Predicting Diversion Program Outcomes Using Drug Testing Information

Introduction

While drug crime-related criminal legal system and victim costs reached $113 billion across the United States in 2007, just $14.6 billion was spent on treating substance use disorder.^[1]
Due, in part, to drug crimes, many U.S. citizens are under correctional supervision, with 1 in 66 adults being on probation or parole in 2020.^[2] In Illinois, 67,587 individuals were on probation in 2020^[3] and 26,426 were on parole.^[4] The supervision population rate of substance use is estimated to be two to three times higher than that of the general population, with nearly half of the people under community supervision having a substance use disorder.^[5]

The Illinois Criminal Justice Information Authority’s Adult Redeploy Illinois (ARI) program was established by the Crime Reduction Act of 2009 to provide financial incentives to local jurisdictions for programs that divert justice-involved individuals from state prisons by providing community-based supervision and individualized services. While researchers have evaluated ARI, models used in some jurisdictions,^[6][7][8][9] research on ARI client outcomes related to the impact of drug testing is limited. The research goals for this study included:

• Quantitatively examining all local ARI program drug test data, including tested drugs, drug test frequencies, and drug test results.
• Systematically examining how ARI drug testing contributes to the possibility of revocation including other factors such as age, sex, and race.
• Proposing recommendations for better program practice to reduce the rate of revocation.

Methods

This study sought to answer the following research questions:

1. How is drug testing being practiced and observed in ARI in terms of its frequency, pass rates, and tested drugs?
2. Does drug testing have a significant impact on ARI participant outcomes when controlling for demographic variables?

The data were derived from ICJIA’s Center for Community Corrections Research ARI Database. The study included 53,159 records of 1,055 individuals collected from October 3, 2011, to June 20, 2019.

Summary of Findings

The median number of drug tests per individual was 19, and the median of the average days between drug tests was 10 days. The most frequently tested drugs also had the highest positive results: heroin (32%), marijuana/THC (30%), cocaine/crack (14%), alcohol (10%), and other opiates (8%). Logistic regression analyses were used to determine what demographic, drug testing, and criminal justice variables predicted program outcomes of completion or revocation.

Among the demographic variables, only age predicted program outcomes. Neither sex nor race emerged as significant program outcome predictors. Drug test positivity rates predicted revocation, as well as drug test frequency (number of times an individual was tested) and average number of days between the drug tests.

Overall, the average drug positivity rate was 29% and most tests were passed with no drug found. Most successful clients who were older women at medium to medium-high recidivism risk and whom tested monthly with lower test positivity rates. Those most likely to experience program revocation were younger men who tested several times per month with higher test positivity rates during their program tenure. Graphing the programs by test positivity, number of tests, and frequency of tests suggests that individuals enrolled in some programs had higher test positivity rates (>50%) and were subject to less frequent drug tests than other programs.

Figure 1

Test Frequency, Drug Presence, and Total Number of Tests

Note: Source was ICJIA’s Center for Community Corrections Research ARI Database.

In Figure 1, each dot represents a county. The size of the dot represents the number of drug tests reported by the site that was included in the study. The larger the dot, the more drug tests reported. The horizontal position of the dot, or Drug Present percentage, was based on the percentage of drug tests where drugs were found. The vertical position of the dot represents the median number of days between tests or Test Frequency. Those counties with dots closer to the zero in the graph had better drug tests outcomes than those in the upper right section of the graph.

Two groups of counties emerged in the drug testing data. Data on the first group, containing Will and Winnebago counties (the larger dots), showed positivity results of less than 25% in the 6,000-25,000 total number of test range, with test frequencies of less than 10 days. Both counties run multi-program courts. These counties had high numbers of tests but relatively low drug test present percentages.

The second group contained Peoria, DuPage, and Macon, all ISP-S programs, showed positivity results of over 50% with less than 1,500 tests taken. St. Clair and McLean counties were outliers, with the highest positivity results, both 64%, after taking between 100 and 400 drug tests and with less than 20 days between each one. These counties had fewer drug tests but relatively high drug test present percentages.

ARI-funded program data showed drug tests occurred every nine days with an average of 29% finding illicit drug use. The drugs most commonly tested for were heroin, cocaine, opiates, marijuana, and alcohol. Age, test frequency, test positivity rates, and risk level predicted program outcome, whereas sex and race were not predictive. Future program outcome studies should include additional variables to improve the regression model.

Conclusion

The research goals for this study were threefold: to quantitatively examine all drug test data, including tested drugs, drug test frequencies, and drug test results; to systematically examine how drug testing contributed to the possibility of revocation when including other factors, such as age, sex, and race; and to propose recommendations for better program practice to reduce the rate of revocation. We determined drug testing frequency and results influenced the possibility of probation revocation.

In the quantitative examination, we discovered a lack of uniformity in drug testing data structures across sites. The data from two of the sites with the largest amount of drug testing data was structured in a way that made it difficult to include information on the actual drugs tested. This is an issue that would require further study for the next study using ARI drug tests. Currently, there are efforts to create a statewide, uniform probation database that would likely include drug testing data. It is our hopes that such data would be used for future studies.

Recommendations garnered from this study include suggesting that the ARI Performance Measurement Committee continue working with the sites to improve the consistency and uniformity of drug test data submitted, continuing to analyze and interpret the results with additional program variables, and exploring how drug test results are most commonly used by ARI grantees, either as a monitoring device that results in sanctions or rewards, as a therapeutic tool to identify treatment needs, or some combination of the two. Recommendations to reduce the rate of revocation can be developed following a future examination on how drug test results trigger rewards, sanctions, or therapeutic adjustments.