This study explores three alcohol-related databases so as to provide a comprehensive understanding of drinking patterns and the prevalence of alcohol-impaired driving in Catalonia (Spain). The rate of alcohol-impaired drivers is compared with the percentage of drinkers in this population, with a particular focus on heavy episodic drinkers. Two strategies adopted by law enforcement agents when conducting alcohol breath tests are examined: (i) non-random and (ii) random approaches to roadblock location and driver selection. We find that heavy drinker profiles (in terms of age and gender) closely match those of alcohol-impaired drivers detected at strategically located, non-random sobriety checkpoints (especially in the case of female drivers), and that they also correlate with the age-gender distribution of drivers involved in road accidents with victims. Different drink driving patterns are detected when sobriety checkpoints are located randomly and drivers are tested at random. Older drivers are identified as a risk group as they abandon the safer driving habits in relation to alcohol shown when they were middle-aged. A combination of non-random and random controls would increase driver perception of their chances of being detected when drink driving. As such, the whole population, regardless of their drinking profile, would be alerted to the serious personal, social and legal implications of alcohol-impaired driving.