Roadside oral fluid tests for cannabis still don't meet accuracy standards | Research

The takeaway

A systematic review of 15 studies found that none of the current roadside oral fluid testing devices meet the minimum 80% threshold for sensitivity, specificity, and accuracy recommended by international research projects.

Cannabis users who drive, policymakers, and anyone interested in roadside drug testing accuracy.

Zero out of nine roadside oral fluid devices met the 80% accuracy benchmark

What the researchers found

Nine oral fluid testing devices were evaluated, and none met the minimum 80% benchmark for sensitivity, specificity, and accuracy set by the ROSITA, ROSITA-2, and DRUID projects. Devices detecting lower THC concentrations performed better than those with higher cutoffs.

Why it matters

As more jurisdictions legalize cannabis and need to enforce impaired driving laws, reliable roadside detection is critical. This review shows the technology isn't there yet, creating a gap between legal frameworks and enforcement capability.

The numbers in context

15 studies reviewed. 9 devices evaluated. Target benchmark: 80% for sensitivity, specificity, and accuracy. None of the devices met all three criteria. Lower THC detection thresholds correlated with better performance.

How the study worked

Systematic review searching eight databases (PubMed, Web of Science, MEDLINE, Engineering Village, Embase, Compendex, CINAHL, Scopus) for studies evaluating on-site oral fluid drug screening devices for cannabis detection. Fifteen articles were selected for review.

What this study cannot tell us

Lack of standardized test protocols across studies made comparison difficult. Confirmation analyses varied between blood and oral fluid. Many subjects had used multiple drugs. Device technology may have improved since the studies reviewed.

How to read the evidence

Moderate: systematic review of 15 studies across eight databases, but limited by inconsistent study protocols.

When this study was published

Published in 2019.

The bigger picture

Unlike alcohol breathalyzers, which are well-established, cannabis roadside testing remains unreliable. This has implications for both public safety and civil liberties, since inaccurate tests can either miss impaired drivers or falsely accuse sober ones.

Questions still open

Can devices with lower THC detection thresholds be optimized to meet the 80% benchmark? Should THC concentration be the primary target, or should tests focus on functional impairment? How do environmental factors like passive exposure affect oral fluid results?

Common questions

How accurate are roadside cannabis tests?

According to this review, none of the nine devices tested met the minimum 80% standard for sensitivity, specificity, and accuracy. The technology is not yet reliable enough for consistent roadside enforcement.

Why is testing for cannabis harder than testing for alcohol?

THC behaves differently than alcohol in the body. It can linger in oral fluid and blood long after impairment wears off, and there's no established concentration that reliably corresponds to impairment level.

Read the original research

Are oral fluid testing devices effective for the roadside detection of recent cannabis use? A systematic review.

Public health, 171, 57-65

View on PubMed ↗Publisher’s paper ↗

Citation

Dobri, S C D; Moslehi, A H; Davies, T C. (2019). Are oral fluid testing devices effective for the roadside detection of recent cannabis use? A systematic review.. Public health, 171, 57-65. https://doi.org/10.1016/j.puhe.2019.03.006