This study examined the association between arrest and admission to specialty drug treatment using a community dwelling nationally representative sample. As such, the study initially focused on the independent contribution of arrest to the probability of a treatment admission in the same year. Finally, the analyses examined if the association between arrest and treatment admission was similar to or different from the association with treatment admission at other points in the criminal justice continuum.
Self reported arrest in the NSDUH appears valid. According to the Criminal Justice Sourcebook , there were 8.2 million adult arrests reported in 2002. Our projections based on the NSDUH data indicate that nearly 6 million individuals accumulated almost 7.8 million arrests in this same year. The remaining differences could be attributed to sampling error, method for collection of arrest data in the NSDUH (maximum of three arrests tallied), under-reporting of arrests by individuals, or that some of the arrested population is likely imprisoned and not part of the survey population. Demographic characteristics between criminal justice data and those reporting criminal justice involvement in the NSDUH sample were also similar, particularly the predominance of white males.
Our findings supports previous findings of the high proportion of drug and alcohol involvement among those involved in the criminal justice system [4, 20, 22]. The arrest group was much more likely to have used an illegal substance in the past year, meet criteria for drug or alcohol dependency and enter specialty treatment than the non-arrest group. This trend remained consistent even when restricting analysis to only those meeting diagnostic criteria for SUD.
However, the examination of arrest versus non-arrest may obscure differences in legal status that might influence treatment entry, conflating what might be individual factors with system factors on treatment admission. Replacement of the two-level arrest/non-arrest variable with the four-level categorical variable representing current criminal justice status in the analyses did not change the individual factors predicting treatment (i.e. unmarried, SMI, and SUD). In fact, examination of multiple groups underscores that those involved in the criminal justice system, irrespective of the point in the continuum, are more likely to be admitted to treatment than those who are not. However, the variation in the likelihood of treatment involvement among those points in the continuum is cause for further examination.
The odds of being admitted into treatment for those in the Arrest Only group are seven fold that of those in the non-CJ group, even though they may not be processed through the continuum further. Although we cannot precisely determine the temporal ordering of arrest and treatment, it is possible that the perceived threat associated with arrest may motivate individuals to enter treatment independent of the coercion associated with mandated treatment through formal sentencing. However perceived threat may increase as legal status changes since those in the Arrest Only group were the least likely among the CJ groups to be admitted into specialty treatment even though the proportion of those who met criteria for an SUD was similar to Arrest & Supervision (45% and 55% respectively) and greater than Supervision Only (22%).
One explanation for those in the Arrest & Supervision group being most likely to receive treatment is the greater level of scrutiny they experience through their involvement in the criminal justice system. Hypothetically as they move from arrest to conviction and sentencing, they experienced court ordered assessments and investigation and multiple court appearances. The recent experience of criminal justice processing may achieve the highest level of perceived threat among the dispositions studied and thus induce the highest level of motivation to participate in treatment.
In a recent study of drug-using offenders on probation, Longshore & Teruya , like others  found a significant relationship between motivation and external pressure in the form of criminal justice coercion. However, Longshore and Teruya  argue that treatment motivation is better understood as two related but distinct constructs – readiness and resistance. Readiness, as the positive side of motivation, was a significant predictor of treatment retention during the 6-month period after intake. Resistance, or opposition to treatment, is considered the negative side of motivation and predicted drug use. The authors point out that opposition to treatment can be based upon sceptical views of treatment or in the case of legal mandate – a resentment over the loss of individual control over decision making.
Considering the dual constructs within measurement of motivation forces us to assess not only how and when coercion is used, but also contemplating how to enhance treatment readiness among offending populations. If treatment is unwarranted or the offender becomes oppositional to treatment – even though they enter – we may have squandered important resources. For example, those in the Supervision Only group were most likely to be admitted to specialty treatment without a diagnosis of either abuse or dependency. It could be that a history of a drug or alcohol problem combined with a recent relapse was the catalyst for a mandated or voluntary treatment admission. Alternatively, treatment could be a proxy for a criminal justice sanction, with treatment involvement as an inducement for avoiding confinement. Conversely, confinement may obscure time frames associated with diagnosing a current SUD. Those in jail or prison for an extended period of time would not report recent use, and hence not meet DSM diagnostic criteria for a substance use disorder simply because they did not have the opportunity to use. However, further studies are needed to examine if treatment entrance was clinically appropriate or more associated with legal coercion.
Inappropriate admission into specialty treatment is not isolated to the intersection between criminal justice and treatment. Our findings suggest, similar to others , that those in the general population are admitted to treatment without meeting diagnostic criteria or when they have relatively mild manifestations of their disorder. In fact, among the four groups, the non-CJ group was second highest among those entering treatment without a SUD diagnosis. This finding may reflect difficulties in relying on self-report, determining clinical diagnosis from structured interviews, or people obtaining treatment inappropriate to the level of care required . However, limiting our analysis to only those who met current diagnostic criteria may have obscured this finding.
Although we speculated that informal social controls such as family relationships might be predictive of treatment admission, marriage did not positively influence treatment admission in any of our models. Reflecting other studies , we found those who were married were less likely to enter treatment than those who were not. Certainly a dichotomous variable does little to inform us as to the quality of the relationship or the influence an individual may have in coercing their spouse into treatment. However, in another study, a more qualitative measurement of relationship that captured some of these additional marital elements (i.e. support) yielded similar results . Furthermore, others have found that marriage may be protective of continued drug use and criminal activity, particularly among offending populations .
Finally, it should be noted that individuals with an SMI were more likely to be in criminal justice populations than in the general population and they comprised approximately 20% of those with a SUD in each of the offending and non-offending groups. The co-occurrence of SMIs and SUDs has been well documented [31–33], but it is of particular concern when the individual is involved in the criminal justice system due to the inattention of co-occurring disorders in both the substance abuse and mental health treatment systems [32, 33]. A supervising agent making a referral to a mental health provider may assume that substance abuse issues would be addressed, or vice versa. Treatment non-compliance and/or a misunderstanding of the interaction of the two disorders may result in a higher level of confinement (jail or prison) than appropriate treatment would indicate. Although it is noteworthy that SMI predicts receiving specialty treatment across models, these multiple morbidities (SMI, SUD, and CJ) require highly specialized services that integrate mental health and substance abuse treatment while simultaneously integrating the requirements and potential constraints that the criminal justice system imposes .
The self-report nature of the data may underestimate socially sanctioned behaviors such as arrest and drug use resulting in lower bound estimates. In addition, the NSDUH survey samples individuals living in the community, and therefore we are missing the subpopulation of those who are arrested and incarcerated. Although a small percentage of those arrested actually go to state and federal prisons, it potentially excludes those who have committed more serious offenses and who may have more serious and chronic substance abuse problems .
Another limitation is our inability to know the temporal sequences. We do not know if the arrest or treatment came first; we only know that both treatment and arrest occurred within the same year. Unfortunately, other potentially useful datasets such as the Treatment Episode Data Set and the now defunct Arrestee Drug Abuse Monitoring program either do not include "arrest" as a referral mechanism or do not have community comparison groups [4, 36].
In addition, we do not know the sequence of events for people who have both arrest and criminal supervision within the same year. Certainly our assumption that they have been convicted of the offense they were arrested for and are now serving out their sentences in the community under probation/parole supervision is probable: there is no data, however, to confirm it. Consequently the four group categorizations were derived to focus on those who clearly identify as those at the arrest phase of the criminal justice continuum without including those under supervision. Finally, this study did not control for treatment access and other barriers to treatment which may have influenced the results if known.