Does Assisted Outpatient Treatment (AOT) Affect Arrest Rates for Persons with Severe and Persistent Mental Illness?

Background: New York State established Assisted Outpatient Treatment (AOT) in 1999 as a means of outpatient commitment for persons with multiple psychiatric hospitalizations and the potential for violence toward self or others. While proponents of AOT suggest mandated treatment helps to prevent psychiatric decompensation and promotes a least restrictive environment, others argue that AOT in practice is a pseudo-criminalization of mental illness, as law enforcement is often utilized to locate and produce noncompliant individuals on AOT for treatment evaluation. This informal connection to the criminal justice system could enable more formal involvement, such as arrest or incarceration, which most mental health advocates would argue is antithetical to best practices for persons with mental illness. This study examines whether involvement in AOT affects arrest rates for psychiatric inpatient discharges in New York City.

Methods: We interviewed 184 participants (111 men and 73 women) aged 18 to 65 in treatment facilities in NYC. This sample includes 76 individuals court-ordered to AOT and 108 who are involved in non-mandated outpatient treatment. Participants were interviewed at baseline, 3, 6, 9 and 12 months; in addition we obtained NY State monthly arrest records for five years prior to the baseline interview and up to three years following the 12-month interview. Using logistic regression and generalized estimating equations, we estimated the person-month probability of arrest for those currently on AOT (referent) compared to 1)Those never on AOT; 2)Those not yet on AOT; and 3)Those who have completed AOT, while controlling for major demographic differences evidenced between the AOT and non-AOT group.

Results: Average person-month arrest prevalence for the four AOT groupings were as follows: Never on AOT-1.2%; Pre-AOT-2.2%; On AOT-.85%; and Post-AOT-1.4%. After controlling for age, race, gender, months of follow up, and category of psychiatric diagnosis, those who were never on AOT had higher odds of arrest compared to those currently on AOT (OR=1.88, 95% CI=1.04-3.41). Those that were pre-AOT also demonstrated higher likelihood of arrest than those currently on AOT (OR=2.58, CI=1.54-4.33). There was no significant difference in odds of arrest between those currently on AOT and those who were post-AOT. Comparing the Wald chi-square of the complete model minus the control variable-only model resulted in a chi-square of 20.99 (p<.001), suggesting that inclusion of AOT status has a significant impact on the relationship to arrest.

Conclusions: Contrary to the pseudo-criminalization argument, AOT involvement appears to be associated with a reduced likelihood of arrest. Although reduction of arrest is not a primary goal of AOT, these findings suggest that AOT does not appear to further formally connect mentally ill individuals to the criminal justice system. It is possible that AOT involvement may result in greater engagement in treatment and symptom reduction, thus reducing the likelihood of negative life events such as arrest. However, because AOT was not randomly assigned, authors caution an interpretation that AOT may serve as crime reduction for the general population of persons with mental illness.