Where are the reports? The state promised its testing results in writing

In September 2025, Colorado's Marijuana Enforcement Division announced something genuinely good: the state would buy cannabis off dispensary shelves, as a customer would, and test whether the products matched their labels and were safe. And the founding bulletin made a specific promise about what the public would get back:

"Following completion of the initial Program, the CDPHE team will collect, organize, and analyze the data. The goal is for CDPHE to publish an initial summary of data approximately 30 days after Program completion and a more detailed report approximately 90 days after Program completion."

That is Industry Bulletin 25-11, the program's own announcement, in the Division's own words. Sampling ran through the end of January 2026. So the summary was due around early March 2026, and the detailed report around early May 2026.

As of this page's publication, neither document exists. Not on the Division's Data and Resources page, not on its Marijuana Sciences page, not anywhere on the public web we or anyone else has found. We re-checked the Division's pages the night this article published.

Where the findings actually live

The results were not suppressed; they were presented. If you attended or watched the June 12 and July 10 Science & Policy Forum meetings, you saw the edibles findings, and they were substantial: every safety test passed, while only 47% of sampled edibles passed both of their label checks.

But "presented at a meeting" and "published in writing" are different things, and the difference is who gets to know. A written report can be read, cited, checked, and linked by anyone, forever. A finding inside a two-hour video recording is available in principle and invisible in practice. This site transcribes those recordings precisely because almost nobody else will.

And the gap is narrower and stranger than pure absence. By June, a data summary already existed in the forum's sign-up-gated folder, and the full report was described at the meeting as being in final publication approval. The document appears to be written. It has simply never been placed where the public can find it, and that is what makes "where is it published?" the exact question to ask.

And the pattern extends beyond the reports. Every 2025 forum entry on the Division's Public Engagement page carries a posted meeting agenda. Every 2026 entry offers a video recording only. The written-agenda practice stopped in the year the surveillance results came due. Slide decks and handouts sit behind a Google Group that requires registration to join.

Meanwhile the results waited for the meetings themselves. The edibles findings, from samples collected starting in October 2025, were described as coming soon in November, January, February, and March before being presented in June. As our year-in-review put it: the transparency argument was never resolved; it was outlasted.

The silences that make it a pattern

One missed deadline is a busy quarter. This is more than one:

• The 30-day summary: due ~March 2026, not published.
• The 90-day detailed report: due ~May 2026, not published.
• The Division's own 2025 annual enforcement report, published in February 2026, after sampling ended: it does not mention the surveillance program, potency, homogeneity, or label accuracy at all. The only appearance of "Reduced Testing Allowance", the program's central compliance issue, is as a revenue line item in the Division's fee income.
• The response ladder is invisible by design. The founding bulletin committed the Division, when a sample failed, to quarantine and corrective-action paperwork first, with recall and embargo only as escalations. The first two rungs produce no public document. As our enforcement article lays out, nothing about the surveillance failures ever surfaced on the public rungs. The public record cannot distinguish quiet diligent remediation from nothing having happened, and that is a design property, chosen before the first sample was bought.

And one more silence sits behind all of these: the decision that consolidated the reporting into a single end-of-program document has no findable author. That one is its own article.

The credit column, and one line we deliberately leave out of it

Hold this next to the missing reports, because both are true. The state kept its November commitment to publish the reference lab's methods: five of them are now public. And when the forum was asked hard questions about its own data, the answers were candid, including the August 7 admission that the gap between licensed-lab results and state retests has no explanation yet.

The meetings being recorded and posted in full is why this archive can exist, and we deliberately state that as a fact rather than a credit. Public bodies' meetings are open by law in Colorado, and we have not yet established how much of the recording and posting is legal obligation rather than choice. Credit belongs to deeds beyond the required minimum; until we know where that minimum sits, this one stays off the ledger.

There is also an innocent reading of the missing reports, and we offer it in good faith: with seven licensed labs and a small pool of manufacturers, publishing de-identified data without small cells that identify individual businesses is genuinely hard. That may be the real blocker. But here is the thing about an innocent reading: it costs one paragraph to say so in public. "The report is delayed because de-identification at this market size is hard, and here is the new date." No such paragraph has appeared anywhere.

What we will do about it, and what you can

Colorado law is unusually specific here. C.R.S. 44-10-204 requires that testing records be publicly available in aggregated, de-identified form. That is not a courtesy the Division extends; it is a statutory duty. Claim strength: the statute's own text; we are not lawyers. And the Colorado Open Records Act gives any person, not just journalists or lawyers, the right to request public records, with a response clock measured in working days.

Our posture, in order:

1. Ask, unambiguously, at September 4. Not whether the report exists; where it is published. A URL from the podium ends this article's complaint on the spot, and if the answer is a blocker (de-identification, approval queues), a stated reason and a date nearly do.
2. Then a dated trigger, not an open-ended hope. If the written surveillance report is not publicly posted ten days before the forum's next meeting after September 4, this site will file a records request under the Colorado Open Records Act for the report and the records of its approval. That is not a threat; it is a calendar. The document appears to already exist. Ten more days of any month is enough time to put a finished document on a public page.
3. And the general rule stands. Going forward, when committed public data goes missing past its own deadline without a stated reason, this site will ask its members to file records requests for it. Any Coloradan can file one. A single request from a small website is easy to wait out; requests from a membership of patients, each with a statutory clock attached, are not. We would rather the reports simply publish. That choice is the agency's.

We will update this page when either thing happens, and we will say so plainly if the answer turns out to be as innocent as the good-faith reading above.

Quotes from Division bulletins are from the bulletins as published. Meeting characterizations are from our machine transcripts of the Division's published recordings, which are the source of record.