Good evening. I appreciate the opportunity to share some information as you consider Bill C-46, an act to amend the Criminal Code.
I'm Tom Marcotte. I'm a professor of psychiatry at the University of California San Diego, and co-director of the University of California Center for Medicinal Cannabis Research. I'm an investigator on two current studies examining the impact of cannabis on driving.
Today I'd like to provide some background on the challenges in determining whether an individual's driving has been impaired by cannabis.
In controlled simulator and on-road studies, it's been well established that acute cannabis intoxication results in slowed reaction times, including delays in braking, reduced ability to maintain one's lane position—in other words, swerving—and reducibilities relating to the judgment of speed and distances. The effects of cannabis are amplified by alcohol, although it's not resolved as to whether this is an additive effect or synergistic, in which the two combined are worse than simply adding the effects together. Also, in contrast to alcohol, cannabis users are more likely to judge themselves to be impaired and to adjust behaviour, by driving more cautiously, as one example. However, of course, this is not universal.
Findings from the real world have been mixed. Some studies have found a twofold increase in crash risk when THC is present, while other studies have found no increased risk once adjusting for factors that often travel with cannabis use and risky driving, such as younger age and being male.
Here's one example of the difficulty in interpreting crash results from the states that have legalized cannabis.
In Colorado, it was widely publicized that there was a dramatic 50% increase in the number of fatalities in which marijuana was present following legalization. However, as seen in this next graph, there was only a marginal increase in the total number of crashes in that same period. This mirrored recent data demonstrating that, at a national level, there was also an increase in fatal crashes.
What is clear is that at this same time, the State of Colorado increased the amount of screening they were doing to detect THC. Therefore, it is unclear whether the increased prevalence of fatalities with THC present represents a situation in which increased cannabis use might have led to more fatalities, or whether it is primarily a case that authorities are more frequently looking for the presence of cannabis and finding it.
On the other hand, a recent report has indicated that there has been an increase in insurance collision claims in states where recreational cannabis has been legalized compared with other states. These are the much more common non-fatal crashes. When examining claim rates in Colorado, Washington, and Oregon, the authors found a 3% increase in claims relative to states that did not legalize use, with there being some variability between the states.
What might be some of the reasons that we see significant effects during controlled studies but a more modest effect in the real world? There are a number of possibilities, but to name just a few, in part, epidemiologic findings are based upon imperfect data. For example, the fatality reporting system in the United States often has incomplete reporting, and there's typically a significant delay between the time of a crash and the collection of blood. In addition, THC can be detectable in the blood long after the impairing effects have resolved. Thus, the impact of acute intoxication may not be readily apparent in these analyses, since the THC-positive group includes a much larger number of individuals who might have smoked much earlier and were not impaired at the time that the blood was collected.
On the other side, it is also possible that in some of our studies, while we're able to detect acute effects of cannabis on tasks such as swerving, they may not be of significant magnitude to dramatically affect real-world driving. As an example, in a study of low-dose THC for the treatment of spasticity in multiple sclerosis, we found a significant effect on driving two to three hours after dosing. However, the magnitude of that effect was not dissimilar to what other studies have found for individuals in the initial phases of starting antidepressants, or the residual morning-after effects of taking a sleeping medication the night before.
Drug recognition evaluations are the current gold standard for establishing substance-impaired driving. We're currently in the midst of a large study, funded by the State of California, to better characterize the impact of cannabis on driving, and to investigate whether there are additional effective approaches to identifying those individuals who are or are not impaired due to cannabis.
As part of this study we're working with DRE instructors to explore the validity of select components of the DRE evaluation, as well as assaying for the presence of THC, its metabolites, and other cannabinoids to determine whether they might provide reliable information regarding the time since the participant smoked or, ideally, relating to driving impairment.
Another unique aspect of this study is that we are utilizing novel iPad-based assessments to see if such tests might serve as a useful adjunct to the DRE evaluation. Unlike alcohol, where impairment readily presents itself physiologically, such as staggering and difficulty walking, cannabis effects are primarily cognitive and a current DRE evaluation includes only modest assessments of these abilities.
Particularly relevant to Bill C-46, studies to date raise concerns regarding the validity of using THC levels in blood to identify cannabis-impaired drivers. For example, a study by the American Automobile Association examined 602 cases in which DREs have identified drivers as being impaired, with THC being the only substance identified in the blood.
In this graph, the level of THC runs across the x or the horizontal axis and a per cent of drivers with that THC level is represented on the y or the vertical axis. As you can see in these impaired drivers, there was a wide range of THC levels. The median value or number where half the drivers were above and half the drivers were below was around five, indicating that 50% of these impaired drivers had values below the five nanograms per millilitre cut point at the time the blood was drawn. Thus, drivers can be impaired, yet have THC blood levels below a cut point that some governments have chosen as being indicative of driving under the influence.
Conversely, the table on the left shows that individuals who are likely unimpaired can also have detectable THC levels in their blood, even days after smoking. In this case, participants stayed in a hospital for 30 days so they could be monitored for any cannabis use. They then smoked cannabis and blood was subsequently drawn each day. As you can see in this table, some individuals were registering values of two nanograms per millilitre of THC, even though it had been up to a week since they smoked.
Why is it that we can have individuals with low levels of THC who are impaired, as well as individuals with low levels who are not impaired? The graph on the right is from Dr. Marilyn Huestis, a researcher in cannabis pharmacodynamics. Across the bottom we see THC levels and on the side we see, in essence, how high the person is feeling. This figure shows time in a counter-clockwise fashion, so as you see 1.8 minutes is the first and second is 4.5 minutes and so forth. After smoking, THC levels rise very rapidly so they reach a peak in about 10 minutes. At the same time the person is increasingly feeling high, so you see going to the right it's increasing, but it's also going up, so they're feeling higher. At this point, however, THC levels begin dropping to the point where about an hour after smoking they're now down to fairly low levels as you move across to the left in this graph.
The person, though, is still feeling high during this time. A few hours after smoking the highness starts diminishing, so it starts dropping down the vertical, but THC levels are not changing dramatically during this period. As you can see, it's between zero and 10. This tells us that someone can be high with elevated THC levels, someone can be high with modest levels, someone can be high with low levels, and someone can also have low levels and not be high. To further complicate this, Dr. Huestis has demonstrated that these patterns vary, depending upon whether one is a frequent or infrequent smoker.
At least for screening, oral fluid instruments hold some promise, they're easy to administer, relatively non-invasive, and may help identify individuals who recently used cannabis. This approach, however, is also not without complications. This graph shows results from a study of oral fluid THC levels in individuals who smoked a 6.8% THC cigarette. More studies are needed and ours is assessing the issue, but in general it's believed that the most significant impairing effects happen within the first few hours of smoking and then dissipate over the following few hours.
As you can see here in this graph, however, at least in this one study, a proportion of individuals were at or above a five micrograms per litre cut point in oral fluid eight to 10 hours after smoking.
I mentioned earlier that we have a study going on. If the group is interested, during the discussion I'd be happy to provide more details, but for this purpose I'll skip it and just end with a few concluding points.
Per se laws can be very effective, but this is particularly true when there is a robust relationship between fluid levels and actual driving impairment, as there is with alcohol. As can be seen in some of the data presented earlier, I don't think this is yet the case for cannabis. I'm also aware from attending many meetings that prosecutors remain concerned that a cut point designating impairment may lead the public to assume that a driver below that cut point is not impaired or is less impaired. As seen in the DRE data I presented earlier, low levels do not necessarily mean low impairment.
Some individuals have also expressed concern that the DRE evaluations may not be adequately sensitive to the effects of cannabis and that one should use fluid levels to identify impairment. I would argue that it is very important to continue to use behaviour as a key indicator of driving-related impairment given the uncertainty in interpreting fluid levels.
Last, I encourage you to support additional research into identifying new methods that might help law enforcement identify both those who are impaired and those who are not impaired due to cannabis. This includes biological, psychophysical, and behavioural approaches.
As you know, the complexities associated with detecting cannabis-related driving impairment also have increased our awareness regarding the continuing problem of impairment due to prescription medications. Perhaps new approaches to detecting impaired driving would end up being applicable to these drug classes as well.
Thank you, and I'm happy to take any questions.