What the register is
Regulation 17 of the Control of Major Accident Hazards Regulations 2015 says the Competent Authority must make this information available to the public and keep it updated, for every COMAH establishment: who the operator is, where the site is, what it does, which dangerous substances are on it, what could go wrong, and how people nearby would be warned. Operators must supply the Competent Authority with that information and with revisions when it changes. In Great Britain the Competent Authority is HSE acting jointly with the Environment Agency, Natural Resources Wales or the Scottish Environment Protection Agency, and the information is published through a search page run by HSE. Two fields on each record are the ones this report is about: the date of the last planned COMAH site visit by the Competent Authority (in the words of the Regulations, the last site visit carried out under the programme of routine inspections), and, for upper-tier sites, whether a safety report has been submitted.
How we read it
The register has no list and no bulk file. A search by full postcode returns the establishments within three miles, and each record is served only inside that search. So we laid a 4.5 km grid over Great Britain, matched each point to its nearest postcode using the Office for National Statistics open postcode data, ran one search for each of the roughly eleven thousand unique postcodes, and then read each establishment's record once. One request at a time, identified as Amberhold in every request, with a pause of about 0.3 seconds between each. Both passes described below ran on 17 September, about ten hours in all and roughly 21,400 searches.
Two coverage checks were run before any number was published. Every establishment found by an earlier, smaller search of industrial postcodes had to appear in the national set: all 195 did. Then the whole grid was run a second time, shifted by half a cell in both directions, so that any site sitting at a corner of the first grid would fall near the centre of the second. The second pass found the same 852 establishments and not one more.
That shows the method is consistent, not that it is complete. A grid point is searched only if a postcode lies within two kilometres of it, so an establishment far from any postcode could be missed by both passes. HSE does not publish a total to check against. The figures here are for every establishment our search found. Northern Ireland, the Isle of Man and the Channel Islands are outside this register.
The numbers
Establishments on the register
852
read 17 September 2026
Upper tier
328
safety report required
Lower tier
524
major accident prevention policy required
Establishments by region
From the postcode on the record; nation or English region
Table view
| Band | establishments | share |
|---|---|---|
| Scotland | 208 | 24.4% |
| North West | 122 | 14.3% |
| Yorkshire and the Humber | 106 | 12.4% |
| South East | 63 | 7.4% |
| East of England | 62 | 7.3% |
| West Midlands | 59 | 6.9% |
| North East | 54 | 6.3% |
| Wales | 50 | 5.9% |
| South West | 49 | 5.8% |
| East Midlands | 48 | 5.6% |
| London | 31 | 3.6% |
The last-visit date
Each record carries one date: the last planned, meaning routine, visit to the site by the Competent Authority, as the public record shows it. On 17 September 2026, 772 records carried a date and 80 carried none. Of the dated records, 548 (71%) showed a date more than three years before the day we read it, and 307 more than five years. The median dated record was 4.51 years old. The oldest date on the register is 11 February 2011.
Age of the last planned visit date, all establishments
Years between the date on the record and 17 September 2026. Bands of three years and more are drawn in the accent.
Table view
| Band | establishments | share |
|---|---|---|
| under 1 year | 86 | 10.1% |
| 1 to 2 years | 62 | 7.3% |
| 2 to 3 years | 76 | 8.9% |
| 3 to 5 years | 241 | 28.3% |
| 5 to 10 years | 254 | 29.8% |
| 10 years or more | 53 | 6.2% |
| no date | 80 | 9.4% |
Upper tier
same bands, shared scale
Table view
| Band | establishments | share |
|---|---|---|
| under 1 year | 54 | 16.5% |
| 1 to 2 years | 29 | 8.8% |
| 2 to 3 years | 39 | 11.9% |
| 3 to 5 years | 113 | 34.5% |
| 5 to 10 years | 87 | 26.5% |
| 10 years or more | 2 | 0.6% |
| no date | 4 | 1.2% |
Lower tier
same bands, shared scale
Table view
| Band | establishments | share |
|---|---|---|
| under 1 year | 32 | 6.1% |
| 1 to 2 years | 33 | 6.3% |
| 2 to 3 years | 37 | 7.1% |
| 3 to 5 years | 128 | 24.4% |
| 5 to 10 years | 167 | 31.9% |
| 10 years or more | 51 | 9.7% |
| no date | 76 | 14.5% |
What a three-year-old date can mean
COMAH 2015 sets no fixed visit interval. The Competent Authority sets how often it visits each kind of establishment in its programmes for routine inspections, based on a systematic appraisal of the hazards (regulation 25). Three years is a reading threshold we chose for this report, not a legal benchmark. A date older than that has at least two honest readings, and the register cannot tell you which applies to a given site.
- No routine visit has taken place in that time, because the programme for that site does not call for one.
- A routine visit has taken place, and the public record has not been updated to show it.
Whichever explanation applies, the public-facing record still displays that historic date. That is the finding. It is a finding about records, not about whether an inspection took place.
The safety-report flag
Every upper-tier operator must prepare a safety report and send it to the Competent Authority, and the public record carries a yes/no field saying whether it has been submitted. Of 328 upper-tier records, 317 say yes, 11 say no, and 0 leave the field blank.
11upper-tier records show “No” for “Safety report submitted”. This field should not, on its own, be interpreted as evidence that an operator has failed to meet a duty under COMAH. The Regulations contain different submission timescales depending on how an establishment becomes subject to the upper-tier requirements, and the public-register entry may not establish the circumstances or timing applicable to an individual site.
485 distinct operators hold the 852 establishments. This report names none of them and lists no establishment. Every figure here is national, by tier or by region.
What this does and does not mean
Straight talk
This report describes a public register on the day it was read. It does not describe the safety of any site, and it draws no conclusion about whether any legal duty has or has not been met. The last-visit field shows the last routine visit as the public record displays it; a stale date does not establish anything about whether the Competent Authority has visited a site, and this page never says that it does.
Reactive visits, investigations and enforcement are not on the register. Nothing here is a league table, and this report names no operator or establishment. These are heavily regulated sites, and nothing on this page says otherwise.
Amberhold is a software company. We build the record of operator rounds and temporary controls for industrial sites, so we have an interest in records that can be produced when asked. We found no published national analysis of this register, and we have published the data so that anyone can check the reading.
Get the data
One row per establishment, as read on 17 September 2026: comah-register-2026.csv. It is a reproducibility file: enough to check every number on this page and nothing more. Columns: region, tier, subject to COMAH, notification submitted, the last planned visit date as displayed, the same date in ISO form, its age in years on the read date, and the upper-tier safety-report flag. Region and the two date columns are ours. There are no operator, establishment or place names; anyone wanting a particular site can search HSE's register directly. If you find an error, tell us and we will correct the file and say so here.
Sources
HSE, COMAH 2015 public information search, notifications.hse.gov.uk/COMAH2015, read 17 September 2026. Postcode centroids: ONS Postcode Directory via postcodes.io. Contains public sector information published by the Health and Safety Executive and licensed under the Open Government Licence v3.0.