The February 6 forum: a risk score for new tests, and the third-party sampling fight arrives early

Colorado's Science & Policy Forum met February 6, 2026, the first working meeting under the new testing rules. We transcribed the Division's published recording (about 2 hours 5 minutes) with a speech-recognition model and drafted this summary from that transcript; our editor has reviewed and approved this page, and the recording remains the source of record. January's plan had February down for a proficiency-testing deep-dive. That's not what happened: the meeting ran on the new-analytes process and pulled the third-party sampling debate forward by four months. That's a sign of where the pressure actually was.

Notes from the editor

The detail in this meeting that should stop you cold: Colorado already had third-party sampler rules (division-approved, March 2014 to November 2015) and repealed them, for reasons nobody in the room could document. The Division offered to research its own repealed rules; no answer appears in any later recording we transcribed. The year's loudest debate was a rematch, and nobody could remember how the first round ended. That's not a gotcha. It's a lesson about institutional memory as regulatory infrastructure.

Start from a fact most people never think about: testing destroys what it tests. A lab grinds up and dissolves the sample to measure it, so nobody can ever test the exact product you buy. Instead, state scientists and regulators write detailed collection procedures for every product type (flower, concentrates, edibles), all aimed at one goal: the small sample that gets used up in the lab should faithfully represent the whole batch headed to shelves. That's as true for potency as for contaminants like pesticides. Concentrates raise the stakes further: the same process that concentrates the cannabinoids and terpenes concentrates the contaminants right along with them: flower that would pass can become an extract that fails.

Sample adulteration is anything that makes the sample less representative of the batch it's supposed to stand for. And the temptations are built into the system's structure. A producer that has earned a reduced-testing allowance chooses which batches face a lab at all. And every operator has a feel for which batches are likely to pass. Mixing product from a suspect batch into product known to pass by a wide margin is adulteration too. None of this requires a cartoon villain: in the public enforcement record, one Colorado manufacturer told investigators it diluted failing batches until they passed and genuinely believed that was allowed; another surrendered its licenses over product that shipped on a testing exemption it hadn't earned.

Now the thing to hold onto: everyone in this meeting, regulator and industry alike, agreed this is real and ongoing. And note who said the hard parts out loud: it was the Division that surfaced its own repealed sampler rules, and the Division that told the room adulteration "happens pretty consistently," on a range it described as running from sloppy sampling to deliberate concealment of contamination. That is an enforcement agency admitting, in public, that its enforcement isn't deterring the behavior. That openness deserves credit. A manufacturer put it at half of recent enforcement actions and called it a cancer. A year later there is no enforcement proposal on any table. The agreed-upon problem is the one nothing was done about.

Since this meeting

Where the threads this meeting opened stand as of August 2026, from the recordings that followed.

Resolved. The analyte worksheet found its final form: the risk scores were dropped in March after this meeting's where's-the-decision-rule objections, the evidence tiers were rebuilt for toxicology in April, and the finished submission form went live in July. The February 19 technical method review happened, first of a continuing series. And third-party collection got its answer in June: declined, no consensus.

Still in progress. Of the two worksheet pilots commissioned here, chromium reported back in April; the fusarium mycotoxins half does not appear in the later recordings we transcribed.

Put off or neglected. The repealed-rules history was never explained on the record. The enforcement gap this meeting exposed (consistent administrative action, persistent adulteration) drew no proposal all year. And the video-surveillance sampling design a lab representative outlined here survived only as July's fallback proposal, still unadopted.

What was brought up

A worksheet for deciding what to test for [0:10:11]. The Division presented its minimal version: a spreadsheet of required fields for any proposed new analyte (health reasoning, literature, proposed limits, impacts on sample sizes and retesting pathways). Then a manufacturer presented the fuller version he'd built with collaborators: a risk-assessment worksheet modeled on failure-mode analysis, scoring each candidate analyte 1–5 on severity of harm, consumer exposure, and market prevalence, multiplying them into a risk-priority number, then layering context: can labs actually test for it, do non-testing controls exist, what does it cost the system [0:15:14–0:30:23]. The stated goal: turn "you should test for X" from an assertion into a structured argument anyone can interrogate.

The pushback that sharpened it [0:35:24]. One stakeholder objected on principle (health-and-safety criteria should emanate from public health authorities, not industry) and the Division clarified the tool's actual role: last year several proposed analytes died for lack of information, and this defines what a complete proposal contains. A lab representative added a hard prerequisite: the state lab should have a validated method for any analyte before it becomes mandatory, or enforcement is impossible [0:40:25]. Two participants pressed the open wound: scores without a decision rule are still subjective, so where is the line that triggers action? The author's honest answer: there isn't one; the worksheet structures the argument, and the Division still decides [0:45:30, 0:55:38].

Third-party sample collection, four months early [1:05:40]. The Division's presentation carried a piece of institutional memory worth the whole meeting: Colorado had division-approved sampler rules from March 2014 to November 2015, and repealed them, for reasons nobody at the meeting could document [1:15:42]. The national picture: of 40 reporting jurisdictions, 16 allow licensee self-collection like Colorado; 24 require testing-lab personnel to collect [1:20:47]. And the Division put its own position on the record: sample adulteration "happens pretty consistently out there," on a spectrum from sloppy sampling to deliberate concealment of contamination [1:36:02].

The scale claim [1:30:55]. A manufacturer said a review of the last four years of orders to show cause found fully half involved sample adulteration, called sample fraud "a cancer on our industry," and argued it should end licenses rather than draw fines. Other operators countered with the costs: labs would have to become logistics companies, production timelines would stretch, and (the California lesson, offered by a manufacturer who'd interviewed operators there) third-party collection without massive oversight investment "doesn't solve sample fraud, it just pushes it around" [1:25:52, 1:36:02].

Practical designs on the table [1:41:03]. A lab representative laid out the working model from other states: testing-facility-only collection treated as an assay (with mock sampling events like Oregon's), final-form samples only, full batch presented and matched to the inventory system, randomized collection, video surveillance of every sampling event (the evidence that actually shut down California labs), GPS-tracked transport, full chain of custody, and compliance onus on the lab. Another participant noted labs already offer pickup, and that labs have their own client-pressure conflicts, so surveillance of sampling matters regardless of who collects [1:56:09].

What was nailed down

• The technical method review got a date: February 19, 9–11 a.m., hybrid, in-person seats prioritized for technical experts, RSVP required: CDPHE sharing its reference-lab methods for expert scrutiny [1:56:09].
• The agenda reshuffle itself: third-party collection continues as the lead item in March; the proficiency-testing deep-dive moved later into the year [2:01:11].
• The Division's enforcement history went on the record: clear sample containers, inventory-system adulteration reporting, and consistent administrative actions, with the acknowledgment that adulteration persists anyway [1:36:02].
• The state lab's honest status: the rooftop-unit repairs proved bigger than expected, metals testing was down for lack of exhaust, and the lab expected to be back in one to two weeks: to finish building the pesticide and solvent methods, not yet to resume surveillance [1:56:09]. The edibles report: still in review, promised by March.

What was left open

• The decision rule. The risk worksheet scores analytes; nobody could say what score means "add it." The Division took "map the score to a pathway" back as feedback [0:55:38].
• Why 2014's sampler rules died. The Division offered to research its own repealed rules; a cultivator asked for that history before the March discussion [1:46:07].
• Whether third-party collection addresses interlab consistency at all. The meeting's sharpest analytical point: it addresses sample fraud; a fraudulent sample makes labs look inconsistent, but honest labs testing honest samples can still disagree. The state lab's bridge: "the test result is only as good as the sample" [1:25:52, 1:41:03].
• Cost, geography, and survival: rural and Western Slope logistics, who pays, and whether an industry with two remaining testing labs can absorb any of it [1:36:02, 1:51:08].
• Re-evaluating existing analytes: raised again (with Oregon's dropped Aspergillus testing as the example); CDPHE pointed to its 2024 testing-rationale document and left the door open [0:55:38].

This summary was drafted by our AI desk from a machine transcript and has been reviewed and approved by our editor (2026-08-13). The Division's published recording is the source of record.