We have received a lot of great feedback from programs on how to improve the data collections elements on the card. We have noted these over the last months, and are ready to make changes. (The card was last updated in June 2023.) We're giving analysts 2 months lead time to get ready. Given all the analysis implications listed below, we do not want to make another card change for a long time! We will update this page with the final version prior to launch.

You can find an archive of previous card versions and launch dates at the UNC institutional repository. For this round of changes, we did a data analysis of 8,311 samples analyzed by October 2024 to see patterns of missingness and see if we could make circle choices for common write-in notes for sensations, texture and expected drugs.

The current codebook may be helpful in understanding the analysis implications. The codebook will be updated in January in advance of the launch. The following changes are expected to roll-out in mid-to-late January 2025, as card version 9. Card version numbers are noted in the bottom left of the card.

Card Changes

Big Things

• Added nitazene and carfentanil to list of expected drugs
• Added stimulant / opioid type to OD panel
• Added test strip results for benzos, fent, xylazine, and meth

Little Things

• Changed scoop visual for collection method to be scoop and not spatula
• Added “important!” to collection method to get more completion (this is very important for the lab to know how to analyze the sample, such as needing to concentrate residue samples)
• Added “fine” to texture (common write-in)
• Added “rock” to texture (common write-in)
• Changed “weird” to “unusual” in sensations because the former was considered pejorative by some participants
• Changed “unusual taste” to “odd smell" and "odd taste” in sensations to differentiate it from "unusual" above
• Removed “fake pill” and kept “pill”
• Changed “normal” to “typical” in sensations

Analysis Implications

These are some anticipated downstream implications of the card changes on data.

Analyzing sensations, datasets will be retrospectively processed to aggregate “unusual” or “weird” to be "unusual" only.

Analyzing sensations, "smell" and "taste" can be disaggregated in sensations.

Analyzing sensations, “normal” or “typical” will be aggregated. We will update datasets to replace all “normal” with “typical” retrospectively in variable sen_strength and sensations. Variable sen_weird will be renamed to sen_unusual.

In sensations, “more down” and “sedating” were redundant, so only “sedating” was kept to make room on the card. Variable sen_down will remain as it is but will include both “more down” and “sedating” responses.

Variable overdose will stay the same, but od_type will be added to distinguish stimulant, opioid, and no response.

Analyzing texture, “rock” was commonly listed as a write-in “other” category, and now will be in the main description text of sensations. This will be automated retrospectively to make it consistent with earlier datasets so no additional analysis to search free text (variable texture_notes) . Same with “fine.”

Test strip results will be a new set of 4 verbatim variables for test_benzos, test_fent, test_meth, and test_xylazine. The variables will take values of -1 for circled negative, 1 for circled positive, missing if not specified.

We hope these changes will help our participants describe their experiences better, reduce hassle for staff filling out the cards, and provide analysts with tools for more nuanced analysis.