The causes underlying traffic offender behaviour might be the tip of the iceberg that represents a persistent burden on global health and serves to detect other serious social, family, work problems. Better understanding and prevention of recidivism following a First-time driving under the influence conviction and analysis of High Risk offender driver behaviors are needed. This study seeks to identify the factors that predict an offender driver profile, analysing not only drivers’ general use of alcohol and/or drugs, but also their ability to dissociate the use of alcohol and other substances from driving. A total of 315 drivers — 97 offenders (95.9 % men) and 218 non-offenders (47.2 % men) responded to a battery of self-report evaluation questionnaires on: alcohol use habits (measured with the AUDIT); drug use (measured with the DAST-20); the ability of dissociating alcohol and substance use from driving; educational level; self-reported traffic violations, errors and lapses (measured with the DBQ, Driver Behaviour Questionnaire); and sociodemographic questions. The model showed good fit indicators (R2 = 0.74). It also displayed good predictive power: (1). It correctly classified 91 % of participants as offender or non-offender drivers. (2). More specifically, its sensitivity was 88 %, having correctly classified offenders as offenders. And its specificity was 92 %, having correctly classified non-offenders as non-offenders. The findings underscores that not only is alcohol use a key predictor of offending behaviour, but so is drug use. These addictions are a health problem, and their consequences are more serious when the afflicted individual drives a vehicle. This article makes evident that repeat offenders have a lower ability of dissociating alcohol and substance use from driving than do non-offender drivers. Non-offenders are shown to be more capable than offenders of dissociating consumption from driving: offender drivers are unable to dissociate the consumption of more than 5 spirit drinks, or the consumption of cannabis, from driving. More than half of non-offender drivers admitted to driving after having had 1 or 2 beers. Along these same lines, we found that offenders believe they have fewer lapses than non-offenders. This could be due to the optimism bias of bold offender drivers, who overestimate their abilities and underestimate their lapses. We also found that offender drivers have a lower educational level. Revoking the High Risk offender’s driver’s licence may not be enough. Instead, High Risk Offenders should be given the support they need to give up the consumption of alcohol and drugs, at least while driving. This could be achieved if they receive motivational interventions and are referred for detoxification treatment.