What has marijuana legalization done to Driving Under the Influence (DUI) cases in Colorado? Important answers are now emerging. July 2018, the Division of Criminal Justice (DCJ) released a 106-page report Driving Under the Influence of Drugs and Alcohol for the year 2016. This first-of-its-kind report and its database enabled us to draw twelve important conclusions.
1 Colorado’s report is the first to show causes of DUI, not just presence of drugs
Coroners routinely test fatal crash victims for presence of alcohol and drugs. That makes it easy to report the presence of impairing substances in fatal crashes, which is why we see so many reports of “marijuana-involved” fatal crashes, primarily based upon coroner reports. Reports from the Washington Traffic Safety Commission and Rocky Mountain High Intensity Drug Trafficking Area are recent examples of such studies which show excellent trend graphs. But since coroner’s subjects are never charged with or convicted of DUI, such reports are often criticized on the grounds that mere presence of drugs does not prove impairment.
Colorado’s report was created by linking forensic toxicology data with the state’s judicial data of DUI charges and convictions. It is the first large-scale report to show the causes of all DUI charges, not simply the presence of drugs in fatal crashes. This is a huge step forward to understanding DUI and drugged driving.
2 We can’t understand our drugged driving problem if we don’t test for it.
As the DCJ report repeatedly pointed out, Colorado does a terrible job testing drivers arrested for DUI, which limits the power of the report. Of the 27,244 cases of DUI charges in 2016, only 8.7% were tested for both alcohol and drugs, and the test methods and reports were not uniform. Nevertheless, the following conclusions are evident from the 2016 data.
3 Polydrug impairment was far more common than marijuana impairment.
Of the 2,383 cases tested for both drugs and alcohol, 2,340 had at least one intoxicant. Less than one-third (32.6%) of those positive cases had a single intoxicant, either alcohol or a single drug. The balance (67.4%) were polydrug cases, where drivers were positive for multiple drugs. Only 5.8% of the 2,340 cases were positive for marijuana’s THC only. THC means only delta 9-THC, not its inactive metabolite.
4 Marijuana accounted for a little more than half of all drug impairment
There were 2,885 THC-positive cases, and 2,246 cases positive for other drugs. By that measure, THC constituted 56.2% of the drug-impaired (non-alcohol) cases. For cases of a single drug only (excluding both polydrug and alcohol cases), THC constituted 61.1% of the cases. Clearly, drugged driving is not just about marijuana.
5 Drug users were more likely than drinkers to be polydrug users
90% of drug users and 89% of marijuana users were polydrug users. 73% of drinkers were polydrug users. The latter number is likely an exaggeration of the facts, since the number is based on drivers tested for both drugs and alcohol. Police test drivers for drugs only if they have a good reason for drug testing. Therefore drivers only on alcohol were likely under-represented in the sample. A broader database would likely lower this number. For example, Washington’s study of fatal crashes found that over 70% of drug users (including marijuana users) were polydrug users, but only 39% of drinkers were polydrug users.
6 Drunk drivers were far more likely to be convicted of DUI than drugged drivers
93.3% of drivers who had a BAC above .05 were convicted, compared with 83.8% of drugged drivers. These numbers include polydrug cases, and also include convictions for both DUI as well as DWAI. Colorado has two levels of impaired driving charges: DUI (substantially incapable of safe driving) and Driving While Ability Impaired, or DWAI (impaired to the slightest degree, less capable of safe driving).
7 Marijuana-impaired drivers were less likely to be convicted than others
93.3% of drunk drivers and 84.7% of drivers testing positive for non-marijuana drugs were convicted. But only 82.8% of drivers testing greater than 1 ng/ml THC (the lowest forensic laboratory reporting limit) were convicted. This is striking since Colorado’s 5 ng/ml THC permissible inference law was intended to make THC convictions easier. That doesn’t appear to have happened.
8 Stoned drivers over 5 ng/ml THC were impaired, yet only 60% were convicted of DUI
It is important to look at the small minority of drivers who tested positive for only THC or alcohol. This view completely avoids any question about whether or not it was the THC that caused impairment, or alcohol or another drug in a drug “cocktail” that caused impairment. For cases of THC only that were above 5 ng/ml, 99.7% were found guilty of DWAI. They were proven to be impaired. Yet only 59.8% were convicted of DUI. The only reason that overall THC-positive drivers had a higher conviction rate (82.8%) was because of the 88.4% conviction rate for polydrug impairment, not because of Colorado’s 5 ng/ml permissible inference law.
Table 1 Convictions by single drug
| Convictions by drug | Case Count | ||
| DUI | DWAI | ||
| Alcohol only | 13,323 | ||
| ≥BAC .08 | 92.5% | 99.9% | 11,857 |
| BAC .05-.079 | 23.6% | 85.0% | 1,189 |
| THC only | 878 | ||
| ≥5 ng/ml | 59.8% | 99.7% | 621 |
| 1.0-4.9 ng/ml | 14.1% | 91.4% | 241 |
- Colorado DUI/DUID adjudicated cases in 2016
- No polydrug cases included
- Convictions include Guilty, Deferred, Deferred/Dismissed
- DUI: substantially incapable of safe driving
- DWAI: impaired to the slightest degree, less safe
Incidentally, Table 1 also proves that police were able to accurately identify marijuana-impaired drivers even without the support of the preliminary breath testers often used for alcohol impairment. It also proves prosecutors were successful prosecuting impaired driving.
9 Stoned drivers under 5 ng/ml THC were impaired, yet only 14% were convicted of DUI
For THC only drivers who test below 5 ng/ml, the results were even worse. 91.4% were convicted of DWAI, proving that they were impaired. Yet only 14.1% were convicted of DUI. The low DUI conviction rate would be acceptable if the drivers were not impaired, but DWAI convictions prove otherwise. In contrast, drunk drivers below alcohol’s .08 gm/dl ‘legal limit’ had a 67% higher DUI conviction rate than stoned drivers below THC’s 5 ng/ml ‘legal limit,’ even though drunk drivers had a lower conviction rate of DWAI than stoned drivers.
10 Victims of THC-impaired drivers under 5 ng/ml did not see justice
Fortunately traffic injuries are rare and traffic fatalities are even more rare. In 2016, half of the twenty DUI vehicular homicides were charged with DUI caused by alcohol only, the other half were caused by various drugs or drug and alcohol combinations. Only two were charged with vehicular homicide due to driving under the influence of THC only. Both cases tested below 5 ng/ml and neither driver was found guilty. In contrast, five were charged with vehicular assault due to driving under the influence of THC only, and four were found guilty. All were above 5 ng/ml. The vehicular homicide and assault numbers are too small to be conclusive, but they are fully consistent with the above observations that were based on larger numbers.
Table 2 Vehicular Homicide/Assault-DUI Convictions
| Cases | Number | Rate | Note |
| VH-alcohol only | 10 | 80% | |
| VH-THC only | 2 | 0% | All < 5 ng/ml |
| VH-single other drug only | 1 | 0% | |
| VH-polydrug | 7 | 71% | |
| VA-alcohol only | 79 | 75% | |
| VA-THC only | 5 | 80% | All > 5 ng/ml |
| VA single other drug only | 3 | 67% | |
| VA-polydrug | 50 | 88% |
Table 2 also proves that marijuana can kill and maim innocent victims. That is not a new observation, but it is one that some still deny.
11 It takes an hour to take a blood sample. Two hours if a victim is involved.
A major reason that Colorado’s 5 ng/ml THC ‘legal limit’ in blood is scientifically invalid is that blood is never impaired by THC. Only the brain is impaired. Since THC is fat soluble, it is very quickly absorbed by the brain and other highly perfused fatty tissues. That is why the maximum blood level of THC drops an average of 73.5% within the first 25 minutes after beginning to smoke a joint. Within an hour, the THC level drops below 5 ng/ml for occasional users, and within two hours it drops below 5 ng/ml for chronic users[1]. And that is for cases of a driver smoking a joint at the time of a crash or an arrest. THC blood levels are even lower for a stoned driver who stops smoking an hour or more before driving. Colorado’s 2016 data revealed that the median time between a traffic stop and blood collection was between 55 and 60 minutes, but was 114 minutes between dispatch of an officer to an injury crash and blood collection. Times for fatal crashes were not reported. With such sample collection delays, blood testing cannot reliably measure the concentration of a drug causing impaired driving.
12 A 5 ng/ml law is prosecutorially ineffective, scientifically invalid and judicially unsound
Colorado’s 5 ng/ml THC permissible inference law targets a very small proportion of the state’s DUI drivers; 862 out of 27,244 in 2016. And for those, less than half were convicted of DUI according to Table 1. The law is clearly not effective in achieving prosecutions.
Observation #11 stated only one reason that THC’s ‘legal limit’ is scientifically unsound. Another is that drivers impaired only by marijuana edibles reach a maximum blood THC level of under 3 ng/ml, and that is if they take five times the normal 10 mg dose!
Observation #10 demonstrates how judicially unsound Colorado’s law is. Fortunately, very few victims were killed or injured by marijuana-impaired drivers, but those who were killed or injured deserved the same chance for justice as victims of drunk driving.
[1]Wood E. Weakest in the Nation: Colorado’s DUID laws are the weakest in the nation; why and how to fix that. July 2018. Figure 8 page 28