Divergent trajectories in fatal drug harms across high-income settings: a comparative synthesis of public surveillance indicators, 2015–23

Abstract

Background

Fatal drug-related harms have risen across high-income countries over the past decade, yet national surveillance systems differ substantially in case definitions, reporting timeliness, and indicator coverage. These differences make it difficult to determine whether observed cross-national differences reflect epidemiological divergence or heterogeneous measurement.

Methods

To address this, we compared public surveillance indicators of fatal drug-related harm across the USA, Canada, England and Wales, and European countries reporting to the European Union Drugs Agency (2015–23), using three domains: drug-supply toxicity, surveillance latency, and treatment-system integration.

Results

Fatal drug-harm indicators suggested divergent trajectories: the USA experienced rapid mortality escalation through 2022 followed by a modest decline, while England and Wales reported continued increases in registration-based deaths. In Canada, deaths involving three or more substances increased from 20% in 2018 to 42% in 2023, with opioid–stimulant combinations accounting for ~60% of deaths by 2023. European estimates remained provisional due to incomplete national reporting. Differences between occurrence-based and registration-based surveillance influenced the apparent timing of trends. Continuity of opioid agonist treatment across incarceration transitions emerged as a measurable treatment-integration signal.

Conclusions

High-income settings appear to follow distinct fatal drug-harm trajectories shaped by drug-supply toxicity, surveillance architecture, and treatment-system integration. A structured multi-domain framework can help distinguish epidemiological divergence from variation caused by measurement and reporting. It can also guide improvements in timely surveillance, treatment-continuity monitoring, and cross-jurisdictional learning.

Introduction

Substance use disorders (SUDs), including disorders related to alcohol and other psychoactive drugs, remain major contributors to preventable morbidity and mortality globally and across high-income countries.^1,2 At the same time, drug markets are evolving rapidly, with increasing availability of potent synthetic opioids and complex polysubstance exposure patterns, creating new challenges for surveillance and health-system response.³ Global syntheses have highlighted both the scale of health loss attributable to alcohol and drug use and the persistent gap between need and access to effective care.^1,4 Within this evolving landscape, the present analysis examines illicit drug-related mortality, emerging polysubstance toxicity patterns, and selected continuity-of-care indicators. In Canada, the trajectory from prescription-driven opioid harm to a fentanyl-dominated illicit supply illustrates how rapidly drug-market composition can shift within a single high-income setting.⁵

Across these dimensions, high-income jurisdictions have experienced substantial shifts in acute drug-related harms over the past decade, yet trajectories have diverged considerably across settings rather than converging toward a shared pattern of risk and response. Surveillance systems in the USA, Canada, the UK, and European countries reporting to the European Union Drugs Agency (EUDA) document distinct mortality and exposure patterns despite broadly comparable health-system resources.^6–13 Collectively, these patterns illustrate that high-income country status does not translate to uniform risk environments, service coverage, or outcomes but rather to increasingly heterogeneous overdose trajectories.

Against this heterogeneity, evidence-based interventions for opioid use disorder are well established. World Health Organization (WHO) guidance recommends opioid agonist maintenance treatment with methadone or buprenorphine, supported by psychosocial care and overdose prevention measures.¹⁴ Cohort-based syntheses demonstrate substantially lower all-cause and overdose mortality during periods on opioid agonist treatment (OAT) compared with periods off treatment,¹⁵ with risk concentrated around treatment initiation, cessation, and high-risk transitions including release from incarceration. Despite this evidence base, treatment availability and uptake remain uneven across jurisdictions,^16–18 and patterns of engagement differ by substance type, geography, and service capacity. Access gaps are often structural and policy-mediated rather than purely clinical.

This structural variation makes cross-jurisdictional comparison essential for understanding whether epidemiologic burden, treatment access, and acute harms align across health systems. In practice, however, meaningful comparison is complicated by substantial differences in surveillance infrastructure, indicator definitions, and reporting practices across jurisdictions.^11,19,20 These differences can generate apparent variation driven by measurement rather than true underlying risk. Mitigating this requires transparent documentation; triangulation across multiple public surveillance systems can improve interpretability and identify robust cross-national signals relevant to public health and service planning.

Despite the policy salience of these issues, prior comparative work has fallen short in three respects. First, comparisons rely on single indicators without integration across surveillance, supply, and treatment-system dimensions. Second, no structured method distinguishes measurement-driven from epidemiologically driven divergence. Third, comparisons rarely incorporate continuity indicators across high-risk transitions. A structured interpretive approach is therefore required to move beyond simple juxtaposition of heterogeneous indicators and to identify discrete targets for surveillance improvement and policy response.

To meet this need, we propose that divergence in fatal drug harms across high-income settings is shaped by three interacting system domains: (i) drug-supply toxicity and exposure complexity; (ii) surveillance latency and measurement timeliness, anchored in the timeliness, data quality, and stability attributes of the Centers for Disease Control and Prevention (CDC) surveillance-appraisal framework²¹; and (iii) treatment-system integration and continuity across high-risk transitions. These domains were selected because they link proximate mortality risk, the visibility of that risk, and the system capacity to reduce it. The three domains interact: high-toxicity supply combined with delayed surveillance can produce silent escalation, and fragmented treatment continuity at high-risk transitions can amplify supply-side toxicity (Fig. 1).

Figure 1.

Three-domain conceptual framework for divergence in fatal drug harms across high-income settings. Three input domains (top), drug-supply toxicity and exposure complexity, surveillance latency and measurement timeliness, and treatment-system integration and continuity, converge on observed cross-jurisdictional divergence in fatal drug-harm trajectories (bottom). Curved arrows denote cross-domain interactions identified in the discussion: high-toxicity supply combined with delayed surveillance produces silent escalation; latency combined with treatment fragmentation produces delayed feedback; and treatment fragmentation amplifies supply-side toxicity at high-risk transitions (transition-point amplification). The framework is anchored in the timeliness, data quality, and stability attributes of the CDC Updated Guidelines for Evaluating Public Health Surveillance Systems.²¹

In this study, we conduct a retrospective comparative ecological synthesis of publicly available surveillance and system indicator data to characterize patterns of fatal drug-related harm across high-income jurisdictions.^{6–13,16–18,20,22–24} Reporting is guided by the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) framework.²⁵ Rather than seeking causal attribution, we interpret observed variation as structured comparative signals within the three-domain framework. Understanding these divergent trajectories is particularly salient for European health systems navigating an increasingly complex drug supply.

Methods

Study design and reporting

We conducted a retrospective comparative ecological synthesis of publicly available surveillance and health-system indicator data from national and regional sources. The study compares trends in drug-related mortality, patterns of polysubstance involvement in fatal events, and selected treatment continuity indicators across high-income jurisdictions. Reporting is guided by STROBE²⁵ and, where relevant to routinely collected surveillance sources, the RECORD statement.²⁶ Interpretation of cross-jurisdictional surveillance differences was structured using the attributes of the CDC Updated Guidelines for Evaluating Public Health Surveillance Systems,²¹ particularly timeliness, data quality, and stability.

Jurisdictions and study period

The four jurisdictions examined, the USA, Canada, the UK, and European countries reporting to EUDA, were selected on conceptual rather than purely convenience grounds. They represent contrasting surveillance architectures (occurrence-based provisional reporting in the USA; registration-based reporting in England and Wales; opioid-specific federal aggregation in Canada; harmonized supranational reporting via EUDA) and contrasting drug-market profiles (mature fentanyl-dominant in North America; heroin-predominant with emerging synthetic-opioid signals in Europe). High-income status was defined using the World Bank country income classification.²⁷

The primary analytic window was 2015–23, chosen to maximize overlap across surveillance systems while enabling evaluation of pre- and postpandemic periods. Where sources provided provisional or registration-based updates beyond 2023, including provisional US overdose counts and the ONS 2024 registration–based mortality release, these were used descriptively as context but were not treated as fully comparable to finalized annual series.^6,7,12,28

Data sources

All sources were publicly available and were accessed in December 2025, with citation details verified on 9 February 2026. US indicators came from National Survey on Drug Use and Health (NSDUH) for substance-use prevalence and self-reported treatment receipt^20,22 and CDC/National Center for Health Statistics (NCHS) for overdose mortality and provisional counts.^6,7,29 Canadian data came from Public Health Agency of Canada (PHAC) for opioid- and stimulant-related harms and polysubstance toxicity,^8,30,31 Health Canada for federal context,⁹ and CSUS for population-level prevalence.²⁴ UK indicators came from Office for National Statistics (ONS) for drug-poisoning mortality (2023 and 2024 registrations) and substance-use prevalence^12,13,28,32 and Office for Health Improvement and Disparities (OHID) for adult treatment-system activity.¹⁷ European sources included EUDA products: the European Drug Report,¹⁰ Statistical Bulletin indicators for drug-induced deaths,^11,33 prevalence,²³ and treatment demand,¹⁹ and EUDA OAT indicators.¹⁶ Cross-national system capacity was drawn from WHO ATLAS-SU indicators.¹⁸ The primary outcome domain was drug-related mortality, captured through these analogous but not identical indicators; treatment access was treated as a domain supported by multiple indicators rather than a single pooled metric, given construct and denominator heterogeneity. Where standardized OAT coverage and retention data were not consistently available across jurisdictions, OAT continuity is presented as a policy/system signal rather than a population-level coverage metric.

Interpretive rules for indicator nonequivalence

Two interpretive rules were applied to ensure consistency across nonequivalent indicators. First, Canadian opioid-specific data were interpreted as a within-jurisdiction trend indicator and a lower bound on all-drug mortality, never as directly comparable to US all-drug counts; analogously, within-jurisdiction time trends were not equated with cross-jurisdictional level differences. Second, where rate comparisons were presented across jurisdictions, the differing standard populations (US 2000 standard; ONS 2013 European Standard Population) were flagged in table footnotes, and Supplementary Fig. S1 was annotated to discourage league-table interpretation.

Data extraction, harmonization, and quality checks

For each source, we extracted the most recent time series available within 2015–23, the corresponding metadata and definitions, and published measures of uncertainty where available. Data were compiled into a harmonized analytic file including jurisdiction, year, domain, indicator definition, denominator or age group, estimate, and uncertainty if reported. Quality checks included verification that annual values matched the originating source tables, explicit annotation of calendar versus registration year,^12,28,32 and cross-checking directional consistency across related indicators.^6–8,10,11

Statistical analysis

Analyses were descriptive and comparative. We summarized annual estimates and calculated absolute and relative percent change where consistent annual series were available. All nonsurvey surveillance indicators were used as officially reported, without re-estimation.^{8,11,12,17,19,29} Findings were presented as trend plots and summary tables. No meta-analytic pooling was performed given heterogeneity of definitions and data-generating processes.

Ethics and patient/public involvement

All analyses used publicly available, de-identified secondary data. Research ethics review was not required under public-data exemption criteria. No patients or members of the public were involved in the study design, conduct, or reporting.

Results

Data extraction and comparability

We extracted national or European Union (EU)-level indicators of fatal drug-related harm for the most recent overlapping years available, prioritizing annual counts and age-standardized or age-adjusted rates where reported. Because definitions and standardization methods differ across jurisdictions, results are presented using each source’s published metrics and interpreted as comparative patterns rather than strict quantitative equivalence.^11,29,32,33

Divergent mortality trajectories across high-income settings

Figure 2 shows annual US overdose deaths and age-adjusted overdose mortality rates from 2015 to 2023, overall and by sex.²⁹ Drug-related mortality indicators varied substantially across settings (Table 1). The US series showed rapid escalation through 2021, a peak in 2022, and a subsequent decline in 2023 (Table 2), with absolute mortality rising from 52 404 deaths in 2015 to a peak of 107 941 in 2022 (a 105% relative increase across the analytic window) before declining to 105 007 in 2023.²⁹ England and Wales reported continued increases in registration-based drug-poisoning deaths over the most recent comparable period.^28,32 EUDA reported a provisional EU-level estimate for 2023 and emphasized incomplete national reporting and estimation limits for comparability.³³ Canada reported 8049 opioid-related deaths in 2023, using an opioid-specific indicator not directly comparable to all-drug mortality measures.⁸ A cross-jurisdiction comparison of fatal drug-related harm indicators for the latest year is shown in Supplementary Fig. S1.

Figure 2.

Annual number of drug overdose deaths and age-adjusted overdose mortality rates by sex, the USA, 2015–23. Age-adjusted overdose mortality rates (per 100 000 population) are shown for total, male, and female populations (lines), alongside annual counts of overdose deaths by sex (stacked bars). Data are derived from the US National Vital Statistics System as reported by the Centers for Disease Control and Prevention (CDC/NCHS).²⁹ Rates are age-adjusted to the US standard population.

Table 1.

Key drug-related mortality indicators by jurisdiction [latest reported year(s)].

Jurisdiction (source)	Year	Death count (indicator-specific)	Rate (as reported)	Comparability note	Reporting latency to publication
USA (CDC/NCHS)	2023	105 007 overdose deaths	31.3 per 100 000 (age-adjusted)	All-drug; age-adjusted to US standard populationa	~4 months (provisional); 11–12 months (final)
England & Wales (ONS)	2023	5448 drug-poisoning deaths	93.0 per million (age-standardized; ≈9.30 per 100 000)	Registration-based; age-standardized to 2013 ESPb	Median 214 days (England); 279 days (Wales)
EU/Europe (EUDA)	2023	~7500 drug-induced deaths (provisional estimate)	Not directly comparable	EUDA notes incomplete coverage/estimationc	12–18 months (provisional); 1+ year for finalized
Canada (PHAC)	2023	8049 opioid-related deaths	Not reported as standardized rate	Opioid-specific; not all-drugd	Not reported in standard form

Indicators are analogous but not identical; direct rate comparisons should not be interpreted as league-table rankings because case definitions, ascertainment, standardization, and reporting latency differ across jurisdictions. Values shown as published by the source; unit conversion shown only where mathematically direct.

Abbreviations: CDC/NCHS, Centers for Disease Control and Prevention/National Center for Health Statistics; ONS, Office for National Statistics; EUDA, European Union Drugs Agency; PHAC, Public Health Agency of Canada.

^{a 29} ;

^{b 32} ;

^{c 33} ;

^{d 8} .

Table 2.

US drug overdose deaths and age-adjusted mortality rates by sex, 2015–23.

Year	Total deaths, n	Total age-adjusted rate (per 100 000)	Male deaths, n	Male age-adjusted rate	Female deaths, n	Female age-adjusted rate
2015	52 404	16.3	32 957	20.8	19 447	11.8
2016	63 632	19.8	41 558	26.2	22 074	13.4
2017	70 237	21.7	46 552	29.1	23 685	14.4
2018	67 367	20.7	44 941	27.9	22 426	13.6
2019	70 630	21.6	47 881	29.6	22 749	13.7
2020	91 799	28.3	63 728	39.5	28 071	17.1
2021	106 699	32.4	74 301	45.1	32 398	19.6
2022	107 941	32.6	75 814	45.6	32 127	19.4
2023	105 007	31.3	74189	44.3	30 818	18.3

Source: CDC/NCHS Data Brief (NVSS mortality).²⁹

Reporting latency varies markedly across these systems (Table 1, latency column): US provisional overdose data are released ~4 months after date of death^6,7; ONS registration–based mortality carries median delays of 214 days in England and 279 days in Wales^28,32; and EUDA-finalized drug-induced deaths typically lag by 12–18 months.³³ These differences shape the timing at which apparent trends become visible across systems.

Drug-supply toxicity and exposure complexity

Figure 3 illustrates the marked increase in multi-drug involvement in apparent opioid- or stimulant-related toxicity deaths in Canada between 2018 and 2023. Canadian surveillance demonstrates a pronounced shift toward polysubstance toxicity (Table 3).^30,31,34 The proportion of apparent opioid- or stimulant-related toxicity deaths involving three or more substances increased from 20% in 2018 to 42% in 2023, with opioid-stimulant combinations accounting for ~60% of deaths in 2023.^30,31 In the USA, age-adjusted overdose mortality rose from 21.6 per 100 000 in 2019 to 32.4 per 100 000 in 2021, a 50% increase over 2 years (Table 2; authors’ calculation from CDC NCHS Data Brief No. 522²⁹). EUDA synthesis indicates that opioids remain frequently implicated in European drug-induced deaths, without comparable evidence of rapid increases in three-substance involvement at the same scale.³³ US surveillance has also documented increasing stimulant co-involvement in opioid overdose deaths, indicating that polysubstance exposure is not unique to Canada, although the Canadian PHAC data provide the clearest directly comparable 2018–23 indicator.^29,30,33

Figure 3.

Increasing polysubstance involvement in apparent opioid- or stimulant-related toxicity deaths in Canada, 2018–23. Proportion of apparent opioid- or stimulant-related toxicity deaths involving more than or equal to three substances increased from 20% in 2018 to 42% in 2023. The 2023 composition is shown for context, with opioid–stimulant combinations accounting for ~60% of deaths, opioid-only involvement ~10%, and other patterns ~30%. Data are derived from PHAC Health Infobase multi-drug involvement reporting (death investigation data, 2018–23).³⁰

Table 3.

Polysubstance involvement signals (Canada, 2018 vs 2023).

Metric	2018	2023
Toxicity deaths involving ≥3 substances	20%	42%
Share of deaths involving opioids + stimulants (± other psychoactive substances)	—	~60%
Share of deaths involving opioids alone	—	10%

Source: PHAC multi-drug involvement summary.³⁰

Treatment-system integration and continuity of care

Treatment access indicators differed in construct and denominator across systems, limiting direct international comparability.^18–20,22 In England, administrative statistics indicate substantial numbers of adults in contact with structured substance misuse treatment services annually.¹⁷ In the USA, population survey data continue to suggest a persistent gap between estimated need and self-reported treatment receipt.^20,22 European monitoring emphasizes treatment demand indicators and system descriptors, reflecting heterogeneity in service organization across countries.^16,19

As a continuity-of-care proxy for a high-risk transition, EUDA reporting for 2023 indicates that OAT continuity from community to prison is available in 28 of 29 reporting countries (all EU Member States except Slovakia, plus Norway and Türkiye). In-prison OAT initiation is permitted in 24 countries, and postrelease continuation is available in 23 countries. Take-home naloxone in prison expanded from three countries in 2021 to seven in 2023, consistent with WHO community-overdose management guidance.^35,36 This continuity-of-care indicator represents a measurable system-level signal of treatment-system integration across a transition associated with elevated overdose mortality risk.

Discussion

Main finding of this study

Published surveillance indicators do not show a common fatal drug-harm trajectory across high-income settings with broadly similar economic and clinical resources. Between 2015 and 2023, the USA, Canada, England and Wales, and European countries reporting to the EUDA diverged sharply on the indicators reported by their respective national and supranational systems: the USA passed through a peak and entered a decline in overdose mortality between 2022 and 2023 (Table 2)²⁹; England and Wales continued to report rising drug-poisoning deaths into the most recent registrations³²; Canadian surveillance captured a pronounced compositional shift, with deaths involving three or more substances rising from 20% in 2018 to 42% in 2023 (Table 3)³⁰; and European countries reporting through EUDA registered roughly 7500 drug-induced deaths in 2023, with figures that remain provisional and lagged.³³ These signals are not four unrelated trends. Read through the three-domain framework, they describe how drug-supply toxicity, surveillance architecture, and treatment-system integration jointly shape what each country sees, when, and how clearly. Whether observed differences reflect epidemiological reality, surveillance architecture, or both is the central interpretive question this comparison must answer.

What is already known on this topic

The question of whether cross-national differences in fatal drug harms reflect epidemiological reality, surveillance architecture, or both has been engaged by several distinct lines of prior work. Cross-national comparisons in this field are constrained by definitional and standardization differences across drug overdose, drug poisoning, drug-induced death, and opioid-related death indicators.^11,29,32,33 The CDC Updated Guidelines for Evaluating Public Health Surveillance Systems specify nine attributes for surveillance appraisal, including timeliness, data quality, sensitivity, representativeness, and stability, which map directly onto the heterogeneity observed across the systems compared here.²¹ Beyond these architectural considerations, North America has experienced exceptionally high overdose mortality in the context of high-potency synthetic opioids and evolving polysubstance exposure,³ while European monitoring continues to emphasize rapidly changing supply patterns supported by harmonized but variably timely indicators.^10,33 On the treatment side, evidence-based pharmacological treatment for opioid use disorder substantially reduces all-cause and overdose mortality during periods on treatment compared with periods off treatment,¹⁵ with risk concentrated around treatment initiation, cessation, and high-risk transitions including release from incarceration.^37–39

What this study adds

Our analysis contributes a structured method for distinguishing measurement-driven from epidemiologically driven divergence across high-income surveillance systems. Within-jurisdiction temporal signals, most clearly the Canadian polysubstance shift (Table 3), are unaffected by cross-jurisdictional measurement variation and represent the most defensible basis for cautious comparison.^30,40 The magnitude of separation between US and England and Wales mortality is unlikely to be explained by standardization differences alone, although we continue to avoid treating these indicators as directly equivalent. By contrast, aggregate EU-level totals remain materially constrained by incomplete national reporting and should be interpreted as approximations.³³

The surveillance-latency dimension is empirically grounded. Provisional CDC reporting becomes available ~4 months after date of death; ONS registration-based data carry median delays of 214 days in England and 279 days in Wales^28,32; and EUDA-finalized drug-induced deaths typically lag by 12–18 months.³³ The practical implication is that a mortality inflection captured in near-real-time provisional systems may not become visible in registration-based systems until many months later, particularly where deaths registered in a given year include substantial proportions of deaths occurring in previous years. This asymmetry has direct policy consequences when supply-driven mortality changes are unfolding rapidly.

The CDC framework also helps interpret toxicology-panel limitations identified by the data-quality and stability attributes. Toxicology panels in many jurisdictions evolved historically for forensic and prosecutorial purposes rather than epidemiological surveillance, with implications for both substance-specific detection (panels may not include novel synthetic opioids or nitazenes until incorporated through formal updates) and evidentiary thresholds for substance attribution. Examples of public-health responses to these challenges include supranational early-warning mechanisms such as the EUDA Early Warning System and standardized death-investigation platforms such as CDC SUDORS.

Turning from surveillance architecture to treatment-system integration, the third domain is anchored in country-count data. EUDA reporting for 2023 indicates that continuity of OAT from community to prison is available in 28 of 29 reporting countries; in-prison initiation is permitted in 24 countries; post-release continuation is available in 23 countries; and take-home naloxone in prison expanded from three countries in 2021 to seven in 2023, consistent with WHO community-overdose management guidance.^35,36 Given that the early post-release window carries some of the highest known overdose mortality risks,^37,38 OAT continuity at this transition represents an important system-level factor.^15,39

These domains do not act in isolation. High-toxicity supply combined with delayed surveillance can produce silent escalation, and fragmented treatment continuity at high-risk transitions can amplify supply-side toxicity. Socioeconomic context, criminal-justice policy, and housing or social-protection regimes constitute a candidate fourth domain that we have not incorporated, on the grounds that adequately operationalizing it across four jurisdictions would require a structural-determinants framework beyond the scope of the present surveillance-indicator synthesis. Full appraisal of high-income surveillance systems against the CDC’s nine attributes, together with structural-determinants integration, is an important area for future methodological work.

A final consideration concerns the post-study period. Recent declines in US overdose mortality may reflect shifts in the composition or potency of the illicit drug supply, particularly involving fentanyl and related synthetic opioids.^41,42 Peer-reviewed analyses suggest the supply-side contribution explains only a modest fraction of the decline⁴³; regional and demographic patterns vary substantially⁴⁴; and CDC surveillance documents the reemergence of high-potency analogues including carfentanil.⁴² Continued monitoring is therefore warranted as further peer-reviewed analyses of post-2023 data become available.

Limitations of this study

Several limitations qualify these findings. First, indicators are not fully harmonized, and the EUDA estimate for 2023 is provisional and partly modelled due to incomplete reporting.³³ Second, the empirical depth of the analysis is asymmetric across jurisdictions, with North American indicators supporting more granular trend description than European indicators, partly reflecting the maturity and timeliness of the underlying surveillance systems themselves. Beyond data, the analysis does not capture subnational heterogeneity (provincial, state, or regional variation), which may be substantial and policy-relevant, particularly in Canada and the USA.^7,8 The design cannot attribute changes to specific interventions or disentangle interactions among drug supply, socioeconomic conditions, service delivery, and policy environments.^3,10 Generalizability is also bounded: surveillance-architecture challenges identified here are not unique to high-income settings; in lower- and middle-income contexts, enforcement-derived indicators frequently substitute for epidemiological surveillance, presenting distinct considerations beyond the present scope.⁴⁵ Despite these limitations, applying CDC’s surveillance-appraisal attributes to cross-national mortality indicators offers a structured approach for future comparative methodological work.²¹

Conclusions

Public surveillance data indicate substantial variation in fatal drug-related harm and treatment-system indicators across high-income settings during 2015–23. Three concrete priorities emerge for cross-jurisdiction learning. First, within-jurisdiction temporal signals such as the Canadian polysubstance shift from 20% to 42% over 2018 to 2023 are unaffected by cross-national measurement artifact and warrant prioritized attention as a defensible cross-comparable indicator type.^30,31 Second, reporting-latency differences across surveillance systems (Table 1) materially affect the timing of public health response and warrant explicit harmonization effort.^6,7,28,32,33 Third, the EUDA prison OAT continuity infrastructure, available in 28 of 29 reporting countries in 2023, provides a measurable comparator for treatment-system integration that other jurisdictions could draw on.³⁶

Supplementary Material

Figure_S1_Cross_jurisdiction_fatal_drug_harm_2023(1)

Supplementary Figure S1. Cross-jurisdiction fatal drug-related harm indicators, 2023. Each jurisdiction is shown in a separate panel with its own metric and unit: age-adjusted all-drug mortality rate per 100,000 (United States; 2000 US standard population), age-standardised drug-poisoning mortality rate per 100,000 (England and Wales; 2013 European Standard Population), opioid-specific death count (Canada), and provisional drug-induced death count (EU/Europe, EUDA). Indicators are not directly comparable across panels owing to differing case definitions, standardisation, and ascertainment, and should not be interpreted as a league-table ranking. Sources: CDC/NCHS (USA); ONS (England and Wales); PHAC (Canada); EUDA (EU/Europe), 2023.

Ovie Martin Albert, Assistant Professor; Family Physician; Addiction Medicine Physician

Alexander Arthur, Assistant Professor; Family Physician; Addiction Medicine Physician

Author contributions

O.M.A. conceived the study, designed the methodology, conducted the analysis, and drafted the manuscript. A.A. contributed to study design, data interpretation, and critical revision of the manuscript. Both authors approved the final version.

Conflict of interest

The authors declare no conflict of interest.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

All data analysed in this study are publicly available from the sources cited in the reference list. Extracted and harmonized datasets generated during the current study are available from the corresponding author on reasonable request.

Ethics approval

This study used publicly available, de-identified secondary data obtained from national and international surveillance systems. No individual-level identifiable information was accessed. Research ethics approval and informed consent were not required in accordance with standard public-data exemption criteria.