In the present study, we compared drug use profiles, behaviors, and beliefs related to impaired driving among drivers with and without alcohol-related infractions stopped at Brazilian roadblocks, focusing on the characteristics of drivers who refused to take the breathalyzer test. We found that those who refused the test had a higher proportion of alcohol use and drunk driving behavior in the previous year, even though most drivers in this group had already been stopped before at roadblocks and believed in the effectiveness of enforcement.

Several studies have reported a high prevalence and severity of PAS use and dependence among drivers with a history of driving under the influence.^{20 - 23} Current data in Brazil estimate that around 50% of the population aged 18 years or more has used alcohol in the previous year,²⁴ compared to 88.2% in our sample. This discrepancy was even higher in the RD group, where the proportion of individuals reporting alcohol use was almost two times greater than the national prevalence. Although we found no difference in drug use between groups, the overall sample reported a higher proportion of cannabis and cocaine use than those reported for the general population in previous reports (2.4% and 2%, respectively).²⁴ Since all drug use data were self-reported, it is possible that they were underestimated. One hypothesis for this result relies on the fact that the sample was approached during night roadblocks, as evidence suggest that the prevalence of alcohol and drug positive drivers increase on week nights and weekends.²⁵ Besides, the prevalence of alcohol and drug use in Rio Grande do Sul is higher than in most of the country’s states. However, history of drug use is not the sole risk for DUI; indeed, comprehensive assessment of drug use profiles, cognitive traits, and personality characteristics seems to be the best way to identify citizens who are higher-risk drivers, as well as to assess those who are not responsive to traditional enforcement measures.^{20 , 26}

Policies and enforcement strategies against impaired driving vary greatly between countries and are generally drafted based on local needs. In this context, understanding the profile of drivers with a history of drunk or drugged driving is essential to designing specific approaches to this public problem, since the mere importation of policies or programs from other countries would not cater to our current needs in Brazil. Notwithstanding, there are few studies worldwide evaluating the profile of impaired drivers in LMIC. International studies from Europe and the United States that assessed offending drivers showed a high proportion of men,^{27 - 30} and of drivers of cars rather than other models.³⁰ Mean age varied slightly between studies but remained within a range of 30-50 years.^{29 , 30} Some exceptions were also observed with younger²⁷ and older subjects.³¹ Although average income may not be directly assessed through studies, there is a high proportion of employed or self-employed individuals.^{29 , 30} In previous Brazilian and LMIC studies, drivers with traffic infractions also present a similar sociodemographic profile, with most identifying as young male adults from middle or upper social classes.^{21 , 32 , 33} In this sense, the sociodemographic profile of our sample is in line with international and previous Brazilian data, but the fact that we evaluated a sample from southern Brazil must be considered. However, even with a similar sociodemographic profile of traffic offenders, drivers from LMIC and developed counties have different perspectives on traffic safety.

In our results, we noted a significant difference between groups regarding driving behavior and beliefs. The high proportion of drunk and drugged driving among drivers who refused the breath test suggests a potential risk among this population. Moreover, the RD group also had a lower proportion of drivers who believe that traffic enforcement can reduce drunk driving. Previous data show that drivers with low perceived risk of legal consequences commit more alcohol-impaired driving offenses.^{34 , 35} Moreover, drivers with higher perceived risk are more compliant with enforcement and treatment countermeasures and are more likely to report safer driving behaviors.³⁵ Altogether, the higher risk profile of refusing drivers seen in our results, the mild sanctions, and the weak enforcement system in Brazil may be favoring the continuity of risk behaviors. Therefore, our results support the position that driving laws should penalize refusal as a traffic crime in order to achieve reduction of repeat offenses and to improve social deterrence.

Recently, Pechansky et al. showed how the differences in enforcement and sanctions for DUI could be compared to road traffic mortality rates and to social and economic development parameters in different countries.¹⁰ According to these authors, developed countries with strict laws and enforcement actions have higher social deterrence regarding DUI and, consequently, lower traffic mortality rates. In this sense, several authors have discussed the influence of social deterrence on DUI behavior. While there is still no consensus, most studies suggest that social deterrence is an effective way to reduce negative outcomes in traffic.^{36 , 37} Some studies have shown that repeat offenders in Brazil are usually aware of the traffic laws and risks of TCs involving specific behaviors, which supports the fact that there is lack of social deterrence among Brazilian drivers.^{38 , 39} In our study, 43.3% of drivers with alcohol-related infractions reported they had been stopped at a roadblock in the previous year. Although most said they believed in the effectiveness of enforcement, they were being fined for drinking and driving at the time of data collection. Likewise, 69.7% and 87.5% of PD and RD who reported alcohol consumption in the previous year were recidivist DUI offenders, respectively. When combined, these results indicate that the usual traffic enforcement actions may have no effect on this group of subjects.

Although considered an administrative offense, drivers who refuse a breath test and do not exhibit additional evidence of intoxication can see this possibility as a way to circumvent legislation. It is worth noting that all groups were similarly assessed for signs and symptoms of impairment. The literature is conflicting with regard to the validity of signs and symptoms as the sole evidence for being under the influence.^{40 - 44} In our study, drivers with positive BrAC and drivers who refused the breath test had the same number of signs and symptoms as drivers with negative BrAC, which suggests that, in its present form, the assessment performed by traffic agents was not accurate for identifying DUI among drivers. Although some subjects reported being under the influence of PAS other than alcohol, this proportion was not different between groups. It is therefore important to reevaluate the policies regarding recognition of signs and symptoms in drivers and to invest in the specific training of agents to detect them – perhaps training drug recognition experts, as an example of the strategies used by some developed countries.^{45 , 46}

One of the limitations of this study is the fact that it was conducted exclusively with drivers who were not be allowed to return to the road and continue driving, which prevents comparisons with a control group. Another limitation is the fact that data were collected on the roadside, where time and infrastructure for data collection is limited. This study therefore lacks information regarding the personality and cognitive aspects of drivers, which would be important to better define the risky profiles among groups. However, even with a cross-sectional design and a small sample size, we were able to detect initial evidence that distinguishes between drivers committing different infractions, suggesting that drivers who refuse breath tests could be a high-risk group.