Cannabis, crashes and blood: challenges for observational research
Bibliographic record
Abstract
The cannabis crash responsibility study by Brubacher et al. 1 is timely, considering that cannabis for recreational use became lawful in Canada on 17 October 2018 2, raising expectations of increased adult use 3. Some governmental agencies prohibit cannabis use for 24 hours, 28 days or ever before safety-sensitive work 4-7. The occupational medicine recommendation is 24 hours (or longer if impaired) 8. Canadian impaired driving law uses a two-tier system of concentration per-se limits for delta-9-tetrahydrocannabinol (THC) in blood: greater offence at 5 ng/ml THC or 2.5 ng/ml in combination with 0.05 g% ethanol; lesser offence at 2 ng/ml THC 9. International discussion has not produced scientific consensus on what threshold THC concentration should be enforced, varying from 0.2 ng/ml (Sweden) to 5.0 ng/ml (Colorado) 10. For alcohol, per-se limits work because of distribution throughout the total body water compartment, with good correlations between blood and brain concentrations, degree of impairment and crash risk 10. Lipophilic THC does not have well-established correlations: after smoking, maximum blood THC concentration is reached by approximately 10 minutes, then drops precipitously by 90% in 1–2 hours, with low concentrations measurable for up to 12 hours or longer 11. THC clears the bloodstream but accumulates in lipids and brain tissue 12, causing dose-dependent subjective effects for up to 8 hours 13, 14 (10 hours with oral administration 14, 15) and cognitive effects up to 24 hours 15. Because of rapidly declining THC concentration, a driver might have been substantially above a per-se limit at the time of driving or crash but fall below the limit by the time of blood draw 16. Postmortem blood THC concentrations more accurately reflect concentration at the time of crash, because circulation and metabolism ceases with death. In a study of traffic fatalities, blood THC concentrations of at least 5 ng/ml in killed drivers yielded an odds ratio (OR) of 6.6 for crash responsibility 17. Meta-analysis of cannabis crash studies calculated a non-statistically significant OR in non-fatal crashes but a statistically significant OR of 2.1 in fatal crashes, including both killed and surviving drivers 18. The underpowered study by Brubacher et al. does not demonstrate a statistically significant increase in OR, as only 20 of 1825 analyzed subjects had 5 ng/ml or more THC in blood, probably reflecting delayed sample collection. Nevertheless, the authors claim that their findings ‘suggest that the impact of cannabis on road safety is relatively small at the present time’ (p. 1621), despite $1.09 billion in costs, 75 collision deaths and 4407 injuries attributed to Canadian cannabis-related collisions in 2012 19, and the fact that data collection ended almost 2 years before recreational cannabis became lawful. Like many observational studies, Brubacher et al. rely on blood THC concentrations in living drivers without driver-specific data determinative of impairment and crash risk, such as THC dose, route of administration, time of last use and cannabis use history. True crash risk associated with cannabis use cannot be understood based solely upon low THC concentrations in the blood of living drivers.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.259 | 0.544 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.005 | 0.003 |
| Bibliometrics | 0.008 | 0.014 |
| Science and technology studies | 0.004 | 0.006 |
| Scholarly communication | 0.006 | 0.012 |
| Open science | 0.007 | 0.007 |
| Research integrity | 0.005 | 0.007 |
| Insufficient payload (model declined to judge) | 0.009 | 0.003 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; the direct Gemma label and the distilled Codex classifier agree on what is shown here.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".