Source Matters: An Examination of Drug Checking Samples from Police Departments and Community Based Programs in Massachusetts

Aims and SettingIn the U.S., the emergence of new adulterants and novel psychoactive substances continues to complicate approaches to overdose, treatment, and public safety. Information about this changing drug supply is often gleaned from police drug seizures, but community drug checking services, which test the contents of a persons drug supply and share that data, provide another means to understand local drug supplies. However, it is unclear how seized drugs differ from those collected in the community, whether one approach is potentially more instructive, and what can be learned about local drug supplies from each source. We therefore compared drug samples tested from police departments (PDs) and community partner (CP) drug checking programs to examine what, if any, differences existed in sample content, form, submitter characteristics, and emerging substance presence. DesignWe conducted a retrospective cohort analysis of drug samples collected and tested between April 2018 and December 2025 by the Massachusetts Drug Supply DataStream derived from CPs and PDs operating in the same geographic area across eight locations. Bivariate analyses (Chi-square, Fishers exact) tested for differences in sample and submitter characteristics by source. FindingsThere were 2,430 unique samples submitted by CPs (68.1%) and PDs (31.9%) from the same location. Compared to CP samples, proportionally more PD samples showed fentanyl as primary substance (74.2% PD vs. 64% CP, p<.001) and less often contained additives (xylazine 15.0% PD vs. 27.4% CP; medetomidine 0.6% PD vs. 2.2% CP, both p<.001). PD samples were typically powders (73.2% vs. 37.9%) and pills (13.6% vs. 3.6%) while CP samples were more often residue (51.9% vs. 2.1%, p<.001). Submitter characteristics, when reported, differed by source: gender (n=528, male: 78.6% PD vs. 50.1% CP, p<.001), race/ethnicity (n=468, Black: 15.8% PD vs. 7.8% CP; Hispanic: 6.7% PD vs. 13.2% CP, p<.05), and associated overdose (n=242, fatal: 62.9% vs. 10.9%, p<.001). Emergent substances were detected a median of 249 days sooner in CP than co-located PD samples, though drugs exhibiting concerning patterns (e.g., unexpected fentanyl in stimulants) had similar, swift detection times. ConclusionDrug samples differ based on PD vs. CP source in significant ways that may introduce bias when drawing conclusions about drug supply trends but also offer unique insights for public health and responses to emerging drugs. Modern drug monitoring should include a broad range of sources to best prepare for changes the illicit supply may bring to overdose prevention, public safety, and health systems.