Countries appear first, followed by English regions. All mortality bars compare broad drug-poisoning age-standardised rates for 2024 registrations; the latest national headlines retain their own periods and definitions.
Editorial review: 4 November 2026
Scotland1,133Drug use deaths registered · 2025
Northern Ireland251Drug-related deaths · 2024
North East rate135.22025 · per million · highest in England
Wales rate141.32025 · per million · 422 deaths
Why the North East total looks lower
Raw totals rise with population size. The North East registered 349 drug-poisoning deaths in 2025 and remained the highest-rate English region at 135.2 per million, down from 167.6 in 2024. Wales registered 422 deaths at 141.3 per million. The bars below retain 2024 for a common-year comparison across the four nations; current headlines and substance tables retain their separate 2025 labels.
M: published · T: selected · A: selected · R: selected. Area-series availability, not completeness.
Any opioid / heroin-morphine · 2024
131 / 35
Cocaine / benzodiazepines
71 / 128
Pregabalin / nitazenes
101 / Not separately published
Hospital / ambulance
No comparable UK-wide public feed
Evidence gap
Separate reporting and policing systems
Regional substance data
English-region and Wales substance figures refer to 2025 registrations in the ONS selected-substances dataset; the main cross-country bars retain 2024. ONS reports that 62.7% of 2025 registrations relate to earlier deaths; the separate occurrence series ends in 2024 using an 18-month registration window. The male drug-poisoning rate rose to 130.0 per million while the female rate fell to 56.6; the overall decline does not establish safer markets or falling risk for every group. Cocaine includes powder and crack because the mortality tables do not split them. Scotland and Northern Ireland publish different definitions and are shown from their own national tables. A blank or unpublished regional nitazene figure is not treated as zero.
Explore a source-defined annual series; keep counts, population-adjusted rates and reporting years distinct.
Official mortality series
North East · Broad drug-poisoning deaths
Points show the published rates. Whiskers show their published 95% confidence intervals. The chart scale is set for the selected area.
Scroll horizontally if needed. Exact values are also available below.
2015
98.9
2025
135.2
Difference over this period
+36.3deaths per million
Exact annual counts, rates and uncertainty
North East · Broad drug-poisoning deaths; registration year. Rates and 95% confidence intervals per million.
Year
Deaths
Rate
95% interval
2015
245
98.9
86.4–111.3
2016
260
104.9
92.1–117.7
2017
289
116.7
103.1–130.2
2018
354
143.7
128.6–158.7
2019
332
135.1
120.5–149.7
2020
389
158.0
142.2–173.8
2021
402
164.1
148.0–180.2
2022
332
133.9
119.4–148.4
2023
441
174.3
157.9–190.7
2024
429
167.6
151.7–183.6
2025
349
135.2
121.0–149.5
Published patterns · TD interpretation
Persistent differences, changing annual figures
North East: a sustained difference
The North East rate was 98.9 per million in 2015 and 135.2 in 2025. England’s corresponding rates were 63.6 and 89.5. These source-defined rates describe persistent geographical differences; they do not explain them.
Annual changes differ between areas
From 2024 to 2025, London’s rate changed from 76.5 to 58.3 and the East Midlands from 102.1 to 94.6. The North East changed from 167.6 to 135.2 and the North West from 118.9 to 115.7. Other regions moved differently: Yorkshire and the Humber rose from 115.4 to 124.1 and the South West from 83.0 to 96.5. These are observed differences, without a claim of statistical significance.
Counts describe the volume of recorded deaths. Age-standardised rates account for population size and age structure. Registration delays separate the reporting year from the date of death.
Observe change without inventing its cause
These observations cannot isolate changes in supply, exposure, treatment, deprivation or recording. Broad drug-poisoning and narrower drug-use series remain different measures.
Evidence boundary
The selected series retains its source definition. A change in one area cannot be attributed to a specific local policy or drug combination from these totals alone.
04
Substances contributing to regional harms
Recorded involvement is different from sole causation, prevalence or risk among people using a drug.
Substance involvement · overlapping records
North East · selected substances
ONS broad drug-poisoning deaths, 2025 registrations; substances mentioned on the death certificate · 349 deaths in the stated total. Substance counts overlap.
0147
Any opioid (ONS opiate group)147
Heroin/morphine49
Methadone67
Cocaine80
Amphetamine5
MDMA/ecstasy8
Benzodiazepines78
Pregabalin91
Gabapentin16
Synthetic cannabinoids13
TD interpretation · combinations and attribution
Several substances can be recorded in one death
These are selected substances mentioned or implicated under the source’s definition. One death can appear in more than one bar, and grouped categories can include the individual substances below them.
Opioids and depressants
Review local toxicology and overdose intelligence for opioids alongside benzodiazepines, gabapentinoids, alcohol and emerging sedatives. Separate mentions do not identify which combinations occurred.
Cocaine and emerging synthetic opioids
Cocaine mortality tables may combine powder and crack. Nitazene observations are affected by testing coverage; a missing region-level publication does not establish absence.
Compare within a defined series
Prescribing status, route, intention and sole causation cannot be recovered from these counts. Use the original definitions before comparing national datasets.
Evidence boundary
A share of registered deaths is not the risk of death among people using that substance. A comparable exposed population would be needed for that question.
05
Treatment demand and changing needs
Service contact and problem-substance records provide a separate perspective on demand.
Official treatment records · separate from mortality
England · four adult treatment groups
0138,255
Opioid (OHID opiate group)138,25541.9% of the stated cohort
Non-opiate only43,67813.2% of the stated cohort
Non-opiate and alcohol47,75814.5% of the stated cohort
Alcohol only99,95530.3% of the stated cohort
Treatment trend · source-defined population
England · adults in treatment
All adults in specialist treatment during each financial year; not a population-use trend.
Scroll horizontally if needed. Exact values are also available below.
2015–16
288,843
2016–17
279,793
2017–18
268,390
2018–19
268,251
2019–20
270,705
2020–21
275,896
2021–22
289,215
2022–23
290,635
2023–24
310,863
2024–25
329,646
England · separate cohorts and measures
Co-occurring needs and recorded treatment exits
Recorded mental health treatment need
124,76373.6% of 169,542 new adult journeys
Completed treatment exits
70,65446.5% of 152,067 recorded exits
Mental health need was recorded at entry. Completion includes no use or free of dependence at the recorded exit; it is not evidence of sustained recovery.
06
What the regional evidence tells us
TD interpretation of the linked statistical sources; evidence confidence concerns the measure, not a score for the area.
01
Population and age structure
A large count can reflect a large population. Age-standardisation improves comparisons, but does not adjust for every social, health or exposure difference.
02
Deprivation and exposure
Differences can reflect accumulated disadvantage and exposure. Area associations cannot establish what caused an individual death or prove a geographical explanation.
03
Recording and toxicology
Testing changes can alter which substances are identified. Registration delays and different national definitions constrain comparisons.
04
Services and access
Mortality and treatment records cannot alone distinguish unmet need, changes in access, retention, continuity or the effects of service quality.
05
Rapid intelligence
Local alerts and submitted samples can identify specific threats. Selected detections cannot be treated as representative market prevalence.
06
Unresolved explanations
A hotspot or changing trend identifies a question to investigate. It does not establish that local supply, enforcement or one intervention caused the pattern.
How to use the confidence bands
High confidence: robust direct evidence for the stated measure, place and period. It does not mean every explanation of that pattern is established.
Moderate confidence: credible evidence with limitations in coverage, completeness, recency or comparison.
Limited confidence: sparse, selected, preliminary or inconsistent information supports only a narrow conclusion.
These are qualitative editorial judgements. They are not probabilities, statistical confidence intervals or regional harm scores.
07
Implications for local services
Planning considerations to check against local need and current guidance; these charts do not prove intervention effects.
01
Commissioners and public health
Review comparable trends alongside local treatment access, housing, physical health and prevention capacity. Ask which groups are missing from routine service data.
02
Drug and alcohol treatment
Assess substance combinations and co-occurring needs. Review continuity at referral, transfer, hospital discharge and release from custody.
03
Overdose prevention
Review naloxone availability, emergency response and synthetic-opioid preparedness against local alerts and national guidance.
04
Mental and physical health
Check access to assessment for co-occurring mental health needs and drug-related physical symptoms, with agreed responsibility across services.
05
Local intelligence partnerships
Link defined mortality, treatment, toxicology and emergency-care measures. Date alerts, record uncertainties and avoid treating sample counts as prevalence.
06
Review and evaluation
Monitor access, continuity, equity and outcomes with suitable denominators. Compare service changes with wider trends before claiming an effect.
Three questions for a local review
Is the local measure comparable across its stated years and neighbouring areas?
Which people, substances or service transitions are not adequately described by the available records?
What additional evidence would distinguish competing explanations before changing a service?