Abstract
Background: Driving under the influence of alcohol or other substances (DUI) and suicide are both persistent but potentially preventable public health problems. DUI offenders, particularly those who have repeatedly offended, report elevated risk for many suicide risk factors. However, researchers have not systematically studied associations between repeat DUI and suicide.
Methods: Out of a population of 729 repeat DUI offenders, 207 were asked about lifetime suicidal thoughts and behaviors (STB) and lifetime instances of DUI as an attempt at self-harm. We calculated rates of DUI as self-harm and compared lifetime STB rates to rates observed in the National Comorbidity Survey Replication (NCS-R). In addition, we examined associations between known risk factors of STB (e.g. psychiatric disorders, marital status) and lifetime STB outcomes among repeat DUI offenders.
Results: Nearly 4% of repeat DUI offenders reported that at least one past experience of DUI was a deliberate attempt to take their own life and an additional 3.4% reported that at least one past experience of DUI was a form of non-suicidal self-injury (NSSI). Rates of lifetime STB all were significantly higher among repeat DUI offenders than NCS-R respondents. For repeat DUI offenders, several risk factors (e.g. race/ethnicity, age, anxiety disorders, NSSI) shared significant relationships with at least one lifetime STB outcome.
Conclusion: An exclusive focus on drinking habits and problems is insufficient to address mental health risk among repeat DUI offenders, who might require specialized services designed to address STB.
Keywords:
Acknowledgement
Dr. Edson had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Disclosure statement
The authors received primary funding from the Foundation for Advancing Alcohol Responsibility (FAAR; a foundation that receives funding from the alcohol industry). However, FAAR was not involved in reviewing this paper and did not have advanced knowledge of the research question.
Notes
1 Anecdotally, primary reasons for declining to participate were wanting to distance themselves from their DUI, suspicion about providing any identifying information, and not wanting to take time to do the baseline interview.
2 DUI offenders who agreed to participate and were assigned to complete CARS were further randomly assigned to do so either as a computer-guided interview with their counselor or as a self-administered computerized interview. There were very few differences between results from those two conditions, so we combined them for purposes of this paper.
3 We conducted an attrition analysis comparing descriptive statistics for demographics (gender, age, race/ethnicity), marital status, employment status, and three mental health outcomes (whether respondents had experienced drug problems, anxiety, or depression over the last 30 days, as of the baseline interview) between the 207 repeat offenders who completed the baseline interview and the six month follow-up and the 138 repeat offenders who completed the baseline interview but did not complete the six-month follow-up. Repeat offenders who completed the six-month follow up were slightly older (mean = 39.7 years of age, SD = 11.5) than repeat offenders who did not complete the six-month follow up (mean = 36.2, SD = 10.7, t (340)= 2.84, p < 0.05). Repeat offenders who completed the six-month follow up were also less likely to be employed than repeat offenders who did not complete the six-month follow up (55.1% vs. 70.3%, χ2 = 21.3, p < 0.05). The two groups did not significantly differ on any other variables.
4 Overdoses, which are included as a criterion for the NSSI measure, also might in some circumstances be a form of suicide attempt. To test for criterion contamination (i.e., whether the overdose criterion is more so related to the outcome than to the predictor, thus biasing the results), we re-ran our analyses with the overdose criterion removed from our NSSI measure. This removal resulted in no substantive changes to our results. Therefore, we opted to include overdoses in the NSSI measure.
5 As Table 1 shows, screening status for addiction-related disorders had low variation. Therefore, we excluded it as a predictor variable.