Evidence from regression-discontinuity analyses for beneficial effects of a criterion-based increase in alcohol treatment

Recidivism prevention for impaired driving: Longitudinal 5-year outcomes from Quebec's severity-based intervention assignment program

INTRODUCTION

Driving while impaired by alcohol (DWI) is a persistent problem. Tailoring intervention modality to client risk and needs (i.e., risk/needs) is posited to both reduce recidivism more efficiently than uniform approaches and circumvent overtreatment or undertreatment. DWI drivers in Quebec must participate in a severity-based intervention assignment program to be relicensed, but like most tailoring programs it has yet to undergo systematic scrutiny. The current longitudinal cohort study tests two main hypotheses underpinning this approach: 1) drivers classified at higher recidivism risk based on their arrest characteristics (DWI_R) show poorer outcomes over up to 5-years postassessment compared to drivers classified at lower risk (DWI_F); and 2) for both DWI_R and DWI_F groups, assignment of drivers with greater risk/needs to intensive intervention (II) will be advantageous for reducing recidivism risk compared to assignment into brief intervention (BI) for those with lower risk/needs.

METHODS

Drivers who entered the program from 2012 to 2016 were followed to the end of 2018 (N = 37,612). Survival analysis examined the predictive validity of the initial classification into DWI_R or DWI_F groups for documented recidivism over a follow-up of up to 5 years. Logistic regression discontinuity evaluated the relative outcomes of drivers who were assigned to either BI or II. The study explored interaction effects between classification and intervention assignment with age and sex.

RESULTS

In line with the hypothesis, the average hazard of recidivism was 58 % greater in DWI_R drivers compared to DWI_F drivers. In both DWI_F and DWI_R drivers, assignment of drivers with greater risk/needs to II was associated with reduced recidivism compared to assignment of drivers with lower risk/needs to BI, with 57 % and 35 % decreased probability of recidivism, respectively. Younger age was more strongly associated with recidivism risk in DWI_F drivers than in DWI_R drivers.

CONCLUSIONS

The current study found that Quebec's severity-based intervention assignment approach accurately identifies DWI drivers who: i) by their arrest characteristics pose a greater risk for recidivism, which may require expeditious exposure to preventative countermeasures; and ii) as a function of their greater risk/needs, benefit from assignment to more intensive intervention to mitigate their recidivism risk.

Evaluating moderators of beneficial effects of severity-based assignment to substance use treatments in impaired drivers

Remedial programs for impaired driving offenders have proved valuable in reducing subsequent alcohol and other drug use and preventing recidivism in this population. Many of these programs are based on a severity-based assignment scheme, where individuals assessed to have greater problems or be at higher risk are assigned to longer, more intensive interventions. Recent research, using regression discontinuity analyses, provided support for severity-based assignment schemes in demonstrating that those with higher problem or risk levels assigned to longer and more intensive programming showed a significant reduction in drinking days over a follow-up interval, attributable to program assignment. Regression discontinuity analyses can also be used to assess moderators of this assignment benefit. We report an assessment of the impact of eight potential moderators of assignment benefit, derived from a factor analysis of the Research on Addictions Self-Inventory screening instrument. Five of the eight factors were found to moderate the assignment benefit: Negative Affect, Sensation Seeking, High Risk Lifestyle, Alcohol Problems, and Family History. The significance of these results for developing more effective program assignment procedures is discussed.

Multiple "Lower BAC" offenders: Characteristics and response to remedial interventions

BACKGROUND

In recent years, there has been increasing attention to "lower BAC" drinking drivers, typically those whose blood alcohol content (BAC) is under the legal limits defined in criminal law. In 2009, legislation was enacted in Ontario, Canada that enabled police to issue roadside license suspensions to individuals caught driving with BAC between 0.05% and 0.08%, known as the "warn range". Multiple warn range (MWR) offenders are required to attend the Back on Track (BOT) remedial measures program. This study aimed to provide: (1) a preliminary characterization of MWR drivers charged under warn range legislation; and (2) an initial assessment of outcomes associated with BOT participation among MWR offenders.

METHODS

A subsample of 727 MWR offenders was drawn from program records, and compared to samples of 3597 first-time Criminal Code (CC) offenders (those caught driving with a BAC of 0.08% or higher) and 359 second-time CC offenders. To provide an initial assessment of outcomes associated with BOT participation, another subsample consisted of 394 MWR participants from whom pre- and post-workshop questionnaires were collected and successfully matched using probabilistic matching processes.

RESULTS

Similarities in demographic profile and driving history between MWR and first-time CC participants were apparent. MWR offenders scored higher on risk of problem drinking and drink-driving recidivism than either of the CC offender groups. Second-time CC offenders scored higher on these measures than first-time CC offenders. Following BOT participation, MWR participants demonstrated positive change including improved knowledge of and intentions to avoid drink-driving.

CONCLUSIONS

MWR offenders share a similar demographic profile to that of first-time CC offenders and they report significantly higher risk of problem drinking and recidivism. MWR offenders may include high-functioning problem drinkers who are likely to continue drink-driving and who may escalate to a CC drink-driving offense. Like CC offenders, MWR offenders benefited from BOT participation.

Characteristics and predictors of recidivist drink-drivers

OBJECTIVE

The current study compared characteristics of nonrecidivist versus recidivist drink-drivers and of recidivists in their first versus second appearance at Back on Track (BOT), the remedial measures program for impaired drivers in Ontario, Canada.

METHODS

Information from 59,134 convicted drivers who participated in BOT between 2000 and 2010 was examined to identify drivers who completed the program a second time following reconviction.

RESULTS

A total of 586 recidivists were identified. Compared to nonrecidivist drivers, recidivists at first attendance were more likely to be male and had higher scores on measures of alcohol dependence and adverse legal consequences of substance use. Compared to nonrecidivist drivers, recidivists at second attendance were significantly older, had a higher income, were more likely to be retired, and were less likely to be employed part-time. They had fewer legal problems. Recidivists reported fewer drinking days and fewer drinks per occasion but greater use of benzodiazepines than nonrecidivists and had higher scores on a measure of future risk of alcohol- and drug-related problems. Comparison of recidivists' characteristics at first versus second attendance confirmed many of these findings, with second-time recidivists reporting fewer drinks per drinking day and greater use of benzodiazepines and having higher scores on a measure of future substance use problems than first-time recidivists.

CONCLUSIONS

Results suggest that identification of drivers at increased risk of recidivism may be possible at first program attendance by examining indicators of increased alcohol-related problems. In addition, recidivists appear to show a greater readiness to change at second attendance. Implications for remedial program development and recommendations for future research are discussed.

Working in tandem: The contribution of remedial programs and roadside licence suspensions to drinking and driving deterrence in Ontario

In 1998, Ontario implemented a remedial program called "Back On Track" (BOT) for individuals convicted of alcohol-impaired driving. Drivers convicted before October 2000 were exposed to a single-component program ("Edu BOT"); those convicted after participated in a multi-component program ("Full BOT"). We evaluated the impact of BOT, and the preceding 90-day roadside licence suspension, on drinking and driving recidivism, an outcome yet to be examined, using population-wide driver records. A Chi Square Test was used to compare the three-year cumulative incidence of recidivism between three historically-defined cohorts: No BOT, Edu BOT, and Full BOT. Stratified analyses by completion status and by age were also conducted. Analyses of the roadside suspension were conducted using an interrupted time series approach based on segmented Poisson/negative binomial regression. The roadside suspension was associated with a 65.2% reduction in drinking driving recidivism. In combination with indefinite suspensions for non-completion, the BOT program was also associated with a 21% decrease in drinking and driving recidivism in the three years following a CCC driving prohibition, from 8.5% to 6.7%. This reduction cannot be explained by pre-existing trends in recidivism. Conversion of the BOT program from the single-component version to the multi-component program further reduced the three-year cumulative incidence of recidivism to 5.5% (a total reduction of 35% from pre-BOT). Results provide strong converging evidence that remedial alcohol education/treatment programs in combination with other sanctions can produce substantial increases in road safety.