The HGN Comparison Did Not Validate a Standalone Cannabis Impairment Test
The HGN analysis compared small groups defined by self-reported duration. It did not establish a standalone diagnostic accuracy for identifying cannabis-impaired drivers.
What the statistical comparison involved
Forty-four participants underwent horizontal gaze nystagmus assessment, with forty-three examined at three hours after smoking. The paper compared HGN incidence at that time across duration groups. A reported comparison between the one-hour group and the greater-than-three-hour group involved fourteen and five participants, respectively, and yielded p = 0.0491.
A group difference is not sensitivity and specificity
The result concerns an association between HGN and the study’s duration categories. It does not supply the true-positive and false-positive rates of a roadside test against an independent validated driving-performance reference. The category labels themselves relied on participants’ assessments, which should not be silently replaced with measured crash risk or on-road impairment.
Use the finding at its actual scale
The small group sizes, selected participants, smoking conditions and observation window limit extrapolation. A reviewer should retain the exact comparison and avoid treating statistical significance as proof that HGN detects every affected person or excludes impairment when absent. The paper’s disclosed commercial interests and overlapping earlier study data are also relevant to evaluating independent validation.
Related analysis: baseline and post-smoking breath findings.
Research cited
Wurz and DeGregorio (2022). Indeterminacy of cannabis impairment and THC levels in blood and breath. DOI: 10.1038/s41598-022-11481-5.
