Charles Miller Award 2024 for Top Research and Evaluation Stream

Impairment Identified by the Standardized Field Sobriety Test (SFST) after Smoked Cannabis with and without Alcohol: Relating SFST Results to Simulated Driving Performance

By: Christine Wickens, Ph.D., Senior Scientist at the Centre for Addiction and Mental Health (CAMH) and Associate Professor at the University of Toronto

Background

Cannabis has recently matched or surpassed alcohol as the psychoactive substance most frequently found in drivers seriously or fatally injured in a motor vehicle collision. Meta-analyses of epidemiological and experimental data have confirmed that cannabis impairs driver behaviour and performance. Standardized Field Sobriety Tests (SFSTs) were initially developed based on laboratory research on the effects of alcohol on human performance. Research on the ability of SFSTs to identify individuals impaired by cannabis is emerging from laboratories around the world, producing mixed results. Fares et al. (2022) reported the results of a randomized clinical trial examining the effects of alcohol and smoked cannabis, separate and combined, on simulated driving. In that study, alcohol alone resulted in significant increases in standard deviation of lateral position (SDLP) compared to placebo and a significant increase in standard deviation of speed and maximum speed when compared to placebo and cannabis alone. Cannabis resulted in significant increases in SDLP in single- and dual-task conditions, but did not impact speed measures. The combination of both drugs led to a significant increase in SDLP in both single- and dual-task conditions.  

Aims

Fares et al. (2022) included administration of the SFST following simulated driving. The current analyses aimed to: (1) examine impairment identified by the SFST following administration of cannabis and alcohol, separate and combined, under controlled laboratory conditions, and; (2) relate SFST results to driving simulator performance.  

Methods

In a within-subjects, double-blind, double-dummy, placebo-controlled, randomized trial, 28 regular cannabis users (1–7 days/week; 16 males, 12 females) aged 19–29 years attended four drug administration sessions in which simulated driving was assessed and the SFST administered following placebo alcohol and placebo cannabis (<0.1% THC), alcohol (target BrAC 0.08%) and placebo cannabis, placebo alcohol and active cannabis (12.5% THC), and alcohol and active cannabis. Simulated driving 45 min after drug exposure was assessed under both single- and dual-task conditions. The SFST included the one-leg stand (OLS), walk-and-turn (WAT), and horizontal gaze nystagmus (HGN) tests, and was administered by two trained personnel to permit assessment of inter-rater reliability.  

Results

Inter-rater agreement was strong in all drug conditions and in all SFST tests, with agreement ranging from 72.5% for HGN in the alcohol-cannabis condition to 95.2% for HGN in the placebo condition and with most agreement values exceeding 80.0%. Cohen’s Kappa scores were also generally strong, with most values demonstrating moderate (κ>.40) or substantial (κ>.60) agreement. The percentage of participants in each group failing the SFST was: 15% after placebo, 48% after alcohol, 19% after cannabis, and 63% after alcohol-cannabis. Friedman tests were conducted for each SFST test to detect an overall effect of drug condition on test scores; significant effects were found for all SFST tests (most p<.001). Utilizing confidence intervals and error bar plots to assess differences between treatments, higher mean HGN and total SFST scores were found in the alcohol and alcohol-cannabis conditions compared to the cannabis and placebo conditions. Although, mean WAT scores were similar across conditions, mean OLS scores were higher in the alcohol-cannabis condition than all other conditions. No differences were found between the cannabis and placebo conditions. Wilcoxon signed rank tests were conducted to compare driving performance 45 min after drug (i.e., alcohol, cannabis, alcohol-cannabis) versus placebo among those participants who passed and who failed the SFST. Among significant findings, SDLP was higher after drug than placebo among both participants who passed and who failed the SFST in both the cannabis (fail: Z=-2.032, p=.042; pass: Z=-2.096, p=.036) and alcohol-cannabis (fail: Z=-3.336, p<.001; pass: Z=-2.213, p=.027) conditions. Spearman pairwise correlations between SFST score (overall and subtests) and driving variables (change from baseline and post-drug) were analyzed in each drug condition. When found, significant correlations were always in a positive direction and typically moderate in size, with most rs values ranging from .41 to .53. In the placebo and alcohol conditions, changes in driving variables from baseline were not significantly associated with overall SFST score or any of the SFST subtests. However, correlations between post-drug SDLP in both single- and dual-task conditions were correlated with WAT and total SFST scores. Additionally, in the placebo condition only, mean speed under single-task instructions was correlated with WAT scores. In the cannabis condition, there were no correlations between SFST scores and post-drug driving variables. However, there were correlations between changes in driving variables from baseline (mean, maximum, standard deviation of speed) and HGN. In the alcohol-cannabis condition, change in SDLP from baseline was correlated with OLS, HGN and total SFST scores under dual-task instructions and with OLS under single-task instructions. Among the post-drug driving variables, SDLP under dual-task instructions was correlated with OLS and total SFST, and standard deviation of speed under single-task instructions was correlated with HGN.  

Discussion

Although significant differences in SDLP were found between the cannabis alone and placebo conditions, the SFST was not able to detect this impairment. The percentage of participants in each group who failed the SFST, and the HGN, OLS, and Total SFST scores were higher in the alcohol alone and alcohol-cannabis than cannabis alone and placebo conditions, but no difference in SFST (overall or subtest) was found between the cannabis and placebo conditions. These results suggest that the SFST may lack sensitivity to cannabis impairment, which is consistent with several existing studies. SDLP scores were higher after drug than placebo in both the cannabis alone and alcohol-cannabis conditions, but this was true among those who passed and failed the SFST. Positive correlations between SFST scores and measures of simulated driving impairment were found, but within the cannabis condition involved speed-related variables as opposed to SDLP and were exclusive to HGN.  

Conclusions

These findings do not provide strong support for use of the SFST for detection of THC impairment. Research examining the inclusion of other tests of impairment (e.g., finger-to-nose test, Modified Romberg test) or revised scoring to improve sensitivity and specificity for detection of THC impairment has shown promise and is recommended.

Bio: 

Dr. Christine Wickens is a Senior Scientist at the Centre for Addiction and Mental Health (CAMH) and holds an appointment as Associate Professor at the University of Toronto. Through experimental studies, analysis of population-level survey data, and qualitative analysis of interviews and archival data, Dr. Wickens has contributed to knowledge of topics including alcohol- and drug-impaired driving, driver anger and aggression, distracted driving, street racing, and the impact of mental health on driving. She has published more than 110 peer-reviewed journal articles and presents her work at conferences nationally and internationally. Dr. Wickens serves on the Board of Directors for CARSP and for the International Council on Alcohol, Drugs and Traffic Safety.