|Home | About | Journals | Submit | Contact Us | Français|
The use of screening and assessment tools to gauge substance abuse disorders and the risk for recidivism are two widely recommended practices. A national survey of adult prisons, jails, and community correctional agencies was conducted to examine the practices used to place offenders in appropriate treatment services. Study findings indicate that 58.2% of the surveyed respondents report the use of a standardized substance abuse-screening tool, and that 34.2% use an actuarial risk tool. The provision of higher intensity treatment programs, the use of standardized risk tools, and the provision of more community referral services were all independently associated with the use of a standardized substance abuse-screening tool. Because practices vary considerably, agencies desiring to improve correctional programming should consider different dissemination, implementation, and technology transfer strategies.
Re-occurring discussions have emphasized the importance of prioritizing correctional treatment services for high-risk offenders and the value of using standardized risk and need tools that identify criminogenic factors such as substance abuse (Andrews & Bonta, 1998; Lowenkamp, Latessa, & Holsinger, 2006; Taxman, 2006). This is evidenced by two recent special editions of journals devoted solely to this topic (see Byrne, 2006; Taxman & Marlowe, 2006). The recommended service-delivery process is one that integrates risk-and-need assessments into a unified service-delivery process that links offenders to appropriate substance abuse-treatment services. Mechanisms for screening and assessing offenders’ need for treatment have evolved somewhat differently in the criminal justice and drug treatment systems. Developed by criminologists, risk-assessment instruments address criminal propensity or the factors that predict further involvement in criminal behavior. These tools may incorporate measures of substance abuse, as it is widely recognized that substance abuse is a criminogenic risk factor (i.e., the use of illicit substances on a regular basis increases the odds that an individual will be involved in criminal conduct; see Andrews & Bonta, 1998, for a discussion). substance abuse-screening tools have been developed within the drug-treatment and larger public-health system to identify potential problem areas and the severity of the behavior that may affect a person’s functioning. Formal clinical assessments can provide a more complete picture that includes the severity of the problem behavior and specific issues to be addressed in treatment. Taken together, risk screening, substance abuse screening, and assessment are processes that are central to decisions about referring and placing offenders in programs that are suitable to address their risks and needs.
Although the use of standardized risk tools, substance abuse-screening tools, and assessment practices has been widely advocated in both criminal justice and treatment literature for the past several decades, there remain substantial gaps in our knowledge about the use of valid, extant screening tools in correctional settings. This article addresses this gap by presenting results from a national survey of prisons, jails, and community corrections facilities and also discusses implications of the survey findings for improvements in substance abuse-program placement. Survey results are presented to answer the following questions: (a) What are the patterns of use of standardized screening tools in the correctional system? (b) To what degree is the use related to the availability of treatment programs offered in correctional settings? and (c) What other factors are associated with the use of screening and assessment practices in these correctional settings?
Within the criminal justice system, the use of risk-screening tools to make referral and placement decisions has been discussed as a means to improve consistency and equity in the system and to improve offender outcomes (Gottfredson, 1987; Hoge, 2002; Taxman & Thanner, 2006). The trend for the past 60 years has been to move beyond the nature of the offense (legal criteria) as a proxy for risk and to identify factors through an actuarial statistical model that can predict recidivism or criminal activity (Hoge & Andrews, 1996a, 1996b; Taxman & Thanner, 2006). Within correctional settings, specialized actuarial tools that predict other types of risk (e.g., inmate vulnerability, escape propensity) are sometimes used to make housing and facility-placement decisions. Tools that are used to make decisions about the provision of treatment services and/or referrals do not directly address traditional correctional concerns regarding security and inmate management. Their importance has been less evident to many corrections administrators, although most administrators understand that scarce resources, such as treatment placements, should be allocated based on the needs of offenders.
The potential value of structured risk tools has increased substantially as the tools themselves have advanced in quality and content. Described in generational terms (Andrews, Bonta, & Wormith, 2006; Bonta, 2002; Ferguson, 2002), risk tools were initially limited to clinical judgment, with a focus on the nature of the arrest charge or conviction offense as a proxy for severity of the behavior. The second generation of tools augmented clinical judgments with standardized measures that typically comprised static risk factors, such as age at first arrest, number of prior arrests or convictions, and age of first alcohol or drug use.
Third-generation tools incorporated dynamic factors involving attitudes and beliefs (“criminal thinking”) and service needs (e.g., substance abuse, housing, or employment instability). Third-generation tools solidified the move away from clinical judgment toward proven, valid actuarial items that were statistically predictive of criminal recidivism. The addition of dynamic items also identified needs that are amenable to intervention. These tools serve the dual function of assessing risk-severity level and advancing the concept of risk–need–responsivity (RNR), when correctional and program placements are matched to the level of risk and types of criminogenic needs (Andrews & Bonta, 1998; Grove & Meehl, 1996; Hoge, 1999). Common tools in this generation include the Wisconsin Risk and Needs instrument (WRN; Baird, Heinz, & Bemus, 1979), the Level of Service Inventory–Revised (LSI-R; Andrews & Bonta, 1995), and the Correctional Offender Management Profiling for Alternative Sanctions (COMPAS 4.5; Northpointe Institute for Public Management, Inc., 2004).
Recently, a fourth generation of offender assessment tools has evolved that includes measures of the severity of a given problem area (e.g., substance use disorder, antisocial personality, psychopathy) to improve the level of match between need and service interventions (Andrews, Bonta, & Wormith, 2006; Ferguson, 2002; Jones, Harris, Fader, & Grubstein, 2001). With these advances, the concept of risk is one that focuses attention on the severity of the offending behavior given the lifestyle and needs of the offender.
The high rates of substance-use disorders among offenders (Glaze & Palla, 2005; Karberg & James, 2005; Mumola, 1999; Mumola & Karberg, 2006; Taylor, Fitzgerald, Hunt, Reardon, & Brownstein, 2001) as well as the associations between crime and drug use (Arrestee Drug Abuse Monitoring System [ADAM], 2003; Bureau of Justice Statistics [BJS], 1991; National Institute of Justice [NIJ], 1991; White & Gorman, 2000), have been the basis of persistent calls for instituting substance abuse assessment and treatment systems in correctional settings (National Institute on Drug Abuse [NIDA], 2006; Peters & Wexler, 2005; Taxman, 1998). Clinicians have historically relied on criteria set forth in the Diagnostic and Statistical Manual of Mental Disorders–Fourth Edition (DSM-IV-TR; American Psychiatric Association, 2000) along with other clinically relevant information to determine a diagnosis of substance abuse or dependence. Instead of relying on DSM-IV-TR criteria and a formal diagnostic assessment conducted by a trained clinician, correctional agencies have turned to comparatively abbreviated substance abuse-screening tools that can be administered and interpreted by nonclinical staff (Knight, Simpson, & Hiller, 2002; McLellan, Cacciola, Alterman, Rikoon, & Carise, 2006) and tools that require minimal staff time to administer. When employed appropriately, these substance abuse-screening tools are useful for making a preliminary evaluation about the offender’s need for services. Use of more extensive assessment instruments in correctional settings can help determine the severity of the substance-use disorder, and the extent of other service needs, and thus provide a more refined means of allocating limited treatment resources to high-priority offenders.
Progress on substance abuse-screening tools has advanced at an accelerated pace similar to that seen in the criminal justice field for risk screening and multipurpose risk-and-needs instruments. A number of tools have been developed and studied for use with both the general population and offenders. Commonly used screening tools include the Simple Screening Instrument for Alcohol and Other Drug Abuse (SSI-AOD; Center for Substance Abuse Treatment, 1994), the Texas Christian University Drug Screen II (TCUDS-II; Texas Christian University, 2006), and the Substance Abuse Subtle Screening Inventory (SASSI-3; Miller & Lazowski, 1999). The Addiction Severity Index (ASI; McLellan, Alterman, Cacciola, Metzger, & O’Brien, 1992), a widely used multidimensional problem index, provides measures of severity in several domains (e.g., mental health, legal status, family) in addition to substance abuse. The ASI is similar to the LSI-R and COMPAS instruments, referenced above, in that it examines different aspects of a person’s life. Although none of these tools provides a clinical diagnosis of substance dependency, their scores may be used as an indicator of the severity of substance abuse problems.
Within the criminal justice field, a number of recent studies have illustrated the importance of using valid actuarial risk tools, augmented by problem-severity tools, to improve offender outcomes. Lowenkamp, Latessa, and colleagues have conducted extensive analyses exploring the role of offender risk on outcomes in diverse correctional programs: In two studies involving more than 13,000 offenders participating in 97 correctional programs of varying compositions, they found that offenders who scored high on actuarial risk measures consistently obtained more positive outcomes than lower scoring offenders (Lowenkamp & Latessa, 2005; Lowenkamp et al., 2006). The authors concluded that low-risk offenders are likely to regress in structured treatment programs because participation may expose them to high-risk offenders, remove them from their prosocial networks, and involve requirements that include more scrutiny and increased probability of technical violations. The studied programs involved services that addressed substance abuse as well as other criminogenic factors. A number of substance abuse-treatment studies have demonstrated similar improved outcomes when treating high-risk offenders, including studies of therapeutic community programs (Hiller, Knight, Saum, & Simpson, 2006; Knight, Simpson, & Hiller, 1999), drug treatment courts (Marlowe, Festinger, Lee, Dugosh, & Benasutti, 2006), and probation (Taxman & Thanner, 2006; Thanner & Taxman, 2002). Also, in a national study of substance abuse-treatment programs, the authors found that placing addicts with less severe problems in high-intensity programs would be of little benefit compared to using high-intensity programs for those with significant drug-use problems (Simpson, Joe, & Broome, 2002).
Given that substance abuse is defined as a criminogenic risk factor (Andrews & Bonta, 1998), it is not surprising that studies that have focused specifically on correctional-based drug treatment programming have yielded similar findings. A consensus drawn from the findings of several of these studies is that many offenders are ill-suited for the programs in which they are placed, which in turn contributes to the less-than-favorable outcomes (Belenko & Peugh, 2005; Lowenkamp et al., 2006; Taxman & Marlowe, 2006; Taxman & Thanner, 2006). Research involving drug courts found nearly one half of misdemeanor clients in one program (Marlowe, Festinger, & Lee, 2003) and one third of felony drug court clients in another (Marlowe, Festinger, & Lee, 2004) had a “sub-threshold” drug composite score on the ASI, which is similar to the reported rates for a community sample of non-substance abusers (DeMatteo, Marlowe, & Festinger, 2006). Sub-threshold ASI substance abuse scores have also been found for two thirds of pretrial supervisees in a drug treatment and monitoring program (Lee et al., 2001) and for nearly three fourths of the probationers in an integrated drug treatment–probation model (Taxman & Thanner, 2006). Predictably, offenders in the experimental conditions in all of these studies did not fare better than offenders in the no-treatment or standard practice control or comparison groups. By way of contrast, outcomes of improved performance were routinely observed in this collection of studies for the high-risk offenders who did have evidence of substance abuse disorders.
Results such as these have led many to advocate the principle of responsivity—matching service referrals, placements, and programming based on risk level and type and severity of treatment needs—as a key component in effective correctional treatment (see Andrews & Bonta, 1998; NIDA, 2006). Despite significant advances in the development of valid, standardized tools in recent years, in-depth studies of individual programs highlight inadequacies in the implementation of a service delivery process that links screening to referrals and placement (DeMatteo et al., 2006; Farabee et al., 1999; Knight et al., 2002; Lowenkamp et al., 2006; Taxman & Bouffard, 2000; Taxman & Thanner, 2006). However, it is uncertain whether the results from studies of individual programs reflect national practices.
The use of standardized screening and assessment instruments has been identified as a central front in efforts to bridge the gap between science and practice (Andrews et al., 2006; NIDA, 2006; Peters & Wexler, 2005). Presenting data from a national survey of correctional agencies, this article examines the extent of use of screening and assessment tools and explores factors that may influence their use in both institutional (prison, jail) and community (probation, parole) settings. As stated above, the goals of this article are to address questions about the use of valid, standardized screening and assessment tools in the correctional system and how their use is related to the availability of treatment programs offered. Implications of survey findings for expanding adoption and implementation of screening and assessment practices in correctional settings are also discussed.
As part of the Criminal Justice Drug Abuse Treatment Studies (CJ-DATS), a national research cooperative funded by the NIDA, a survey of state and local program administrators regarding substance abuse services, practices, and policies was conducted. In addition, the survey examined a number of organizational factors (e.g., climate, culture, intra-agency communication, etc.) and working relationships between respondent agencies and other stakeholders. The survey consisted of multilevel questionnaires that were targeted to the census of state administrators and clinical directors of state correctional agencies and to stratified samples of administrators of prisons, jails, and local probation, parole, and/or community correctional agencies. As the first national survey to solicit information on substance abuse treatment and other correctional program practices from agencies throughout the adult and juvenile justice systems, the National Criminal Justice Treatment Practices Survey (NCJTP) provides a uniquely detailed understanding of the prevalence of screening tools, assessment practices, and other treatment-related services for offenders in diverse correctional settings.
This article focuses on data from the stratified sample of facility administrators in the adult correctional system (see Taxman, Young, Wiersema, Rhodes, & Mitchell, 2007, for a complete description of the survey methodology). The prisons selected for the survey were drawn from the sampling frame of adult prisons developed by the BJS and used in their 2000 census of prison systems (Stephan & Karberg, 2003). We first selected with certainty the census of 58 prisons that were designated in the BJS sampling frame as specialized, stand-alone facilities devoted to providing substance abuse treatment. However, of the original sample of 58 drug-treatment prisons, five of the administrators of these prisons indicated that the facility no longer offered drug-treatment services, which was confirmed in the data editing process. The remaining census of 53 drug-treatment prisons was surveyed. In selecting a representative sample of conventional prisons holding general population inmates, we stratified by size of the institution using the BJS stratification criteria (fewer than 12,000, 12,001 to 50,000, and more than 50,000 inmates) and region of the country (eight regions that included the four states with the largest correctional systems with the remaining states grouped into four geographical regions). Prisons were selected randomly at levels proportionate to the number falling in each stratification cell (Three Size Categories × Eight Regions). The final sample was 141 prisons (88 general-population prisons and 53 drug-treatment prisons); 98 (69%) of the wardens or directors of these facilities responded to the survey. The original survey sample excluded diagnostic and reception centers as they have a different function than standard prisons. Telephone interviews with these centers revealed that few used standardized substance abuse-screening tools as part of their intake process.
The sample for local correctional agencies (jails, probation, parole, and other local community corrections) is also modeled after the BJS sample. A similar stratification scheme was used—size of the jurisdiction (fewer than 250,000, 250,001 to 750,000, and more than 750,000) and the same eight regions of the country were employed. The Census Bureau reports that there are 3,100 counties in the United States. We used a two-stage strategy in which the first stage cluster was composed of counties or county equivalents as defined by the U.S. Census Bureau and the National Institute of Standards and Technology (see http://geonames.usgs.gov/fips55.html), and the second stage involved the corrections facilities, offices, or programs within the selected counties. A random sample of 72 counties was selected (3 per cell) based on proportions falling in each strata. Researchers then gathered data through various sources (corrections directories, Web sites, telephone calls, faxed queries, etc.) to determine the correctional agencies operating in the target areas regardless of the government entity responsible for administering or operating the agency or facility. This included those facilities administered by state and regional probation and parole units, local governments, and sheriffs or other elected officials. In addition, we surveyed all diversion programs and specialized case-management agencies, such as Treatment Accountability for Safer Communities (TASC) programs, that operate in these 72 counties. The sample for local community corrections and jail agencies totaled 271, of which 191 (71%) directors representing these agencies responded to the survey.
The 31-page survey was mailed to targeted respondents. Traditional techniques were used to improve response rates, including follow-up phone calls, postcards, and resending the survey instrument (Dillman, 2000). An analysis of respondents and nonrespondents revealed no bias in the characteristics of responders; response rates were equivalent across all stratification units as well as different correctional units. Furthermore, the results of this sampling strategy, when extrapolated to account for national correctional population estimates, were similar to the BJS statistics’ estimates of the national inmate population in state prisons (1.23 million) and the national probation and parole population (4.8 million; see Glaze & Palla, 2005).
A total of 289 adult facility administrators completed the survey. All respondents were responsible for managing the prison, jail, probation, or parole facility or other criminal justice office or local agency referred to as a facility in this article. The study sample included administrators who identified their facility as a prison (34.0%), jail (14.2%), community corrections managed under a state correctional system (24.6%), or community corrections run locally (27.2%). (The difference between state and locally administered agencies is based on the government entity responsible for daily operations of the supervision function.) Each respondent group obtained similar response rates from the survey. The respondents were primarily Caucasian (81.4%) males (69.6%), who were an average of 50.4 years old (SD = 6.9). The majority had a college degree (27.9%), completed some graduate course work (17.4%), or had a master’s degree (33.4%). Only 7.0% reported high school as their highest level of education. The majority of participants identified criminal justice (38.8%), psychology (9.3%), sociology (6.9%), or social work (6.9%) as their primary field of study.
Respondents were asked about the substance abuse screening and assessment tools used in their facilities (from here on, we use “standardized screening” to refer to both screening and assessment instruments). Approximately 31.8% reported that there was no substance abuse screening used, 10.0% reported using an unstandardized screening measure or interview developed for their facility, and the remainder (58.2%) reported use of at least one standardized substance abuse-screening tool. Survey respondents were also asked for a quartile estimate of the proportion of offenders in their facility being screened using a standardized substance abuse tool. Overall, 41.9% of respondents reported no use of standardized screening instruments to anyone under their custody, whereas 11.4% reported administering standardized screening tools to about one fourth of their population, 4.8% reported these instruments being administered to about half of their population, and 41.9% reported using them for about three fourths of their population.
The standardized screening tools that respondents employed at their facility (respondents could report use of more than one tool) included the ASI, used by 46.4% of facilities, followed by the SASSI (42.3%), the TCUDS-II (22.0%), Michigan Alcohol Screening Tool (MAST; 20.8%), the Drug Abuse Screening Tool (DAST; 17.9%), and the Alcohol Dependence Scale (ADS; 14.9%). About 29.8% reported the use of some other standardized substance abuse-screening tool.
When examining the type of correctional setting (facilities) separately, those representing jails and local and state-run community corrections agencies reported little use of standardized screening tools, with fewer than one half reporting the administration of the tools to 75% or more of offenders under their custody (19.5%, 39.2%, and 35.2%, respectively). A higher proportion of respondents from prisons (48.4%) reported using standardized screening tools, with at least 75% or more of their population, whereas three fourths of the respondents from the specialized drug-treatment prisons reported assessing at least 75% of their population with standardized screening tools. Overall, 72% of the respondents reported using a standardized screening tool with more than half of their population in their facility.
Table 1 presents the types of screening tools used by different agencies. Results include the responding drug-treatment prisons as identified in the BJS sampling frame and confirmed through the sampling frame process. Drug-treatment prisons were considered separately from other prison facilities because it was believed that drug-treatment-facility respondents would likely report higher rates of use of substance abuse screening and assessments and skew the estimates for the overall prison sample. As expected, significant differences were found between prison facilities that were identified as specialized drug-treatment facilities and general prison facilities that were not devoted to providing substance abuse treatment or serving any special populations. Use of standardized assessment was reported by 86.1% of drug-treatment facilities, as compared to 53.2% of general prisons, χ2 = 10.94, df = 2, p = .004; drug-treatment prisons were thus analyzed separately.
When examining the five types of facilities, significant differences in the use of standardized screening instruments were found among drug-treatment prisons, general prisons, jails, state-administered community correctional agencies, and locally administered community correctional agencies. There was significantly greater use of standardized tools reported by administrators of drug-treatment prisons (86.1%) as compared to those reporting for general prisons (53.2%), jails (58.5%), state-run community correctional agencies (50.7%), and locally run community correctional agencies (55.7%). Similarly, respondents from drug-treatment prisons were significantly more likely to report that standardized screening tools were administered to a majority of their custody population as compared to other respondents (see Table 1).
Table 2 presents information on the nature of services provided in the facilities by the type of screening tool used. Based on the distribution of the sample, it was possible to divide responses on the amount of substance abuse service provision into three general levels: facilities with no treatment, those where lower intensity treatment (i.e., substance abuse education or group counseling sessions for less than 5 hours per week) was provided, and those with higher intensity treatment (i.e., substance abuse group counseling for 5 or more hours per week or therapeutic community treatment). These distinctions were developed by the team of researchers from CJ-DATS that designed the instrument and reflected their best judgment among various dosage units of programming. Overall, 27.8% of the respondents reported that their facilities did not offer any substance abuse programming, 21.5% provided lower intensity treatment interventions, and 50.6% provided higher intensity interventions.
As expected, nearly all of the respondents for drug-treatment prisons reported offering higher intensity services as well as using standardized screening tools. More than half of the respondents for drug-treatment prisons reported use of the ASI (54.8%), followed by the TCUDS-II (38.7%), SASSI (35.5%), ADS, MAST, DAST (all 16.1%), or some other standardized assessment tool (25.8%). Half of the respondents for general-population prisons reported offering high-intensity treatment services for offenders, 32.3% reported low-intensity services, and 17.7% reported offering no treatment services. Among respondents for general prison facilities, analyses showed the use of standardized tools was significantly more prevalent in facilities with high-intensity treatment services (74.2%), as compared to low-intensity treatment (30.0%) or no treatment (36.4%). Of the general prison respondents reporting use of standardized screening tools, 39.4% reported use of the SASSI, followed by the TCUDS-II, and ASI (33.3%, respectively), MAST (27.3%), DAST (21.2%), ADS (15.2%), or another standardized instrument (24.2%).
Results from the jails showed a distribution of high-intensity treatment (56.1%), low-intensity treatment (29.3%), and no treatment services (14.6%) that was similar to the general prisons breakdown. The comparability of the prison and jail findings—and particularly the relatively high rates of higher intense treatment reported by jail administrators—is somewhat surprising but likely because of the working relationship with community treatment providers in jail settings. In jails, the use of a standardized assessment was not significantly associated with the provision of more intensive treatments. Screening tools reported to be used by jails included the ASI (58.3%), MAST (29.2%), SASSI (25.0%), DAST (16.7%), ADS (16.7%), and TCUDS-II (8.3%).
Fewer than half (47.9%) of the respondents for state-run community correctional facilities reported providing high-intensity treatment programs and, when compared to prisons or jails, a larger percentage of these facilities did not report to have any treatment programs (28.2%). Just less than one quarter (23.9%) of the respondents from probation, parole, or community correctional programs report offering low-intensity treatment programs to their offenders. The presence of a treatment program was not related to the use of standardized screening tools in state-run community corrections facilities. The most common tool reported to be used in these facilities was the SASSI (58.3%), followed by the ASI (47.2%), DAST (25.0%), MAST (22.2%), TCUDS-II (19.4%), and ADS (16.7%). About 30.6% of respondents for state community corrections agencies reported the use of some other standardized tool.
Compared to the state facilities, 54.7% of the locally run community corrections respondents reported offering high-intensity treatment, whereas 24.2% reported low-intensity treatment, and 21.1% reported no substance abuse programs in their facility. As with the jails and state-run community corrections facilities, the use of a standardized assessment tool was not related to the intensity of treatment services in the locally run facilities. The SASSI was administered in 45.5% of these facilities, followed by the ASI (43.2%), MAST (13.6%), DAST, TCUDS-II, and ADS (the latter three each 11.4%). Use of some other standardized assessment instrument was reported by 38.6% of the local community corrections respondents.
The survey included questions about five types of services that could facilitate offenders’ ability to access community-based substance abuse programs after release. These were combined into a simple, additive measure of aids to community referral. The items are (a) developing a treatment/discharge plan, (b) providing referral to a community-based substance abuse treatment provider, (c) making an appointment with a community-based treatment provider, (d) developing personal contact(s) with a community provider, and (e) providing contact information about a sponsor in the community. To make analyses and the presentation of results more easily interpretable, we categorized this factor along the same lines as the intensity of treatment factor, into three levels: none (no referral services provided), low (1 to 3 services provided), and high (4 or more services provided). Inspection of the distribution of the additive referral services variable, which ranged from 0 to 5, suggested these cutoff points.
Table 3 shows the relationships between screening practice and community substance abuse referrals for treatment after leaving the prison or jail facility or in the community corrections facility. Overall, 69.6% of the respondents reported that high-intensity community referral services were provided to offenders under their custody, 28.6% reported providing low-intensity community referral services, and 1.8% did not report any community referral services for their population. Overall, the use of standardized screening tool(s) was associated with the provision of high-intensity community referral services. When the five types of facilities were examined independently to determine the relationship between screening tools and community referral services, a similar pattern emerged whereby the use of a standardized screening tool was associated with the provision of more community referral services.
The survey also allowed for an examination of the use of tools to measure an offender’s risk to public safety, a recommended service-delivery practice. As described above, prior studies indicate that the most commonly used risk tools are the LSI-R and some version of the WRN. Overall, the survey respondents reported that the LSI-R was used in 25.3% of the facilities, and the WRN was reportedly used in 12.7% of the facilities. A small proportion (7.6%) reported using some other risk tool. Overall, just more than one third (34.2%) of the respondents reported using one or more of these risk tools in their facility, and 65.8% did not use any standardized risk tool.
Just less than two thirds (63.9%) of respondents for drug-treatment prisons reported no use of a standardized risk tool, compared to 80.6% of general prison respondents and 80.5% of jail respondents; there was no use of a standardized risk tool in 59.2% of the state-run community corrections facilities and 65.8% of the local community corrections facilities. The LSI-R was reported to be used in 30.6% of the drug-treatment prisons, 11.3% of the general prisons, 7.3% of the jails, 19.7% of the community corrections (state), and 25.3% of the local community corrections. The WRN tool was reported to be used in 2.8% of the drug-treatment prisons, 4.8% of the general prisons, 7.3% of the jails, 21.1% of the state community corrections, and 12.7% of the local community corrections.
Use of the LSI-R was significantly associated with use of standardized substance abuse-screening tools, compared to use of an unstandardized tool and no use of a screening tool (28.0%, 13.8%, and 4.3%, respectively; χ2 = 22.11, df = 2, p < .001). Similar to the LSI-R, the use of the WRN and other risk tools was also found to be significantly related to the use of a substance abuse-screening tool. Finally, when risk tools were tallied and categorized into one category, use of any risk tool was significantly more common in facilities where there was a standardized substance abuse screening tool compared to facilities where an unstandardized tool was used or no tool was used (44.6%, 17.2%, and 9.8%, respectively; χ2 = 36.7, df = 2, p < .001).
The above discussion concerned results from analyses of bivariate relationships between the type of criminal justice setting, community referral practices, use of risk tools, and correctional facilities’ use of standardized substance abuse-screening tool. To assess the relative association of these various factors and their combined effect on the use of substance abuse-screening tools, we developed a series of logistic regression models. Predictors entered into the model included facility type (drug-treatment prison, prison, jail, community, local, or state), intensity of substance abuse treatment provided (none, low, high), and number of community referral services (none, low, high) offered by the facility. We also used facility size, measured by the number of new offender intakes processed in the past year, as a control variable. Several assumptions underscore these analyses including that (a) facility size would influence the ability to acquire a screening tool and use it in decision-making processes, (b) facilities that use risk tools would be more likely to use substance abuse-screening tools, and (c) intensity of treatment services provided and/or referrals undertaken are a function of whether screening and assessment are used to determine type of services needed. These assumptions reflect findings from prior studies and the RNR literature. These exploratory models begin to consider the organizational and programmatic factors that may affect the use of a standardized substance abuse-screening tool in correctional settings.
Two models were developed: one to assess factors associated with the use of a substance abuse-screening tool and one to assess implementation of the substance abuse-screening tool to scale (defined as being used for 75% or more of the offenders under the facility’s custody). Table 4 presents the results of these logistic regression models.
The provision of higher intensity treatment programs, the use of standardized risk tools, and the provision of more community referral services were all independently associated with the use of a standardized screening tool. In addition, use of standardized substance abuse tools was more likely to occur in prisons that are designated drug-treatment facilities (increased odds of providing assessment sevenfold). Overall, the model was significant, χ2 = 89.4, df = 10, p < .001, with 76.3% of the sample correctly classified as providing standardized screening tools based on the predictors in model. Size of the intake population in the facility was not related to use of a standardized substance abuse-screening tool.
Similarly, a logistic regression model was developed to identify factors that predicted use of a standardized tool with the majority (75% or more) of offenders under custody. The same independent predictors were simultaneously entered into the model. Significant independent predictors of providing screening to the majority of the population included offering more community referral services and use of a risk tool. In this model, type of facility was associated with the screening criterion, as use of standardized substance abuse-screening tools with the majority of the population was significantly more common in drug-treatment prisons and general prisons compared to the facilities. Intensity of treatment services offered and size of the facility were not significant predictors in the model. The overall regression model was significant χ2 = 65.0, df = 10, p < .001, with 72.6% of the facilities correctly classified on the screening criterion.
Within the correctional system, the importance of using standardized tools to screen and assess offenders for public safety risk and for substance abuse disorders (and other service needs) has been a dominant theme of the best practices dialogue during the past 20 years. The screening and assessment process is widely recognized as critical for the efficient allocation of resources to ameliorate risk to public safety, to address offenders’ criminogenic needs, and to target resources to improve offender outcomes. The number and quality of tools available to the field has grown significantly during this 20-year period, but questions remain as to whether the field is taking advantage of these advances and using these tools for making treatment-placement decisions. As noted by Latessa, Cullen, and Gendreau (2002),
Effective treatment practices require the appropriate assessment of both the risks posed by, and the needs underlying the criminality of, offenders. When such diagnosis is absent, and no classification of the offender is possible, offenders in effect enter a treatment lottery in which their access to effective interventions is a chancy proposition. (p. 48)
The NCJTP survey findings indicate that 58.2% of correctional facilities nationally are using standardized tools to screen or assess offenders for substance abuse disorders and that 34.2% use a standardized risk tool. In specialized drug-treatment prisons focusing specifically on substance abuse treatment, administration of standardized tools appears to be a routine part of standard practice, although few differences were noted among general prisons, jails, and state or local community corrections facilities in terms of the percentage of reported use of tools in these facilities. Facility type was found to be important with regard to use of screenings with a majority of the custody population, with 75% of the drug-treatment-prison respondents and half of the general-prison respondents reporting use of these tools with at least three fourths of their offenders. This is compared to approximately 35% to 40% of the community corrections facilities and just 19.5% of the jails. As a whole, these findings suggest that adoption of standardized screening measures in community corrections and jails is more of a challenge to the field than is adoption of screening tools in prison settings. The comparatively low rates of use of these tools in community corrections facilities may be partially attributable to administrators in these facilities relying on community-based substance abuse programs to conduct these assessments. Jail administrators also may be less disposed to use assessments because many of their inmates pass through the facility quickly (many jails serve multiple purposes, including booking and detaining large numbers of offenders that are released within 24 to 48 hours). Because of the short stays and quick turnaround in the jail population, jail administrators may see no immediate need to screen for substance abuse severity or service need. In fact, it could be that the jail administrators use de facto “screening” of severity of need among incarcerated pretrial offenders, whereby those who are unable to be released are presumed to be “high risk or high need” and are, therefore, assigned to available treatment options.
In the bivariate analyses, use of standardized substance abuse-screening tools was associated with the provision of more intensive treatment in both types of prison but not in the jails or community corrections facilities. This is consistent with the findings and explanations given above. Bivariate analyses also showed that use of standardized screening tools was associated with other consensus and empirically driven practices, such as employing standardized tools to identify high-risk offenders, and using transitional prerelease referral efforts to assist offenders in linking to treatment and support in the community. The association between use of standardized screening tools (both any use and use with 75% or more of the custody population) and the use of risk tools and community service referrals was further observed in the logistic regression results. Thus, in analyses predicting the use of any screening tool, provision of intensive treatment in the facility was found to be related to the use of screening tools.
Comparing these national survey findings with the Substance Abuse and Mental Health Services Administration’s (SAMHSA’s) 1997 survey of prison and jails, we found diverse trends. The SAMSHA 1997 survey found that screening for substance abuse-treatment needs occurred in 67% of the sampled prisons and 64% of the jails; comparable figures from the NCJTP survey were 53.2% and 58.5%, respectively (SAMSHA, 2006). No gains were apparent in the percentage of prisons where treatment-type services (excluding educational or self-help groups) were offered. This figure was 56% in 1997 and is now 50%. However, provision of treatment services in jails, at 32% in 1997, increased to 56% in the current survey. The SAMSHA survey did not include community correctional agencies.
The survey findings demonstrate that correctional facilities that offer more intensive drug-treatment services are more likely to employ a wider range of strategies to place offenders in services based on need and to transition offenders to another level of care. This is probably attributable in part to the federal block grant devoted to Residential Substance Abuse Treatment (RSAT) programs, where more attention has been on intensive and structured in-prison treatment and the characteristics of effective programs (Farabee et al., 1999; Harrison & Martin, 2001, 2003). RSAT has provided guidance for states in implementing behind-the-bars therapeutic communities with postrelease aftercare, as documented in a 2004 Center for Substance Abuse Treatment (CSAT) Treatment Improvement Protocol on substance abusing offenders (Peters & Wexler, 2005). The use of risk-screening and substance abuse-screening tools to make treatment decisions provides an opportunity and a challenge for correctional programs. NCJTP survey findings highlight some of the issues that need further research and consideration as more research examines the adoption and implementation of standardized tools for assessing offender risk and substance abuse service needs in correctional settings. These issues are summarized below.
Slightly more than a third of the facility respondents reported use of a standardized risk tool. Recent work by Latessa and colleagues highlights the importance of risk screening as a critical component of correctional programming. Their findings indicated that high-risk offenders were more likely to make improvements in specialized programs than were low- to moderate-risk offenders attending these same programs (Lowenkamp et al., 2006; Lowenkamp & Latessa, 2005; Taxman & Thanner, 2006), which is a similar finding to studies of drug-treatment programs that examine outcomes based on problem severity (Simpson et al., 2002). Nonetheless, the survey findings also suggested that, at many facilities, actuarial risk screening is not incorporated into decision making for placement in substance abuse services or other correctional programs. The value of risk tools is that they assist program administers to identify offenders that are likely to be prone to criminal conduct and have a greater need for treatment services, permitting—in the latter instance—the efficient use of scarce treatment resources.
The challenge of providing a unified service-delivery system in correctional agencies has been the subject of much discussion. The consensus is that substance abuse screening should identify offenders to be assessed, and then the assessment should be conducted to place offenders into treatment services based on risk and need factors. Yet the system has had difficulty in implementing all of the components of the RNR model, as demonstrated by results from this survey. Also, individual studies indicate that placement of offenders in existing programs and services in correctional settings are often not based on valid, systematic assessments of their needs and risks (DeMatteo et al., 2006; Lowenkamp et al., 2006; Lowenkamp & Latessa, 2005; Taxman & Thanner, 2006).
The logic behind the risk-problem severity decision-making model is that outcomes can be improved if offenders are matched to appropriate treatment programs. But this is based on the presumption that there is variety in the programs and services available to address different problem-severity issues. If the programs do not vary to address these needs, then the field may want to consider an alternative approach that has not received as much discussion in the public policy arena. If, in fact, substance abuse-treatment programs available to offenders are not considerably different or agencies have difficulty using screening tools in program assignment, then researchers and practitioners may desire to consider the use of tools that measure progress when offenders are involved in treatment programs. This concept may be worth researching to determine whether adaptive treatment strategies may be more feasible in the correctional system as compared to the desired RNR model. A beginning of this discussion can be found in some of the articles in this special edition regarding the Performance Indicators for Corrections (PIC), a project of the CJ-DATS cooperative.
Our findings showed that 58% of the agencies responding to the survey have begun the process of organizational change involved in adopting and implementing appropriate screening and assessment practices. This is a positive finding, especially in light of the respondents who indicated that their facilities have some treatment programs and use different referral strategies. Although the survey cannot address the chronological or causal paths between assessment, treatment, and community referral practices, survey findings suggest that they occur in tandem. Efforts to expand use of assessment tools might be most efficiently targeted at the minority of correctional agencies in states and localities that have so far failed to adopt assessment, treatment, or reentry services. The survey results suggest that these efforts are most needed in jails, community corrections facilities, and typically in locally funded corrections agencies. Despite being responsible for the majority (70% to 80%) of the correctional population, jails and community-supervision agencies are much less likely than prisons to employ assessment, treatment, and community referral practices. Much of the federal focus has centered on prisons through RSAT programs and various, recent reentry initiatives (e.g., Serious and Violent Offender Initiative, Department of Labor Prisoner Reentry Initiative, etc.). Jails and community corrections agencies were largely excluded from these initiatives. Furthermore, many probation and parole agencies may consider these initiatives to be prison innovations; the role of community supervision does not routinely embrace screening and assessment into these reentry efforts. Thus, supervision agencies as well as jails may have neglected to adopt these practices on the presumption that community-based treatment providers will execute them. These agencies must come to accept their role as central to addressing offender service needs in the community, as depicted in new models of community supervision (Taxman, 2002; Taxman, Shepardson, & Byrne, 2004).
Assessment results from the NCJTP survey have some notable limitations. Again, it should be recognized that this is a cross-sectional survey that has its inherent limits. A longitudinal survey would be most desirable as well as companion qualitative interviews with respondents to understand some of the nuisances of practice. For the most part, a cross-sectional survey suffers from a lack of contextual information to explain some of the findings. One example is an observation made earlier—that results for the community corrections agencies may be partially attributable to administrators presuming that screening and assessment of offenders under their custody are being done by other agencies or explicitly making arrangements for outside assessments. Another such limitation applies primarily to the general prison facilities, where results may be influenced by the fact that some correctional systems designate specialized facilities for assessment and classification and thus may not support wide use of assessment in their nonspecialized, general-population facilities. Furthermore, we do not know whether assessment findings are affected by agency funding requirements or by legal decisions or court orders on conducting assessments.
Future studies, particularly qualitative ones that examine the issues identified above, can also add precision and depth in measuring factors that we found to be associated with use of assessment. The multivariate analyses in our research included a limited number of factors coded from simple counts and groupings of measures. Additional, more detailed facility-level information on the type and extent of the full range of programming that may be provided in correctional settings would provide a more complete picture of factors associated with the use of assessments. Research on different kinds of variables that may be related to use of standardized assessments is part of the ongoing research agenda of the CJ-DATS network and, specifically, studies involving the NCJTP survey. Results from the survey may assist understanding of the role of organizational factors in fostering or hindering use of assessment.
At a more fundamental level, the current research cannot address whether facilities’ limited use of assessment is the basis for inadequacies in treatment matching and responsivity in corrections—that is, facilities’ frequently narrow, “one-size-fits-all” conception of treatment programming (Latessa et al., 2002). From one perspective, it makes little sense for correctional agencies to invest in systematic use of standardized need and risk-severity tools if they do not provide services that are an appropriate match for the results of these assessments. An alternative, more hopeful view is that wider use of assessments can help drive demand for the kinds of tailored, responsive treatment networks that are needed in correctional systems. A related, and perhaps more pragmatic, conclusion is that standardized assessments are needed for making efficient use of the limited treatment resources available. Expanding use of assessment tools beyond the limited applications evident in the survey, coupled with the provision of appropriately matched services, could lead to more high-need, high-risk offenders being referred to treatment and, ultimately, to improved treatment and public safety outcomes.
AUTHORS’ NOTE: This study was funded (Grant No. U01 DA16190) under a cooperative agreement from the U.S. Department of Health and Human Services, Public Health Service, National Institutes of Health, National Institute on Drug Abuse (NIH/NIDA). The authors gratefully acknowledge the collaborative contributions by federal staff from NIDA and the 10 research centers that were participating sites in the National Survey of Criminal Justice Treatment Practices survey, which is one of the Criminal Justice Drug Abuse Treatment Studies (CJ-DATS) projects, as explained in the introduction to this edition of Criminal Justice and Behavior.
FAYE S. TAXMAN, George Mason University.
KAREN L. CROPSEY, University of Alabama, Birmingham.
DOUGLAS W. YOUNG, University of Maryland.
HARRY WEXLER, National Research & Development, Inc.