Predictors of commencement and completion of the NSW Mandatory Alcohol Interlock Program

By Sara Rahman

AIM To identify factors which predict commencement and completion of the Mandatory Alcohol Interlock Program (MAIP). METHOD We use a dataset from Transport for NSW after comprising 10,209 Mandatory Alcohol Interlock Orders (MAIOs) with an initial disqualification period ending before 20 April 2019. We also examine a subsample of 2,860 MAIOs (with an expiry date prior to 20 April 2019) where a person commenced MAIP. These records were linked to the NSW Bureau of Crime Statistics and Research’s Re-offending Database and MAIP operational data. Logistic regression models were used to identify factors predicting commencement and completion of the program. To assess model performance, we report the Area Under the Curve (AUC) and the percentage correctly predicted. RESULTS Offenders had a decreased likelihood of commencing the program if they were: a) already disqualified at the time of the offence (15 percentage points less likely to start); b) aged 55 years and above (20 percentage points less likely to start the program than 18-24 year olds); c) Aboriginal (12 percentage points less likely than non-Aboriginal offenders and 15 percentage points compared to those with unknown Aboriginality); or d) sentenced to imprisonment at the index contact (15 percentage points less likely to start). Our model has moderate predictive power (AUC=0.68). While the majority of starters completed the program, having an existing disqualification or suspension, a longer interlock period, and having an order extended were all associated with non-completion. Demographic factors, particularly age and Aboriginality, were also significant independent predictors in our model predicting completion, which has acceptable predictive accuracy (AUC=0.70). CONCLUSION Interlock installation can be predicted moderately well by licensing, operational, demographic, and criminal justice information. However, we lack information on other potentially important factors such as risk preferences and the availability of alternative transport. Most of those who commence the interlock program complete it and thus, improving commencement should be a greater priority for policymakers.

(Bureau Brief No. 160). Sydney: NSW Bureau of Crime Statistics and Research 2022. 17p.