Recidivism after drug treatment programs: New Sentencing Commission report

On May 17, 2022, the United States Sentencing Commission (USSC) published new findings on eight-year recidivism rates of 25,142 federal offenders after their participation in Bureau of Prisons (BOP) treatment programs. The report is part of a larger multi-year recidivism study of more than 32,000 federal offenders released in 2010. The programs reduced overall recidivism for people who completed programs relative to eligible non-participants. Despite this, drug-related recidivism was still higher among program completers. The latter finding is perplexing considering that one major goal of the treatment programs was to reduce substance use.

In 1990, Congress passed the Crime Control Act that mandated that BOP make available treatment opportunities for prisoners with treatable substance abuse issues. The BOP provides several substance abuse treatment programs, two of which are examined in this report: the Residential Drug Abuse Program (RDAP) and the Non-Residential Drug Abuse Program (NRDAP). Both used cognitive behavioral therapy (CBT) and aimed to reduce paticipating offenders’ substance use and recidivism.

The study used data regularly collected by the USSC, data from the FBI’s Criminal Justice Information Services Division, and BOP data on offender eligibility, participation, and completion information for both programs. After combining all of the data sources together, the authors matched offenderes based on program participation. The final sample was 25,142 offenders, all of whom had served at least one day of incarceration and re-entered the comunity during 2010. Recidivism was measured using re-arrrest rates.

The authors ran a logistic regression to predict recidivism as a yes/no outcome (1=re-arrested and 0=not re-arrested), while controlling for factors such as offense type and criminal history. They found that RDAP and NDRAP completers both had lower rates of overall recidivism than respective eligible non-participants; RDAP completers were 27% less likely to be re-arrested and NDRAP completers were 17% less likely. Interstingly though, rates of drug-related recidivism were higher for both completers and non-completers of these programs relative to eligible non-participants.

The Residential Drug Abuse Program (RDAP)

The RDAP is the most intensive drug treatment program offered by the BOP, and it contains three treatment components: a unit-based or residential component, follow-up services, and transitional drug abuse treatment (TDAT) in a residential community center (RCC). Each component of the program used cognitive behavioral therapy to treat substance abuse and facilitate transition into the community. The residential component required participants to receive treatment in a specialized unit separate from the rest of the prison population. During that time, participants received 500 treatment hours which were completed through half-day sessions over a 9-12 month period. After the residential treatment, offenders returned back to the general prison population and began follow-up treatment in the form of monthly, hour-long group meetings. Follow-up treatment continued for up to one year until the participant transferred to an RCC. During the time in the RCC, participants received TDAT, which consisted of at least one hour of drug treatment per month.

Offenders who posed serious threats to the community were ineligible for residential placement and therefore were ineligible for RDAP. To participate, offenders also had to have a verifiable substance use disorder and at least 24 months remaining on their sentence. Incentives for RDAP included limited financial awards, preferred living quarters, and in some cases, early release (up to a 12-month reduction). Not all participants who enrolled completed the program, so in addition to program completers and eligible non-participants, a third group included the program non-completers. Some voluntarily withdrew, while others were removed due to unsatisfactory progress, misconduct, threats of violence, or attempted escape.

There were a total of 8,474 offenders who were eligible for RDAP. 67% of them (n=5,677) completed the program, 29% (n=2,456) enrolled but did not complete the program, and 4% (n=341) did not enroll at all. There were some slight differences between groups regarding race, but other demographics were fairly similar. RDAP completers were mostly white (39%), while non-completers and non-participants were mostly black (45.2% and 44.7%, respectively). Males comprised more than 80% of each group, and the median age at release was 36 years for each group. There were no differences between groups in terms of medical or mental health care requirements.

Regarding criminal history and offense characteristics, there were important differences between groups. RDAP completers had less extensive criminal histories (median of 3 points) than non-completers (5 points) and eligible non-participants (8 points). Accordingly, a larger percentage of RDAP completers were in a lower criminal history category (CHC) than the other two groups. When it comes to offense type, 80.8% of RDAP completers were sentenced for drug trafficking, compared with roughly 60% of non-completers and eligible non-participants. Further, there were fewer firearms offenders among the RDAP completers (4.3%) compared with RDAP non-completers (17.2%) and eligible non-participants (18.8%).

RDAP completers had shorter original sentences than those in the other two groups, with a median of 60 months, compared to 63 months for the non-completers and 77 months for the eligible non-participants. Completers also served less time overall, with a median of 40 months, compared with 46 months and 49 months for non-completers and eligible non-participants (respectively). Part of the reason for this is likely because RDAP completers were eligible for sentence reductions up to 12 months, which further reduced their time served.

RDAP completers were re-arrested at a lower rate (48.2%) than non-completers (59.2%) and eligible non-participants (68%). This was consistent across age categories and lengths of stay. Additionally, RDAP completers took longer to recidivate (median of 24 months) than non-completers (20 months) and eligible non-participants (16 months). Further, completers also had lower rates of violent recidivism (25.3%) than non-completers (33.2%) and non-participants (47%). In contrast though, RDAP completers had a higher rate of drug-related re-arrests (30%) than non-completers (25.2%) and non-participants (21.6%).

These results seem promising, but there are other things to consider that could have affected recidivism rates. First, differences in offense and criminal history between groups may be a partial explanation: as criminal history category increased for each group, so did recidivism rates. Further, there was a smaller percentage of firearms offenders among completers in comparison with the other two groups. Prior work by the USSC shows that firearms offenders have higher rates of recidivism than non-firearms offenders.

Because of the differences between completers and eligible non-participants, it’s possible that differences in recidivism are related to these factors, making completers less likely to recidivate regardless of program participation. In theory though, these differences shouldn’t affect results due to the modeling techniques used — the authors statistically controlled for the impact of criminal history, crime type, and sentence length in their analysis. However,  it’s important to note that large differences in sample sizes across the groups (n=5,677 versus n=341) may have caused artificial findings. In other words, it’s hard to not find significance when there are vast differences in sample sizes between groups.

To summarize, offenders who completed RDAP were less likely to recidivate than eligible offenders who did not participate in RDAP. Completers had lower recidivism rates across all criminal history categories, all lengths of time served, and in nearly every age group at release. Additionally, RDAP completers took longer to recidivate and had fewer median recidivism events compared to the other groups. Despite this, completers had the highest rates of drug recidivism, while non-participants had the lowest rates. This is perplexing; if the treatment were effcetive, then presumably, drug recidivism rates would be lowest among people who completed treatment. This recidivism finding also might reflect BOP’s policy of placing offenders with persistent substance abuse issues in its most intensive substance abuse program. It still doesn’t explain how non-completers ended up with lower drug recidivism rates though; they would likely also have “persistence substance abuse” issues, but they still fared better in terms of drug-related recidivism than those who completed treatment.

The Non-Residential Drug Abuse Program (NRDAP)

The NRDAP consisted of drug treatment provided in 90-120 minute group sessions over the course of 12 to 24 weeks. Similar to RDAP, NRDAP followed a cognitive behavioral therapy model to reduce susbtance abuse and facilitate successful re-entry. NDRAP targeted specific participants, specifically those who: were waiting to enter RDAP, did not meet RDAP requirements but wanted treatment, were referred to treatment by staff, had a judicial recommendation for treatment, were requried to detox in custody, or used drugs or alcohol while incarcerated.

To be eligible, offenders had to be able to read at an 8th-grade level.  Like RDAP, not all participants complete the program, as they were able to voluntarily withdaw, or sometimes were removed due to disruptive behavior or unsatisfactory progress. Incentives for NRDAP included limited financial awards and in some cases, NDRAP offenders were approved for the maximum period of RCC placement.

A total of 4,446 offenders were eligible to participate in NDRAP. 61.9% (n=2,752) of these people completed the program, while 18.1% (n=803) enrolled in the program but did not complete it, and 20% (n=891) didn’t enroll at all. There was slight variation among the groups in terms of demographics. NRDAP completers and eligible non-participants were mostly white (41.2% and 38.7%, respectively), while NRDAP non-completers were mostly black (43.3%). Males comprised more than 85% of completers and non-completers, but only 75% of eligible non-participants. The median age at release was 38 years for NRDAP completers, which was slightly older than the other two groups, whose median age at release was 35. There were no differences between groups in terms of medical or mental health care requirements.

When it comes to criminal history and offense type, there were also some small differences between groups. NRDAP non-completers had the most extensive criminal histories (median of 5 points), followed by completers (4 points), and eligible non-participants (3 points). Accordingly, a larger percentage of NRDAP eligible non-participants were in a lower criminal history category (CHC) than the other two groups. Regarding offense type, a slightly larger percentage (62.4%) of NRDAP completers were sentenced for drug trafficking, compared with approximately 57% of offenders in the other two groups. Further, there were more firearms offenders among NRDAP non-completers (19.3%) in comparison with the other two groups (about 14%).

NRDAP completers and non-completers had longer original sentences (median of 63 and 60 months, respectively) than eligible non-participants (40 months), which is likely because their criminal histories were more serious. Completers also served the most time in comparison to the other two groups (median of 44 months) compared to non-completers (37 months) and eligible non-participants (25 months). The differing lengths of time served may account for why NRDAP eligible non-participants did not attend NRDAP during their time in BOP custody.

Outcome differences among the NRDAP-eligible participants were not as large as the differences among RDAP-eligible participants. NRDAP completers had the lowest re-arrest rates (49.9%), followed closely by eligible non-participants (54%) and non-completers (59.4%). This was consistent across all lengths of time served and nearly every criminal history category. Regarding age, NRDAP completers in nearly every age category had lower recidivism rates than the other two groups, apart from offenders who were under age 29. For those who were younger than 29, NRDAP completers had the highest recidivism rates. NRDAP completers tended to be older overall though. Taken together, these findings suggest that if the average age for completers was lower, the findings may have been different — in other words, the success of the NRDAP offenders could be partially atributable to differences in the average age.

Additionally, NRDAP completers took longer to recidivate (median of 22 months) than non-completers (20 months) and eligible non-participants (19 months). Non-completers had the highest rates of violent recidivism (34%), followed by non-participants (31.8%) and completers (29.8%). When it comes to drug-related recidivism though, NDRAP offenders had a slightly higher rate of re-arrests (26.3%), followed by non-participants (25.4%), and non-completers (24.1%).

These results seem somewhat promising, but are hard to make sense of. First, differences between groups were not very large. Second, NRDAP completers fared better in terms of overall and violent recidivism, but not for drug-related recidivism. Third, NRDAP non-completers had the highest rate of violent recidivism, the second highest rate of recidivism, and the lowest rates of drug-related recidivism. It is possible that differences in offense, criminal history, and age may be a partial explanation. Similar to the RDAP offenders, as criminal history category increased for each group, so did recidivism rates. Further, the non-completer group had the largest percentage of firearms offenders, which could explain the increased violent recidivism rates. This link is supported by the USSC’s prior work. It still doesn’t explain the differences in overall recidivism, though. Another possible explanation for the NRDAP completers’ lower recidivism rates is that median age for this group was also higher. When parsing out results for age categories, younger people (age 21-29) in the NRDAP group recidivated at higher rates (73%) than those of comparable ages in the other two groups (about 62%).

Authors did control for demographic factors and criminal history in their model, which in theory should controlled for differences between groups. Still though, there is a possibility that large differences in sample sizes may have affected the model’s validity (there were 2,752 completers, 803 non-completers, and 891 eligible non-participants). As stated above, large differences in sample size can skew results, regardless of statistical controls, provided the differences are large enough. The differences in sample size aren’t as vast as those documented for the RDAP offenders, but they are still pretty large. This might also explain why the outcome differences are larger for the RDAP group, as differences in sample size were also larger.

To summarize, NRDAP completers had the lowest overall and violent recidivism rates and eligible non-participants had the highest rates. Furthermore, NRDAP completers had a longer median time to recidivism, compared to those who did not complete the program. However, among younger offenders (age 21-29), who are at the highest risk of re-offending, recidivism rates were higher for NRDAP completers (a difference of nine percentage points). NRDAP completers also demonstrated a slightly higher rate of drug-related recidivism than the other two groups. This is perplexing; if the treatment were effcetive, then presumably, drug recidivism rates would be lowest among people who completed treatment. This recidivism finding also might reflect the fact that the program targets people with persistent substance abuse issues. . It still doesn’t explain how non-completers ended up with lower drug recidivism rates and higher violent recidivism rates, though. It is possible that the program targeted the most serious people and was slightly helpful for people for people who completed it (particularly those older than age 29), hence why non-completers did worse (but only by a small margin) than completers.

Conclusion

The treatment programs reduced the likelihood of recidivism for offenders who completed them, but they did not seem to impact drug-related recidivism. Further, for the NRDAP offenders, results seemed to differ for younger offenders (age 21-29). For those who enrolled but did not complete treatment, fidings were very mixed. Generally, people who did not participate in programs at all fared the worst in terms of overall and violent recidivism, but this was not always the case.

Importantly, there is always potential for a selection bias problem, which may have impacted results. In other words, the type of people completing these types of programs are likely different from those who do not, so it is hard to know whether successes are related to the programs themselves or if there is something about those individuals (e.g. greater motivation to change) that makes them more successful. For example, students may initially enroll in a program for extrinsic reasons, but those who actually graduate are intrinsically motivated and have more enthusiasm about the process. Relatedly, it is not surprising that people who complete the program fare better than those who do not, and this is a common finding in past research. Unfortunately also exacerbates the potential for selection bias in terms of estimating causal effects.

Based on this report, the programs seem somewhat effective in reducing recidivism, but they don’t seem to affect drug-related recidivism as much. While the results are promising, it’s important to keep in mind that results will not be the same for everyone. As such, we should still avoid using blanket statements about in-prison substance abuse programs being effective. More research is needed to better understand the types of people that will succeed in treatment versus those who will not. It is likely that people have different risk factors for criminal activity, some of which might be modifiable and amenable to treatment, while others might be unable to be changed.