An Evaluation of Data from Drivers Arrested for Driving Under the Influence in Relation to Per se Limits for Cannabis

AAA Foundation for Traffic Safety · 2016 · AAA Foundation for Traffic Safety

archive: archived pipeline: cataloged verified

Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)

Summary

This 2016 report by the AAA Foundation for Traffic Safety evaluates whether scientific data supports establishing a quantitative "per se" legal threshold for delta-9-tetrahydrocannabinol (THC) in the blood to define impaired driving. Motivated by the increasing legalization of cannabis and legislative efforts to criminalize cannabis-impaired driving, the study aimed to determine if specific blood THC concentrations correlate with observable signs of impairment. The authors sought to identify if a numerical cutoff could reliably distinguish between impaired and non-impaired drivers, thereby providing a scientific basis for per se laws similar to those used for alcohol. The researchers analyzed two distinct datasets. The first comprised 602 drivers arrested for impaired driving with THC as the sole detected substance, who underwent Drug Recognition Expert (DRE) assessments, alongside a control group of 349 drug-free individuals. These assessments included physiological indicators (e.g., pupil size, pulse, blood pressure) and psychophysical tests (walk-and-turn, one-leg-stand, and finger-to-nose). The second dataset included 4,799 drivers arrested for DUI who tested positive for cannabinoids, providing demographic and toxicological data without behavioral assessments. Statistical analyses, including bivariate correlations, chi-squared tests, and logistic regression, were used to examine the relationship between THC concentrations and impairment indicators. The study found significant differences between cannabis-positive drivers and drug-free controls in both physiological and psychomotor performance. Cannabis-positive subjects exhibited higher rates of red eyes, eyelid tremors, dilated pupils, elevated pulse, and higher systolic blood pressure. They also performed significantly worse on field sobriety tests; for instance, only 6.0% of cannabis-positive drivers completed the walk-and-turn test without errors, compared to 55.5% of controls. However, when examining the correlation between blood THC concentration and these impairment indicators, the study found minimal association. Neither the walk-and-turn nor one-leg-stand tests showed increasing error rates as THC concentrations rose from 1 to 47 ng/mL. Only the finger-to-nose test showed a correlation with higher THC levels. Logistic regression demonstrated that DRE indicators could not predict whether a driver’s THC level was above or below the 5 ng/mL threshold adopted by several states, nor did any indicator meet an 80% sensitivity threshold for correctly identifying impairment status. The analysis of the larger arrest population revealed that the median THC concentration was 4.0 ng/mL, meaning half of all arrested drivers had levels below the 5 ng/mL per se limit. Among drivers positive only for cannabis, the median was 7.8 ng/mL. The study concluded that no objective THC threshold exists that reliably establishes impairment based on the collected data. While a 1 ng/mL threshold offered the highest sensitivity (80.4%), it also had low specificity, resulting in many non-impaired drivers being classified as impaired. Consequently, the authors determined that a quantitative per se limit for THC cannot be scientifically supported, as blood concentrations do not consistently correlate with the degree of observed impairment.

Key finding

No blood THC concentration—including the 5 ng/mL per se thresholds used in several states—reliably distinguished impaired from non-impaired DRE/SFST performance; DRE indicators could not predict THC above 5 ng/mL, and only 30–49% of arrested drivers would exceed that threshold.

Methodology

modeling

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_aaa_foundation on 2026-05-23 (5 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified_with_issues.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• cannabis
• illicit drugs
• polydrug
• dui enforcement
• alcohol