The distracted mind on the wheel: overall propensity to mind wandering is associated with road crash responsibility.

Responsibility for the crash

We determined responsibility levels in the crash using a standardized method adapted from the Robertson and Drummer crash responsibility tool (11). The adapted method takes into account mitigating factors likely to reduce driver responsibility: road environment, vehicle related factors, traffic conditions, type of accident, traffic rule obedience, and difficulty of the driving task. Each factor scores from 1 (not mitigating, i.e. favorable to driving) to 3 or 4 (mitigating, i.e. not favorable to driving). All six scores are summed to provide a responsibility score (multiplied by 8/6 to be comparable with the eight factor score proposed by Robertson and Drummer). This method has been previously validated in the French context (5,12–15). Indeed, two factors such as “level of fatigue” and “witness observation” are unavailable in French Police records. The higher the score, the lower the responsibility. Responsibility scores are classified into three categories: 8- 12=responsible; 13-15=contributory; >15=not responsible. Drivers displaying any degree of responsibility for the crash were classified as cases (score ≤15); drivers who were judged not responsible (score >15) served as controls. The interviewer was unaware of the responsibility status while interviewing the participants since responsibility scores were computed during the analysis. Risk factors

Participants were asked to describe their thoughts just before the crash and the question was coupled with a numeric scale from 0 to 10 that captured the self-estimated level of perturbation. In order to reduce memory bias and halo effect, two opportunities were offered during the interview to report thoughts which were subsequently classified as being related or not to driving. The Mind Wandering State was defined as the report of any thought unrelated to driving. A Disturbing Thought (DT) corresponded to a Mind Wandering State with a perturbation rating higher than 4. Perturbation level was indeed the answer to “How disturbed / distracted was this thought?”. Mind Wandering Trait was built from a scale comprising four items selected based on their clinical significance. Two items are part of the Day Dreaming Frequency Scale (DDFS): Daydreams and fantasies make up X % of the day, and Recalling things from the past, thinking of the future, or imagining unusual kinds of event occupies X% of my day (16). Two items were developed from literature data: In general, when you drive, how often do you happen to think about something else? And In general, when you read, how often do you happen to think about something else. For each question, the related time spent each day was measured from 0 to 100 percent. If the frequency was higher than 50% for at least one item, the patient was defined as in the high category of the boolean MWT variable.

The analysis also included well-known risk factors for road crash and potential confounders such as patient characteristics (age, sex, socioeconomic category), alcohol consumption during the 6 hours before the crash and self-reported psychotropic drug use the day before accident. Characteristics of the crash were also reported (location, vehicle type). The variable Distractive Activity was obtained by asking participants about their activities just before the crash (this included use of a mobile phone, listening to radio/television, talking with or listening to a passenger, manipulation of electronic devices, manipulation of objects, grooming, smoking, eating, drinking, reading). Patients were also asked to evaluate their pain at the time of the interview with a numeric scale; A painful participants was defined as with a self-rated pain value strictly superior to 3. Participants were also asked whether they had been distracted by a distracting event that occurred inside or outside the vehicle. Sleep Deprivation was evaluated with The Epworth Sleepiness Scale (ESS) (17).

Statistical Analysis

Univariate analysis was conducted to investigate the link between crash responsibility and risk factors using Student t-test for continuous variable and Chi-square test for categorical variable. Multivariate analysis was then performed with a step by step backwards selection procedure keeping all significant variables (p < 0.05) and all confounders (variation of β > 20%). We then tested interactions between independent variables kept in the final model. Finally, we performed sensitivity analyses to assess the robustness of the results: 1. by stratifying on pain; 2. by changing the cut-off for responsibility score to 14 and 16; 3. by stratifying on the existence of chronic disease.