shortlist; DAD, Discharge Abstracts Database; ICD-10-CA, International Statistical Classification of Diseases and Related Conditions, 10th edition, Canada; ICD-9-CM, International Statistical Classification of Diseases and Related Conditions, 9th Edition, Clinical Modification; MSP, Medical Service Plan Payment Information File; NACRS, National Ambulatory Care Reporting System.
Sociodemographic factors and substance use
We derived age and sex from the Population Data BC Consolidation File, which contains demographic and geographic information on BC residents. We used area income quintile, adjusted for household size, as a measure of socioeconomic status (area-level income). This measure is obtained from census summary data linked to individuals’ geographical residence at dissemination area-level, which is the smallest geographical unit (400–700 people) for which all census data are disseminated. Dissemination areas within a census region are ranked by income level and grouped into five quintiles, from the lowest (Q1) to the highest (Q5). Individuals are then assigned an area-level income quintile based on where they live. Residency in the Downtown Eastside, a neighbourhood in Vancouver’s inner city characterised by complex problems including a high prevalence of substance use and housing instability, was characterised by the first three digits of the postal code being V6A.21 Among individuals with 5 years of available data prior to baseline,22 substance use disorder (SUD) was identified through a combination of three BC algorithms all based on healthcare utilisation patterns (online supplemental table 1).23,26
Statistical analyses
We used χ2 and Kruskal-Wallis tests for the descriptive analyses. For each year of follow-up, we calculated age-standardised rates of non-fatal overdose events per 1000 person years (PY) for people with and without HIV, stratified by sex, using the 2016 census population of Canada as reference. We compared rates using rate ratios. We graphically presented yearly rates of all non-fatal overdoses, as well as separately for opioids and stimulants (including calendar years with full-year data only). To test whether the trends in yearly non-fatal overdose rates were significant, we used two approaches. First, we assessed whether the overdose rates showed a statistically significant linear trend using a test for trend across calendar years. Second, we used a permutation test for joinpoint regression to identify statistically significant changes in the trend over time. For variables with missing values, we incorporated a missing category. Individuals with missing data for sex were excluded from the incidence calculations. In the results section and tables, small cell sizes (<5) are presented as <5 to mitigate any risk of re-identifiability. To mask these numbers, the second smallest cell size is presented as a range.27 We conducted all our analyses using SAS V.9.4 (SAS Institute Inc., Cary, NC, USA). All tests were two-sided and p<0.05 was considered statistically significant.
Patient and public involvement
Community members are actively involved in the COAST study at large; an advisory board is currently being formed and there are several completed and ongoing community-co-led projects.15 Patients and the public were not involved in the design, conduct, reporting or dissemination plans of this analysis.
Results
Study population
The selection process for the study population is outlined in online supplemental figure 1. In total, we included 11 050 PWH of whom 82.3% were male and 473 952 people without HIV of whom 50.3% were male in this analysis (table 2). These groups contributed 68 035 and 3 285 824 PY of follow-up, respectively. A total of 1221 PWH and 5747 people without HIV experienced at least one non-fatal overdose during follow-up. At baseline, people with and without HIV who experienced a non-fatal overdose were significantly younger than those who did not. Specifically, among PWH, the median baseline age was 42.9 years (Q1–Q3: 35.2–49.0) for those who had a non-fatal overdose and 46.8 years (Q1–Q3: 38.0–54.1) for those who never experienced a non-fatal overdose. Similarly, the median baseline age for people without HIV who experienced a non-fatal overdose was 34.4 years (Q1–Q3: 23.2–51.2) while those who never experienced an overdose were 42.1 years (Q1–Q3: 26.3–57.9). PWH who experienced a non-fatal overdose were more likely to be female (29.2–6% vs 16.8%), from a low-income area (57.4% vs 36.3% resided in an area with the lowest quintile income) and from the Downtown Eastside neighbourhood in Vancouver (26.7% vs 11.3 %) compared with those who never experienced an overdose.
Table 2. Baseline characteristics of people with and without HIV included in the analysis, stratified by whether they experienced at least one non-fatal overdose between 1 April 2012 and 31 March 2020.
| PWH N (%)* or median (Q1, Q3) N=11 050 |
People without HIV N (%)* or median (Q1, Q3) N=473 952 |
|||
|---|---|---|---|---|
| No non-fatal overdose N=9829 |
≥1 non-fatal overdose N=1221 |
No non-fatal overdose N=468 205 |
≥1 non-fatal overdose N=5747 |
|
| Age, years | 46.8 (38.0, 54.1) | 42.9 (35.2, 49.0)† | 42.1 (26.3, 57.9) | 34.4 (23.2, 51.2)† |
| Sex | ||||
| Female | 1650 (16.8%) | 357–362 (29.2–6%)† | 232 759 (49.7%) | 2648 (46.1%)† |
| Male | 8179 (83.2%) | 859 (70.4%)† | 235 427 (50.3%) | 3099 (53.9%)† |
| Missing | 0 | <5 | 19 | 0 |
| Area-level income | ||||
| Q1 (lowest) | 3189 (36.3%) | 651 (57.4%)† | 77 751 (20.3%) | 1613 (30.8%)† |
| Q2 | 1721 (19.6%) | 141 (12.4%)† | 77 218 (20.1%) | 1124 (21.4%)† |
| Q3 | 1627 (18.5%) | 168 (14.8%)† | 76 921 (20.1%) | 949 (18.1%)† |
| Q4 | 1262 (14.4%) | 109 (9.6%)† | 77 143 (20.1%) | 818 (15.6%)† |
| Q5 (highest) | 982 (11.2%) | 65 (5.7%)* | 74 553 (19.4%) | 738 (14.1%)* |
| Unknown/missing | 1048 | 87 | 84 619 | 505 |
| Residence | ||||
| Vancouver’s Downtown Eastside | 1046 (11.3%) | 321 (26.7%)† | 5455 (1.4%) | 179 (3.3%)† |
| Other | 8181 (88.7%) | 883 (72.%)† | 395 165 (98.6%) | 5275 (96.7%)† |
| Unknown/missing | 602 | 17 | 67 585 | 293 |
| Substance use disorder | ||||
| Yes | 743 (9.0%) | 212 (18.8%)† | 6948 (1.9%) | 898 (17.4%)† |
| No | 7552 (91.0%) | 914 (81.2%)† | 356 627 (98.1%) | 4242 (82.6%)† |
| Unknown/missing‡ | 1534 | 95 | 104 630 | 607 |
| Follow-up, years | 8.0 (4.1, 8.0) | 8.0 (6.1, 8.0)† | 8.0 (7.3, 8.0) | 8.0 (7.7, 8.0)† |
| Number of non-fatal overdoses | ||||
| 0 | NA | – | NA | – |
| 1 | 696 (57.0%) | 4329 (75.3%) | ||
| 2 | 223 (18.3%) | 763 (13.3%) | ||
| 3 or more | 302 (24.7%) | 655 (11.4%) | ||
Small cell sizes (<5) are presented as <5 to mitigate any risk of re-identifiability and to mask these numbers the second smallest cell size is presented as a range.
Percentages are calculated excluding missing values.
Statistically significant; p<0.05.
Individuals without a 5-year lookback window (5 years historic administrative data prior to baseline) were classified as unknown.
NA, non-applicable.
Additionally, participants who experienced at least one non-fatal overdose were more likely to have an indication of SUD at baseline compared with those who never experienced an overdose. Among PWH, we found baseline SUD in 18.8% of those who experienced a non-fatal overdose compared with 9.0% of those who never had an overdose. For people without HIV, we observed baseline SUD in 17.4% of those who had an overdose compared with 1.9% of those who did not. Lastly, among those who ever experienced a non-fatal overdose, PWH were nearly twice as likely to experience multiple overdoses than people without HIV (43.0% of PWH vs 24.7% of people without HIV) (table 2).
Overdose incidence
PWH had a significantly higher age-adjusted incidence rate of non-fatal overdose compared with those without HIV, across both sexes. In particular, the age-adjusted incidence rate was 36.4 per 1000 PY (95% CI 34.0 to 39.0) among males with HIV and 3.12 per 1000 PY (95% CI 3.03 to 3.20) among males without HIV. The corresponding incidence rate ratio (IRR) was 11.7 (95% CI 10.9 to 12.5). Likewise, among females, the age-adjusted incidence rate was 61.4 per 1000 PY (95% CI 55.9 to 66.8) for those with HIV and 2.33 per 1000 PY (95% CI 2.26 to 2.40) for those without HIV. The IRR comparing females with and without HIV was 26.3 (95% CI 24.0 to 28.7). We observed a similar pattern when non-fatal overdoses were classified by substance type, including opioid and stimulant overdose (table 3).
Table 3. Age-adjusted incidence rates (per 1000 person years) for any non-fatal overdose, opioid overdose and stimulant overdose among males and females with and without HIV.
| Males with HIV aIR (95% CI) |
Males without HIV aIR (95% CI) |
IRR | Females with HIV aIR (95% CI) |
Females without HIV aIR (95% CI) |
IRR | |
|---|---|---|---|---|---|---|
| Any non-fatal overdose | 36.4 (34.0 to 39.0) | 3.12 (3.03 to 3.20) | 11.7 (10.9–12.5) | 61.4 (55.9 to 66.8) | 2.33 (2.26 to 2.40) | 26.3 (24.0–28.7) |
| Opioid overdose | 16.5 (15.1 to 18.0) | 1.00 (0.96 to 1.05) | 16.5 (15.0–17.9) | 28.3 (25.1 to 31.4) | 0.50 (0.46 to 0.53) | 56.7 (50.4–63.1) |
| Stimulant overdose | 2.11 (1.48 to 2.73) | 0.16 (0.14 to 0.18) | 13.3 (9.4–17.2) | 4.02 (2.58 to 5.47) | 0.08 (0.06 to 0.09) | 51.6 (33.0–70.9) |
aIR, age-adjusted incidence rate; IRR, incidence rate ratio.
Incidence trends
The age-adjusted incidence rate for non-fatal overdose increased significantly between 2013 and 2019 (calendar years with full-year data) among males (p-value for trend test (ptrend) < 0.05) and females (ptrend<0.05) without HIV (figure 1a). This was not the case for males and females with HIV. When examining by substance type, we observed a significant increase in age-adjusted incidence rate of opioid overdose among males without HIV (ptrend<0.05), but not among females without HIV or males and females with HIV (figure 1b). For stimulant overdoses, the age-adjusted incidence rate did not change significantly in any of the groups (figure 1c). There was an increase in age-adjusted incidence rate for any non-fatal overdose as well as opioid overdoses up until 2016–2017, followed by a decrease in males and females with HIV. However, this pattern was statistically significant only for males with HIV (p-value for joinpoint regression (pjoinpoint) < 0.05) in the age-adjusted incidence rates for any non-fatal overdose (figure 1a). A similar pattern was observed among males with HIV for opioid overdoses, although not statistically significant (pjoinpoint=0.09) (figure 1b).
Figure 1. Age-adjusted rates (per 1000 person years) of any non-fatal overdoses, opioid overdoses and stimulant overdoses among males and females with and without HIV. (a) Age-adjusted non-fatal overdose rates among males and females with and without HIV. (b) Age-adjusted opioid overdose rates among males and females with and without HIV. (c) Age-adjusted stimulant overdose rates among males and females with and without HIV.
Discussion
We assessed the incidence of non-fatal overdose recorded in healthcare settings, and compared patterns among PWH and a random sample of the BC general population without HIV. The rates for any non-fatal overdose, opioid overdose and stimulant overdose were significantly higher among both males and females with HIV compared with their counterparts without HIV. However, between 2013 and 2019, the age-adjusted non-fatal overdose rate increased significantly among people without HIV but not among PWH. The non-fatal overdose rate was highest among females with HIV. These findings are in line with and extend previous studies in BC, reporting on a higher rate of fatal overdoses among PWH7 8 and reflecting a higher prevalence of drug use among PWH.28
The higher non-fatal overdose incidence among females compared with males with HIV that we observed contrasts with the higher overdose incidence among males in the general population.19 29 This is likely a consequence of the synergism between the HIV and substance use epidemics: vulnerable populations, which include women in urban poor, rural and racialised communities, are more likely to be affected by the intertwined epidemics of HIV, substance abuse and violence.30 31 This could be a consequence of sociostructural factors, such as stigma, discrimination and housing instability30 and reflected by a higher prevalence of substance use among women than men with HIV. Although our data only consist of administrative health information and do not include direct measures of these sociostructural factors, evidence from previous literature suggests they likely contribute to the elevated overdose risk among women with HIV. For instance, in 2022, injection drug use was the exposure category of 13.3% of infections among males and 36.1% among females in Canada.6 32 Women who use drugs may encounter gender-based barriers to accessing care—including harm reduction and safer drug consumption services.31 Furthermore, those who inject drugs are often dependent on others for access to and preparation of drugs for injection and are frequently injected by their partners.31 33 Together, these factors likely increase both substance use and overdose risk among women with HIV.12,1431 34
Although the incidence of non-fatal overdose was consistently higher among PWH than people without HIV, the incidence persistently increased over calendar time among males and females without HIV only. In contrast, the non-fatal overdose incidence rate among PWH peaked in 2016/2017, decreasing thereafter. This pattern was statistically significant among males with HIV for any non-fatal overdoses and also observed for opioid overdoses (although not statistically significant), whereas it was not discernible for stimulant overdoses. The decrease after 2016 may be a consequence of policy changes and measures coming into place around that time, when a province-wide public health emergency was declared in BC.35 Although a province-wide take-home naloxone programme began in 2012, it was expanded in 2016 to include more distribution sites and to provide naloxone to individuals likely to witness an opioid overdose, in addition to those at risk of experiencing an overdose. As a result, the number of kits distributed increased sharply from approximately 6000 in 2015 to 1 40 000 in 2017 and continued to increase to nearly 400 000 in 2020.36 PWH may have been uniquely positioned to access naloxone kits and training through HIV clinics, AIDS service organisations and other community-based organisations they were engaged with. This may have led to a higher proportion of overdoses among PWH being reversed and not needing further intervention and not captured in our study.
A strength of this study is that it provides insight into the occurrence of non-fatal overdoses at the level of the population by identifying subgroups who are potentially most at risk for non-fatal overdose. Limitations are related to the use of administrative data: these data are collected for financial and administrative purposes, therefore are prone to incompleteness and lack in detail and accuracy; only those events that lead to a healthcare encounter will be captured.37 While we can identify people who: (1) attend an emergency department, (2) were admitted to a hospital or (3) attended by a healthcare practitioner in other settings following a non-fatal overdose, we do not have data on non-fatal overdoses that are bystander-attended or attended to only by a paramedic with no subsequent transport to an emergency department. Although provincial reports suggest that a considerable number of overdoses fall within this category, to our knowledge, there is no evidence to suggest that these events differ by HIV status, making it less likely for the under-capture of such events to explain the observed differences in our study. Another limitation is the exclusion of participants who had no follow-up data after the study start date. Most of these individuals had either died or were lost to follow-up prior to the start date. Although data was available from 1992, we restricted our start date to 1 April 2012, to ensure consistent capture of overdose-related emergency department visits through the National Ambulatory Care Reporting System, which is only available from 2012 in BC. While this exclusion was necessary, it may introduce selection bias if individuals excluded differ systematically from those included in our analyses. Furthermore, there is a potential for misclassification, particularly in the classification of overdoses by drug involved. Similarly, information on SUD is limited by the shortcomings of available algorithms: any algorithm applied to administrative healthcare records relies on healthcare utilisation patterns; thus, people who do not seek healthcare will not be captured.23 25 26 However, since we used the same data sources for people with and without HIV, we assume misclassification bias may have been similar for each group. Also, our dataset did not include information on stigma and discrimination; therefore, we were unable to adjust for these factors in our analyses. Lastly, we did not have data on gender identity and relied on biological sex.
Conclusions
We observed an alarmingly high non-fatal overdose rate among PWH, with the highest rate among females with HIV. These findings expand earlier reports of a higher rate of fatal overdoses among PWH and support a need for the expansion of harm reduction and overdose prevention services for PWH potentially at risk for overdose, especially for females. Further research is needed to better understand the sociostructural and behavioural factors that contribute to increased risk of non-fatal overdose among people who use drugs, particularly those with HIV, and to assess how this risk differs by sex.
Supplementary material
Acknowledgements
The authors would like to thank the COAST study participants, BC Cancer Agency, BC Centre for Excellence in HIV/AIDS, BC Ministry of Health, BC Vital Statistics Agency, PharmaNet and the Institutional Data Stewards for granting access to the data, and Population Data BC, for facilitating the data linkage process.
Footnotes
Funding: COAST is supported by the Canadian Institutes of Health Research (130419 and 143342) and the BC Centre for Excellence in HIV/AIDS. This project is supported by CIHR (W12-179952) and NIH-NIDA (R21DA057058 (PI: Hogg)). KWK is supported by CIHR (HIV-181935 and IF8-190450), Michael Smith Health Research BC (RT-2022-2559), and a Merck/CTN Postdoctoral Fellow Award. MM is supported by NIH-NIDA (T32DA031099 (PI: Martins, Hasin) and R36DA061635 (PI: Marziali)). JM is supported by grants paid to his institution by the BC Ministry of Health, Health Canada, Public Health Agency of Canada, Vancouver Coastal Health, Vancouver General Hospital Foundation and Genome BC. The funding sources were not involved in study design, data collection, analysis, interpretation of study data, writing of the report or the decision to submit the article for publication.
Prepublication history and additional supplemental material for this paper are available online. To view these files, please visit the journal online (https://doi.org/10.1136/bmjopen-2025-103778).
Provenance and peer review: Not commissioned; externally peer reviewed.
Patient consent for publication: Not applicable.
Ethics approval: The study complies with the BC Freedom of Information and Protection of Privacy Act and did not require informed consent as it is conducted retrospectively for research and statistical purposes only using anonymized data. The COAST study has been approved by the University of BC/Providence Health Care Research Ethics Board and Simon Fraser University Office of Research Ethics (#H09-02905).
Data availability free text: Access to data provided by the Data Stewards is subject to approval but can be requested for research projects through the Data Stewards or their designated service providers. The following data sets were used in this study: National Ambulatory Care Reporting System, Discharge Abstract Database (Hospital Separations), Medical Services Plan, Vital Events and Statistics—Deaths, and Consolidation File. Further information regarding these datasets can be found on the Population Data BC project webpage at: https://my.popdata.bc.ca/project_listings/18-223/. All inferences, opinions, and conclusions drawn in this publication are those of the author(s), and do not reflect the opinions or policies of the Data Steward(s).
Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting or dissemination plans of this research.
Data availability statement
Data may be obtained from a third party and are not publicly available.
References
- 1.Krausz RM, Westenberg JN, Ziafat K. The opioid overdose crisis as a global health challenge. Curr Opin Psychiatry. 2021;34:405–12. doi: 10.1097/YCO.0000000000000712. [DOI] [PubMed] [Google Scholar]
- 2.Canada PHA of Key findings: opioid- and stimulant-related harms in Canada — Canada.ca. 2019. [19-Mar-2025]. https://health-infobase.canada.ca/substance-related-harms/opioids-stimulants/ Available. Accessed.
- 3.BC Coroners Service Statistical reports on deaths in British Columbia - province of British Columbia. 2025. [19-Mar-2025]. https://www2.gov.bc.ca/gov/content/life-events/death/coroners-service/statistical-reports Available. Accessed.
- 4.BC Coroners Service Illicit drug toxicity type of drug data. 2023. [19-Mar-2025]. https://www2.gov.bc.ca/assets/gov/birth-adoption-death-marriage-and-divorce/deaths/coroners-service/statistical/illicit-drug-type.pdf Available. Accessed.
- 5.Perlman DC, Jordan AE. The Syndemic of Opioid Misuse, Overdose, HCV, and HIV: Structural-Level Causes and Interventions. Curr HIV/AIDS Rep . 2018;15:96–112. doi: 10.1007/s11904-018-0390-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Canada PHA of HIV in Canada: 2022 surveillance highlights. 2023. [4-Feb-2025]. https://www.canada.ca/en/public-health/services/publications/diseases-conditions/hiv-2022-surveillance-highlights.html Available. Accessed.
- 7.Salters KA, Parent S, Nicholson V, et al. The opioid crisis is driving mortality among under-served people living with HIV in British Columbia, Canada. BMC Public Health. 2021;21:680. doi: 10.1186/s12889-021-10714-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.St-Jean M, Dong X, Tafessu H, et al. Overdose mortality is reducing the gains in life expectancy of antiretroviral-treated people living with HIV in British Columbia, Canada. Int J Drug Policy. 2021;96:103195. doi: 10.1016/j.drugpo.2021.103195. [DOI] [PubMed] [Google Scholar]
- 9.Marziali ME, Kooij KW, Zhang W, et al. Fatal overdoses among people with HIV experiencing nonfatal overdoses in British Columbia, Canada. AIDS. 2025;39:328–30. doi: 10.1097/QAD.0000000000004039. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Milloy M-JS, Kerr T, Mathias R, et al. Non-fatal overdose among a cohort of active injection drug users recruited from a supervised injection facility. Am J Drug Alcohol Abuse. 2008;34:499–509. doi: 10.1080/00952990802122457. [DOI] [PubMed] [Google Scholar]
- 11.McLeod KE, Slaunwhite AK, Zhao B, et al. Comparing mortality and healthcare utilization in the year following a paramedic-attended non-fatal overdose among people who were and were not transported to hospital: A prospective cohort study using linked administrative health data. Drug Alcohol Depend. 2021;218:108381. doi: 10.1016/j.drugalcdep.2020.108381. [DOI] [PubMed] [Google Scholar]
- 12.Boyd J, Collins AB, Mayer S, et al. Gendered violence and overdose prevention sites: a rapid ethnographic study during an overdose epidemic in Vancouver, Canada. Addiction . 2018;113:2261–70. doi: 10.1111/add.14417. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Moallef S, Genberg BL, Hayashi K, et al. Day-to-day impact of COVID-19 and other factors associated with risk of nonfatal overdose among people who use unregulated drugs in five cities in the United States and Canada. Drug Alcohol Depend. 2022;241:109633. doi: 10.1016/j.drugalcdep.2022.109633. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Marziali ME, Kooij KW, Martins SS, et al. Removing barriers to care: The importance of addressing nonfatal overdoses among women living with HIV in British Columbia, Canada. Can J Public Health . 2025;116:422–6. doi: 10.17269/s41997-025-01026-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Budu MO, Kooij KW, Heath K, et al. Cohort Profile Update: Reflecting back and looking ahead: Updating the Comparative Outcomes and Service Utilization Trends (COAST) Study to include 28 years of linked data from people with and without HIV in British Columbia, Canada. Int J Popul Data Sci . 2025;10:2496. doi: 10.23889/ijpds.v10i1.2496. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Drug Treatment Program | BC centre for excellence in HIV/AIDS. [6-Sep-2024]. https://www.bccfe.ca/drug-treatment-program Available. Accessed.
- 17.Population data BC. 2023. [3-Feb-2025]. https://www.popdata.bc.ca/ Available. Accessed.
- 18.British Columbia Ministry of Health Vital events deaths. V2. 2021
- 19.MacDougall L, Smolina K, Otterstatter M, et al. Development and characteristics of the Provincial Overdose Cohort in British Columbia, Canada. PLoS One. 2019;14:e0210129. doi: 10.1371/journal.pone.0210129. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Public Health Agency of Canada . Ottawa: 2021. Substance-related poisonings and homelessness in Canada: a descriptive analysis of hospitalization data. [Google Scholar]
- 21.BC community health data. [19-Dec-2024]. http://communityhealth.phsa.ca/CHSAHealthProfiles/CHSAHealthReport/Downtown%20Eastside Available. Accessed.
- 22.Nanditha NGA, Dong X, McLinden T, et al. The impact of lookback windows on the prevalence and incidence of chronic diseases among people living with HIV: an exploration in administrative health data in Canada. BMC Med Res Methodol. 2022;22:1. doi: 10.1186/s12874-021-01448-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Janjua NZ, Islam N, Kuo M, et al. Identifying injection drug use and estimating population size of people who inject drugs using healthcare administrative datasets. Int J Drug Policy. 2018;55:31–9. doi: 10.1016/j.drugpo.2018.02.001. [DOI] [PubMed] [Google Scholar]
- 24.Greenwald ZR, Werb D, Feld JJ, et al. Validation of case-ascertainment algorithms using health administrative data to identify people who inject drugs in Ontario, Canada. J Clin Epidemiol. 2024;170:111332. doi: 10.1016/j.jclinepi.2024.111332. [DOI] [PubMed] [Google Scholar]
- 25.Homayra F, Pearce LA, Wang L, et al. Cohort profile: The provincial substance use disorder cohort in British Columbia, Canada. Int J Epidemiol. 2021;49:1776. doi: 10.1093/ije/dyaa150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.BC Ministry of Health Substance-use-disorder.pdf. BC Chronic Dis. Regist. Case Defin. 2024. [28-Nov-2024]. http://www.bccdc.ca/resource-gallery/Documents/Chronic-Disease-Dashboard/substance-use-disorder.pdf Available. Accessed.
- 27.Working in the SRE - masking small cell sizes. [3-Nov-2025]. https://my.popdata.bc.ca/html/SRE/working/cell_sizes.html Available. Accessed.
- 28.Shiau S, Arpadi SM, Yin MT, et al. Patterns of drug use and HIV infection among adults in a nationally representative sample. Addict Behav. 2017;68:39–44. doi: 10.1016/j.addbeh.2017.01.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.BC Coroners Service Unregulated drug deaths - BC. Unregulated drug deaths death rate 100000 popul. 2014-2024. 2024. [28-Nov-2024]. https://app.powerbi.com/view?r=eyJrIjoiZThmOTkxMzgtZWUzNS00ODk1LWJiZjItYzMyMTFjNmY0MzJiIiwidCI6IjZmZGI1MjAwLTNkMGQtNGE4YS1iMDM2LWQzNjg1ZTM1OWFkYyJ9 Available. Accessed.
- 30.Meyer JP, Springer SA, Altice FL. Substance Abuse, Violence, and HIV in Women: A Literature Review of the Syndemic. J Womens Health (Larchmt) 2011;20:991–1006. doi: 10.1089/jwh.2010.2328. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Azim T, Bontell I, Strathdee SA. Women, drugs and HIV. Int J Drug Policy. 2015;26 Suppl 1:S16–21. doi: 10.1016/j.drugpo.2014.09.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Cescon A, Patterson S, Chan K, et al. Gender differences in clinical outcomes among HIV-positive individuals on antiretroviral therapy in Canada: a multisite cohort study. PLoS One. 2013;8:e83649. doi: 10.1371/journal.pone.0083649. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Wagner KD, Bloom JJ, Hathazi SD, et al. Control over drug acquisition, preparation and injection: Implications for HIV and HCV risk among young female injection drug users. ISRN Addict. 2013;2013:289012. doi: 10.1155/2013/289012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Argento E, Shannon K, Fairbairn N, et al. Increasing trends and incidence of nonfatal overdose among women sex workers who use drugs in British Columbia: The role of criminalization-related barriers to harm reduction. Drug Alcohol Depend. 2023;244:109789. doi: 10.1016/j.drugalcdep.2023.109789. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Ministry of Health Provincial health officer declares public health emergency | BC Gov News. 2016. [18-Jul-2023]. https://news.gov.bc.ca/releases/2016HLTH0026-000568 Available. Accessed.
- 36.Moustaqim-Barrette A, Papamihali K, Mamdani Z, et al. Accessing Take-Home Naloxone in British Columbia and the role of community pharmacies: Results from the analysis of administrative data. PLoS One. 2020;15:e0238618. doi: 10.1371/journal.pone.0238618. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Hashimoto RE, Brodt ED, Skelly AC, et al. Administrative database studies: goldmine or goose chase? Evid Based Spine Care J. 2014;5:74–6. doi: 10.1055/s-0034-1390027. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Canadian Institute for Health Information CED-DxS the Canadian emergency department diagnosis shortlist. 2021