In the three jurisdictions TRI studied, approximately 50 percent of the felony and misdemeanor drug court clients met criteria for being at high risk, meaning they had APD, a prior drug treatment history, or both. These findings suggest that no more than half of drug offenders might reasonably be expected to perform adequately in the type of low-intensity, nonjudicially managed diversionary intervention exemplified in recent State policy initiatives such as Proposition 36 or Proposition 200 (see "Statewide Diversion to Treatment for Nonviolent Drug Offenders").

A substantial proportion of drug offenders could be at risk for failing in such an intervention, suggesting that criminal diversion statutes should, perhaps, incorporate some mechanism to permit poorly responding individuals to be readily transferred to a judicially managed program. The findings have further implications for best treatment practices and for ethical guidelines for drug treatment providers. Clinicians are typically socialized to maintain strict confidentiality and nonporous professional boundaries between themselves and criminal justice authorities. The data suggest this might, indeed, be therapeutic for low-risk clients who may need a safe and discreet setting to focus on their problems. Such an approach, however, would appear to be contraindicated for high-risk clients who could deliberately evade detection of infractions or might exploit gaps in communication and monitoring.

Ideally, then, there might be (at least) two tracks in treatment programs, involving different service arrangements with courts and probation and parole offices.

The content of treatment might also be tailored to clients' risk levels. Highly structured behavioral or cognitive-behavioral interventions are ideally suited for many offenders, particularly those identified as "high risk" (e.g., Cornish and Marlowe, in press; Gendreau et al., 2001). In contrast, insight-oriented or group-process interventions have been associated with increased rates of drug use and recidivism among high-risk offenders, and educational or drug-awareness sessions have been shown to have no effect for any offenders (e.g., Pearson and Lipton, 1999; Taxman, 1999b).

The worst outcomes have been seen with insight-oriented treatments that

The best results have been obtained from programs that focused on

Of course, observable and diagnosable symptoms of depression or anxiety should also be targeted in conjunction with any treatment regimen.

Furthermore, in the most successful programs, staff members have been in a position to reliably detect clients' accomplishments and infractions in the program and to apply rewards for desired behaviors and negative sanctions for undesired behaviors (e.g., Harrell and Roman, 2001; Marlowe and Kirby, 1999; Taxman, 1999b).

For instance, the most effective programs regularly monitor clients' substance use through random breathalyzers and urinalyses.

Drug-free test results are met with rewards such as

Drug-positive results, on the other hand, are met with such sanctions as

If a particular program's philosophy or structure cannot easily accommodate such an approach, that program might consider having a separate, intensive, behavioral or cognitive-behavioral track for high-risk offenders or might consider not accepting referrals to treat such offenders.

Pharmacological interventions are seriously underutilized in the criminal justice system despite the fact that several medications have demonstrated success for reducing substance use and crime among offenders (e.g., Cornish and Marlowe, in press).

Methadone maintenance treatment, in particular, has been consistently demonstrated in numerous experimental studies to reduce drug use and criminal activity among opiate addicts, with effects many times the size of hospital-based detoxification, drug-free outpatient treatment, and residential treatment (e.g., Platt et al., 1998). In a controlled experimental study, researchers at the University of Pennsylvania similarly found that Federal probationers who were randomized to receive naltrexone in combination with psychosocial counseling had lower rates of opioid-positive urines and were less likely to be reincarcerated for probation violations than those receiving psychosocial counseling alone without naltrexone (Cornish et al., 1997). Subsequent studies by the same investigators are examining the effects of oral and depot naltrexone among State parolees, probationers, and drug court clients. Preliminary data from those studies suggest that oral naltrexone may be more effective in retaining parolees in treatment than standard psychosocial treatment alone.

It is possible that opioid-antagonist medications such as naltrexone may be more palatable to policymakers and criminal justice practitioners because they are not perceived as substituting one addictive substance for another, as is commonly ascribed to methadone. Further research is needed to evaluate the acceptability and effects of these types of medications in correctional settings, and to identify and resolve barriers to the use of efficacious medications with criminal justice clients.