Twenty-three areas out of thirty are green or blue on four standards: how Scotland closed a five-year mission against drug deaths

In January 2021 the Scottish First Minister answered a drug-death rate that had been climbing for two decades with a National Mission and dedicated money. The treatment half of that answer was ten MAT standards, a floor of minimum expectations for anyone walking into a specialist drug service: medication the same day, a choice of it, assertive outreach after a high-risk event, harm reduction, support to stay in treatment, then psychological support, shared care, advocacy, mental health care and trauma informed practice. Five years on, Public Health Scotland has published what the accompanying statistical summary calls the final benchmarking report in this format, while the report itself says only that the reporting format will be reviewed, page 2. It is a careful and unusually self-critical document. It also grades a programme built to stop people dying without, after page 12, printing a single mortality figure.

What the colours are made of

The single number a reader carries away, 23 of 30 areas green or blue, is a rounded composite of three unequal things. The process stream is local policies and procedures, and the report states that only new material was assessed: areas that had previously submitted process evidence and been given the maximum score were not required to resubmit, so part of a 2026 colour can rest on documents scored in an earlier year. The numerical stream is service activity. The experiential stream is the machinery for collecting feedback, and its score is applied uniformly across all ten standards. The report records one direction of that arrangement, namely that areas with insufficient evidence of engagement with people who have lived and living experience were marked down across all ten standards, and the dashboard puts a number on it: three partnerships did not receive a full score on the experiential stream and so did not rise above provisional green on any standard. Areas chose their own reporting window: a single three-month consecutive period between November 2025 and March 2026 for standards 1, 3, 5.2, 5.3, 8 and 9, and a snapshot of 28 February 2026 for standards 2, 4, 5.1 and 7. Initial scores could be appealed to a ratification panel. The assessing body and the implementation support team that spent five years helping areas meet the standards sit inside the same organisation, which the report describes plainly as support, oversight and feedback in one role, section 5.2.

None of this makes the result soft. The criteria were raised in October 2025 and most areas still cleared them, and the report insists, page 28, that any apparent slippage be read as a revised threshold rather than a decline in delivery. Assertive outreach after a high-risk event was delivered consistently across all 30 areas, page 30. Under a standard requiring more than 75% of new referrals to be screened for mental health, 95% were. Those are real gains, and the early ones came alongside an extra 5 million pounds a year of resource, accompanied in the same year by a Ministerial Direction.

What is not counted, and what the report says about it

The gap is not hidden, it is structural. A benchmark that grades whether the front door opens has nothing to say about what happens to the person who walks through it, and the standards were designed as a floor, not as an outcome measure. No such admission is made on the report's 41 pages; the accompanying data dashboard makes it, on its Home tab, in these words: "It does not report directly on outcomes but recommends actions to address areas where service delivery may be able to improve". Separately, and about something else, section 5.3 calls for a shift from a compliance orientation to an intelligence-led system so that services can keep pace with changing drug trends, and it concedes that its criteria remain oriented around opioid agonist therapy while prescribing of that therapy declines and benzodiazepines, cocaine, ketamine and adulterated supply grow. Immediately after that concession, page 30, it objects that the process, and the collection of experiential evidence in particular, is far broader in scope in terms of relevant outcomes. The one place where an outcome could still have been checked, the retention target under standard 5, was moved out of the report into that dashboard along with the rest of the tables and graphs. The dashboard was opened for this analysis, and the figure is there: 19,773 people reported in total at the snapshot of 28 February 2026, of whom 17,036 had been retained in treatment for six months or more, 86.2% against a target of at least 75%, with an area minimum of 62.5%, a median of 91.2% and no partnership reporting implementation difficulties under this standard. The number whose absence would otherwise carry this argument exists. It simply does not live in the report.

The report and its own summary do not say the same thing

Two documents were published on 7 July 2026. The three-page statistical summary states, under Background, that the report is published in the context of continuing harms and that it notes 1,017 drug-related deaths in 2024 while suspected drug deaths increased by 8% in 2025. The first half is true, page 9. The second is not in the report: the words suspected and 8% are not found on any of its 41 pages. Two qualifications follow. On page 31 the report does carry a hyperlink anchored (RADAR 2026) leading to Public Health Scotland's suspected drug deaths measure, so it points at that series without printing a number from it. And the 8% itself is checkable, though not against the RADAR output of April 2026, which gives only quarterly comparisons: 330 suspected deaths between 1 December 2025 and 1 March 2026, 31% more than the previous quarter. Its source is the Scottish Government release of 17 March 2026, Suspected drug deaths in Scotland: October to December 2025, which reports 1,146 suspected deaths in 2025, 8% or 81 more than the 1,065 of 2024. The summary adds a second thing the report does not carry: the word nitazene is not found on any of its 41 pages either. The direction is worth noticing, because it is the opposite of the usual one. Here the short summary is the more candid document and the full report is the quieter one. I followed the report, and named the dashboard and the government release where they carry what the report does not.

What the authors call their main conclusion

It should be stated as they state it, because the reading above is not theirs. Section 6, pages 34 to 35: there has been substantial and notable progress in implementing the standards across the five years, laying a firm foundation for future challenges. The sentence that follows is narrower than it looks. Progress was made, they write, at the same time as continuing to incorporate and adapt to annual upgrades to the evidence being required, and the most substantial of those annual upgrades was the introduction of experiential evidence aligned with the Charter of Rights published in 2024. That, and not access or prescribing, is what they single out, and they call it a significant shift away from stigma and criminalisation towards public health and human rights, now embedded to the point where compliance in most areas can be regarded as business as usual. The challenge they name for the future is extending the same standards and rights to non-opioid substances, including alcohol. A prison assessment was submitted for the first time this year and is due in autumn 2026, with experiential reporting from spring 2027.

What a clinician can take from this is narrow and useful. A national system can be made to open its door within one calendar day, offer a choice of medication and chase people after a high-risk event, and this can be verified area by area within five years. Whether that system then keeps people alive is a different question, answered by different data, and this report does not attempt it.