4Screening and Assessment:An Evidence-Based Processfor the Management and Careof Adult Drug-Involved Offenders

Matthew L. Hiller, Steven Belenko,Wayne N. Welsh, Gary Zajac, and Roger H. Peters

AbstractValid and reliable assessment of risk and needs is a cornerstone of evidence-based practices with offenders who use and abuse drugs. They provide theneeded clinical information upon which the case planning and services refer-ral and delivery processes observed in criminal justice settings are based.However, recent surveys of nationally representative samples shows criticalgaps remain, with many criminal justice programs either forgoing assessmentof risks and needs or using instruments that have not be externally validated.To encourage more widespread use of risk and substance abuse instrumentsthat have been shown to be reliable and valid, the current chapter reviewsa number instruments within the context of the Risk-Need-Responsivity(R-N-R) model for assessment and services planning. Descriptions of theseinstruments as well as their reliability and validity when used with offendersamples are presented. Discussion focuses on the need to adhere to evidence-based practices and processes when assessing and managing offenders withdrug abuse problems.

KeywordsCriminogenic risk • Substance abuse • Assessment • Review of instruments

Introduction

Based on the time-tested medical model forscreening and assessing patients for disease, inorder to refer and triage cases into appropriate

M.L. Hiller (�)Department of Criminal Justice, Temple University,Philadelphia, PA 19122, USAe-mail: mhiller@temple.edu

care, the screening and assessment of crimino-genic risk and needs represent an evidence-basedprocess to guide the care and management ofdrug-involved offenders in correctional institu-tions and in the community. Several recent stud-ies and monographs illustrate the need for suchevidence-based procedures.

For example, guidelines from the NationalInstitute of Corrections (Clawson, Bogue, &

45C. Leukefeld et al. (eds.), Handbook of Evidence-Based Substance Abuse Treatmentin Criminal Justice Settings, Issues in Children’s and Families’ Lives 11,DOI 10.1007/978-1-4419-9470-7_4, © Springer Science+Business Media, LLC 2011

46 M.L. Hiller et al.

Joplin, 2005) identify the development and main-tenance of a system for ongoing offender riskscreening/triage and needs assessments as its firstevidence-based principle for effective correction-based treatment. Further, the National Instituteon Drug Abuse (NIDA; 2006) indicates that“assessment is the first step in (the) treatment (ofdrug-involved offenders)” (p. 2).

Indeed, the development and testing of screen-ing and assessment tools for use with criminaljustice samples was a major focus of the recentlycompleted NIDA-funded Criminal Justice DrugAbuse Treatment Studies (CJDATS; Simpson &Knight, 2007).

Despite such national leadership, a recentnational representative survey, the NationalCriminal Justice Treatment Practices Survey(NCJTPS) (also completed as a part of theCJDATS research cooperative) found thatalthough 58.2% of correctional facilities (includ-ing prisons, jails, and probation, parole, andcommunity corrections agencies) used at leastone standardized tool to screen for substanceabuse problems, 10% reported using nonstan-dardized and/or internally developed screens,and 31.8% reported no use of a substance abusescreening tool (Taxman, Cropsey, Young, &Wexler, 2007; Taxman, Young, Wiersema,Rhodes, & Mitchell, 2007). Findings from anational survey of adult drug court programs,implemented as a part of the eCourt project ofthe CJ-DATS research cooperative, were similar(Taxman, Perdoni, Young, Belenko, Hiller, inpress). Specifically, 68% of drug courts surveyedreported using a standardized substance abusescreening tool, while 21% used a standardizedrisk assessment, but only 4% used a standardizedmental health screen. Taxman and colleagues(in press) noted, “Survey results demonstratethat ‘legal criteria’ (e.g., type and severity ofcharge and whether the defendant has a historyof violent offenses) rather than clinical criteria(e.g., drug dependence severity, risk) are the mostdeterminative factors in deciding who is allowedentry into drug courts and who is not.” Ninety-two percent of drug courts involve members ofthe “legal team” (including the judge, prosecutorand defense attorney) in the admission process

and nearly 80% of courts involve coordinatorsand case managers, while 48% of drug courtsinvolve treatment providers.” (p. 14).

Given that substantial gaps are evident inhow existing practice conforms to evidence-based guidelines for screening offenders andthat research shows that standardized screeningand assessment information often is not used toinform services delivery, the focus of the currentchapter is on a commonly cited model, Risk-Need-Responsivity (R-N-R; Andrews, Bonta, &Hoge, 1990), for offender assessment and reha-bilitation. Commonly used instruments in thecriminal justice system are reviewed regardinghow they complement the R-N-R framework,and promising instruments, recently developedas a part of the CJDATS research collaborative,are reviewed regarding their potential to improveoffender assessment and care.

The perspective taken here, regarding theassessment of offenders with drug and alcoholproblems, is that assessment is an ongoing anditerative process. It begins when brief screeninginstruments are administered to offenders to iden-tify those who may have problems and who arein need of more in-depth assessment and diagno-sis. In turn, when diagnostic assessments identifysignificant clinical problems, this information isused to develop an individualized treatment plan.Subsequent to the initial treatment plan, ongo-ing assessment monitors treatment progress andidentifies any needed modification to the treat-ment plan (Knight, Simpson, & Hiller, 2002;Roberts, Contois, Willis, Worthington, & Knight,2007; Simpson & Knight, 2007). Unfortunately,ongoing assessment is not often accomplishedin the criminal justice system (Peters & Wexler,2005), particularly for those offenders movingbetween different points in the system (e.g.,courts, jails, prisons, community corrections).Programs often fall short in the provision ofinitial assessments, making subsequent reassess-ments impossible and jeopardizing any possibil-ity of continuity of care as the offender progressesthrough the criminal justice system. Indeed, sys-tematic program reviews have found that the lackof rigorous assessment is often one of the greatestweaknesses within offender programs (Hubbard,

4 Screening and Assessment: An Evidence-Based Process for the Management and Care of Adult. . . 47

Travis, & Latessa, 2001; Latessa & Holsinger,1998; Lowenkamp, Latessa, & Smith, 2006).

Risk-Need-Responsivity:An Evidence-Based Modelfor Offender Assessment

One conceptual model used to guide offenderassessment and rehabilitation, Risk-Need-Responsivity (R-N-R; Andrews et al., 1990;Andrews, Bonta, & Wormith, 2006; Bonta &Andrews, 2007; Ogloff & Davis, 2004), providesa framework for focusing on specific offenderattributes in relation to planning services forrehabilitation and for preventing recidivism.Developed by Andrews and colleagues (1990),and grounded in general personality and socialpsychology theory, this model was introducedin the late 1980s and early 1990s as a set ofevidence-based guidelines to direct the imple-mentation of treatment services in correctionalsettings. This approach ran counter to theprevailing antirehabilitation sentiment of thattime. Incorporating elements of differentialassociation, operant conditioning, and sociallearning theory, R-N-R is specifically interestedin the assessment of individual characteristics(like criminal history and antisocial attitudes)predictive of criminal behavior, “making it aparticularly useful guide for both assessing therisk of recidivism and planning rehabilitationattempts” (Ogloff & Davis, 2004, p. 230). Itprovides several empirically based principles formatching services to an offender’s assessed risksand needs (e.g., the Risk Principle, the NeedsPrinciple; Andrews & Dowden, 2007).

The empirical literature on the prediction ofrecidivism provides the evidence-base for theassessment of an individual’s risk (Andrewset al., 1990; Andrews & Dowden, 2007; Andrewset al., 2006). Risk factors may be either staticor dynamic characteristics, with the latter rep-resenting the best possible targets for rehabili-tation. Andrews and Bonta (2006) and Andrewsand Dowden (2007) highlight the “Big 8” riskfactors that account for the largest amount ofvariance in the prediction of recidivism. These

include a history of antisocial behavior, antiso-cial personality patterns, antisocial cognitions,antisocial associates, family problems, low lev-els of performance in school/work, low levels ofinvolvement with noncriminal leisure activities,and substance abuse. The level of an individ-ual’s assessed risk (i.e., high versus low) informsthe application of the “Risk Principle,” whichindicates that more intensive services shouldbe reserved for those at high risk, while thosewith lower risk levels receive either minimal orno intervention (Andrews et al., 1990; Taxman,Thanner, & Weisburd, 2006; Thanner & Taxman,2003). This principle also suggests that placinglow-risk offenders in a very intensive interventionmay have unintended harmful effects.

Needs (also referred to as treatment needs,Ogloff & Davis, 2004) are comprised of twosets of dynamic characteristic, criminogenic andnoncriminogenic. Criminogenic needs are a spe-cific subset of dynamic (modifiable) risk factorsthat are predictive of recidivism. In particu-lar, these include antisocial attitudes, antisocialfeelings, substance abuse, poor parental bond-ing and parenting skills, and antisocial peers(Dowden & Andrews, 2000). Noncriminogenicneeds are not directly related to the proba-bility of reoffending, and include poor self-esteem, anxiety, psychological distress, feelingsof alienation, and socially disorganized neigh-borhoods (Ogloff & Davis, 2004). Because cor-rectional and other criminal justice interventionsare expected to reduce recidivism, the “NeedPrinciple” states that criminogenic needs shouldbe the primary treatment targets. Although theNeed Principle does not specifically discour-age addressing noncriminogenic needs, it cau-tions researchers and clinicians that expectedoutcomes related to addressing these shouldnot include recidivism. Some evidence existsthat programs that primarily target criminogenicneeds, as opposed to noncriminogenic needs,and that target more criminogenic needs, havebetter outcomes (Latessa & Lowenkamp, 2006;Lowenkamp & Latessa, 2002). Ongoing assess-ment of criminogenic needs during treatment isrecommended to monitor change.

48 M.L. Hiller et al.

Responsivity refers to characteristics of theservices delivered (general) and attributes of theindividual offender (specific) that can modify theimpact of treatment. For example, with respect tospecific responsivity, one’s level of motivation fortreatment has been shown to affect engagement intreatment (Hiller, Knight, Leukefeld, & Simpson,2002) as well as treatment outcome (Broome,Knight, Knight, Hiller, & Simpson, 1997). Otheroffender characteristics that may affect treatmenteffectiveness include intellectual ability, learningstyles, and self-esteem (Ogloff & Davis, 2004).General responsivity refers to factors related tothe treatment episode, such as intensity and typesof services received, therapeutic relationships,and interactions with counselors (Andrews et al.,1990; Broome, Flynn, Knight, & Simpson, 2007;Ogloff & Davis, 2004; Simpson, Joe, & Broome,2002).

Specific application of R-N-R clearly indi-cates that the most important targets for assess-ment are the dynamic risk factors (needs) of indi-vidual offenders. Treatment plans should specif-ically address these needs, reserving the highestlevels of treatment for those with the highestrisk and greatest needs. Beyond initial assess-ment of the individual, R-N-R highlights the needfor ongoing assessment of changes in crimino-genic needs, to provide feedback on how treat-ment plans should be modified to improve thelikelihood of favorable outcomes (i.e., reducedrecidivism). For example, the Level of ServiceInventory-Revised, LSI-R, includes versions forbaseline and followup administrations. In addi-tion, assessment also should include measure-ment of both individual differences and char-acteristics of the services provided to identifyways to improve the offender’s response to theseservices.

Screening and Assessmentin the Criminal Justice System

A wide variety of screening and assessmentinstruments is available to guide custody andtreatment decisions for offenders with drug prob-lems in the criminal justice system but, as

discussed below, only a few of these are incommon use. These instruments can be placedinto two broad categories, risk assessment andsubstance abuse assessment, with the formertending to focus on a wider spectrum of the“Big 8” and the latter on a more narrowlydefined subset, typically only on substance abuseproblems. Substance abuse assessments, unlikerisk assessments, typically focus on measuringsymptoms, based on a diagnostic standard likethe Diagnostic and Statistical Manuals of theAmerican Psychiatric Association.

As noted by Andrews et al. (2006), risk assess-ment has evolved over the past several decades,beginning with first-generation (1G) assess-ments, which emphasized unstructured profes-sional judgments in rating an individual’s riskfor recidivism. These ratings were shown to beinconsistent across practitioners and ineffectivefor accurately quantifying risk. To address this,a second generation (2G) of instruments adoptedan “actuarial approach” whereby static risk fac-tors for recidivism (offender characteristics likeage and prior offense history which predictedrecidivism but were not amenable to change)were aggregated into summative indices as anobjective prediction of the extent to which onewas at risk for recidivism. Because 2G instru-ments focused on aspects of the person that couldnot be changed through therapeutic interventions,third- generation (3G) instruments were devel-oped to include dynamic risk factors, includingmany of the “Big 8” (described above), that couldbe specifically targeted for change. The mostwell known of these is the Level of ServicesInventory – Revised (LSI-R), which is describedin more depth below.

Recently, fourth-generation (4G) instrumentshave emerged, which extend the collection ofassessments beyond initial baseline measuresto include case planning and management andresponsivity assessment. Because of their new-ness, relatively little empirical information isavailable and, thus, these will not be a focus of thecurrent chapter. One example of a 4G instrumentthat is currently being studied and shows promiseis the Correctional Offender ManagementProfile for Alternative Sanctions (COMPAS)

4 Screening and Assessment: An Evidence-Based Process for the Management and Care of Adult. . . 49

tool (Brennan, Dieterich, & Ehret, 2009; Fass,Heilbrun, DeMatteo, & Fretz, 2008). Another isthe Level of Service/Case Management Inventory(LS/CMI; Andrews, Bonta, & Wormith, 2004;Andrews & Bonta, 2006), which is based on the3G LSI-R.

There are numerous instruments for screeningrisk and substance abuse in the criminal justicesystem and a comprehensive review of all of theseis beyond the scope of this chapter. Instead, thechapter reviews specific instruments that consti-tute “common practice” (i.e., are the most com-monly used) among criminal justice programs(including drug courts) and discussess the evi-dence base for each. As will be shown below, notall instruments commonly used in the criminaljustice system have a strong body of empiricalresearch that supports their use with these pop-ulations. Most notable among these is the com-monly used Substance Abuse Subtle ScreeningInventory (SASSI), for which empirical researchhas failed to confirm the validity of the parts ofthe instrument related to the subtle assessment ofdrug problems.

Whether an assessment is viewed as beinga part of “common practice” is based on find-ings from recent surveys of treatment practices(including screening and assessment) on nationalsamples of correctional programs and drug courts(see Taxman, Cropsey et al., 2007; Taxman et al.,under review). As shown in Table 4.1, despitethe availability of best practices standards fromNIC and NIDA, which include screening andassessment as central components, a majorityof agencies and programs do not use a struc-tured risk assessment instrument, and many donot use a standardized substance abuse tool.For structured risk instruments, the LSI-R is inmost common use (e.g., 25.3% of the NCJTPSand 17.7% of the eCourt sample) followed bythe Wisconsin Risk/Needs Assessment (WRN).With respect to substance abuse, the AddictionSeverity Index (ASI) is the most commonly usedassessment in both correctional systems and drugcourts (46.4 and 44.7%, respectively) followed bythe Substance Abuse Subtle Screening Inventoryin 42.3 and 23.4% of programs respectively.Other substance abuse screening instruments

Table 4.1 Use of screening and assessment instrumentsreported in National System Surveys of Institutional andCommunity Corrections and Drug Courts

Instrument NCJTPSa eCourtb

Risk-needs

Use no standardized risk tool 65.8 79.0

LSI-R 25.3 17.7

WRN 12.7 4.3

Substance abuse

Use no standardized substanceabuse tool

42.8 32.0

ASI 46.4 44.7

SASSI 42.3 23.4

TCUDS 22.0 5.0

MAST 20.8 9.9DAST 17.9 5.7aBased on Taxman, Cropsey et al. (2007).bBased on Taxman et al. (in press).

in common use in the criminal justice systemare the Texas Christian University Drug Screen(TCUDS), the Michigan Alcohol Screening Test(MAST), and the Drug Abuse Screening Test(DAST).

The most commonly used risk assessments(i.e., the LSI-R and WRN) are reviewed next,followed by screening and assessment tools forsubstance abuse problems (i.e., ASI, SASSI,TCUDS, MAST, and DAST), presented in theorder of most to least commonly used. An instru-ment’s being commonly used does not neces-sarily mean it is evidence-based; and specialemphasis is placed on describing the evidencebase (i.e., reliability/validity) of each instrument.Screening and assessment tools must be validand reliable in order for them to be useful forinforming clinical planning and practice (Knight,Simpson, & Hiller, 2002; Peters, Greenbaum, &Edens, 1998; Peters et al., 2000). In addition,information is provided on whether each is inthe public domain, requires specialized training,is interviewer- or self-administered, and an esti-mate is provided of the length of time required tocomplete each. These reviews are followed by adescription of a number of instruments that wererecently developed and tested in NIDA’s CJDATSresearch cooperative.

50 M.L. Hiller et al.

Risk Assessments

Level of Supervision Inventory-R (LSI-R). TheLevel of Service Inventory-Revised (LSI-R;Andrews & Bonta, 1995), a 3G instrument, hasbecome a popular assessment for the classifica-tion and management of offenders both withincorrectional institutions and in the community.As an objective, actuarial instrument that is basedon theory (i.e., the R-N-R principles of effec-tive correctional intervention described above)and empirically validated on diverse samples ofoffenders, the LSI-R attempts to meet the dualtask of managing offenders and assessing theirneeds (Bonta, 2002).

The LSI-R is a 54-item instrument comprisedof subscales measuring 10 different risk/needareas: criminal history, education and employ-ment, financial circumstances, family and maritalsituation, accommodations/housing, leisure andrecreation, companions, drug and alcohol abuse,emotional and personal characteristics, and atti-tudes and orientations. The 10 domains, takenindividually, allow correctional agents to identifyand target areas of criminogenic need. For exam-ple, if the offender scores high on the alcohol anddrug problem subscale, this should trigger a refer-ral to substance abuse treatment programming inprison, the community, or both.

The instrument has been validated on a widerange of offender groups including probationers(Andrews, Kiessling, Mickus, & Robinson,1986), male inmates (Bonta & Motiuk, 1987,1990; Lowenkamp, Hodsinger, & Latessa,2001; Loza & Simourd, 1994; Simourd, 2004),female inmates (Coulson, Ilacqua, Nutbrown,Giulekas, & Cudjoe, 1996), juvenile offenders(Shields & Simourd, 1991), and sexual offenders(Simourd & Malcolm, 1998). A meta-analysissummarizing 30 predictive studies of the LSI-R(Gendreau, Goggin, & Smith, 2002) found thatLSI-R total scores were significantly corre-lated with both general and violent recidivism.Although fewer studies have examined thepredictive validity of individual subscales, Kellyand Welsh (2008) found that both the drug andalcohol subscale and the LSI-R total score sig-nificantly predicted reincarceration in a sample

of drug-dependent adult offenders released fromprison. The LSI-R and its subscales have beenshown to be both reliable and valid (Simourd,2006). There are costs associated with the train-ing for, and use of, this instrument. The LSI-Ris not in the public domain, requires specialtraining, and is most commonly administered asan interview. It takes between 45 minutes and 2hours to complete.

Wisconsin Risk/Needs Assessment (WRN).Among the first 2G actuarial risk assessmentsto be developed, it has experienced a renewedinterest (Baird, 2009; Eisenberg, Bryl, & Fabelo,2009). It focuses primarily on static risk factorslike age at first conviction, number of proba-tion/parole supervision periods and revocations,and number of address changes in the previ-ous 12 months (Baird, Heinz, & Bemus, 1979,1981; Eisenberg et al., 2009). Typically scoredby staff using information from an individual’sPresentencing Investigation Report, the instru-ment represents an efficient alternative to lengthyinterviews. However, the evidence base forthis instrument, including descriptions of valid-ity and reliability, comes primarily from pub-lished reports from Department of Correctionsresearchers, including Wisconsin and Texas, withonly a few examples evident in the scientific lit-erature (e.g., Clear & Gallagher, 1983; Knight,Simpson, & Hiller, 1999). Overall, findings showthat the scale is reliable and valid, and can makeadequate predictions regarding risk for recidi-vism. It is in the public domain and can thereforebe used at no cost.

Substance Abuse Instruments

Addiction Severity Index (ASI). Perhaps the mostwell-researched psychosocial assessment instru-ment for use with substance abusers both in thecommunity and within criminal justice treatmentsettings, the ASI was originally developed in theearly 1980s and has since undergone several revi-sions (McLellan, Luborsky, Cacciola, & Griffith,1985; McLellan, Luborsky, Woody, & O’Brien,1980; McLellan et al., 1992). It is organizedinto seven assessment areas, including medical

4 Screening and Assessment: An Evidence-Based Process for the Management and Care of Adult. . . 51

problems, employment, alcohol use, drug use,legal problems, family/social relationships, andpsychiatric problems. A set of interviewer sever-ity ratings (ISRs) and composite scores (CSs)can be calculated from the information col-lected during this interview, which can be com-pared to those typical for clients in the pro-gram, to help inform treatment planning. Severalof the ASI assessment areas clearly map intothe “Big 8” criminogenic needs (see Table 4.2),including a history of antisocial behavior, disor-ganized family/marital relationships, low perfor-mance in education/work, and substance abuse.In addition, other assessment areas captureinformation relevant for identifying noncrimino-genic needs (e.g., mental health). The ASI CSscores have been found to be highly reliableand valid for numerous populations, includingthose in the criminal justice system (Alterman,Bovasso, Cacciola, & McDermott, 1994; Hanlon,O’Grady, & Bateman, 2000; Leonhard, Mulvey,Gastfriend, & Shwartz, 2000; Mäkelä, 2004). TheISR scores have been found to be less reliableand valid, varying with the level of interviewertraining (Alterman et al., 2001; Mäkelä, 2004).Recent research shows that ASI CS scores arehighly predictive of DSM-IV alcohol and drugdependence diagnoses (Rikoon, Cacciola, Carise,Alterman, & McLellan, 2006). The ASI is in thepublic domain, should be completed in an inter-view, and specialized training is available. Whenadministered by a trained interviewer, it takesapproximately 40 minutes to complete.

Substance Abuse Subtle Screening Inventory(SASSI). Originally released in 1985 and revisedtwice (Miller, 1985; Miller & Lazowski, 1999;Swartz, 1998), the SASSI is a 93-item paper-and-pencil self-report screening instrument thatincludes a total of nine direct and indirect scalesfor identifying individuals, who may have sub-stance abuse disorders, as candidates for morein-depth assessment. Direct scales for measuringsubstance abuse problems include the face-validalcohol, face-valid other drug, and symptomsscales. Reflecting the belief that some individu-als, like offenders, have a vested interest in lyingabout their drug use, scales that are purported tobe indirect measures of substance abuse problems

(i.e., obvious attributes, subtle attributes, anddefensiveness) also are included to detect sub-stance abuse problems among those who havea vested interest in concealing it. Two otherscales, family versus control subjects and cor-rections, are also indirect measures of substanceabuse problems because they measure its cor-relates (family problems, criminal history). Afinal scale, random answering pattern, is includedas a validity scale. As shown in Table 4.2, thethree direct substance abuse scales and the fam-ily and correctional scales relate most closely tothe criminogenic risk factors of the R-N-R modeland, thus, measure domains relevant to reducingrecidivism.

A comparative analysis of the SASSI andseveral other screening instruments that areused with offenders determined that the SASSIwas the least effective in identifying sub-stance dependence disorders (Peters et al., 2000).Furthermore, a recent monograph, reviewing theliterature on screening and assessment amongoffenders, published by the National GAINSCenter (Peters, Bartoi, & Sherman, 2008) indi-cates that the SASSI is significantly less effectivethan other instruments in detecting substance usedisorders among offenders. The monograph rec-ommends that the SASSI should be avoided foruse in criminal justice settings due to concernsabout its validity. The common use of the SASSI,in the criminal justice system, underscores theimportance of knowing whether the instrumentbeing used has an adequate evidence base: espe-cially because best practices emphasize the needto use instruments that are both reliable and validfor the population for which they are being used(Knight, Simpson, & Hiller, 2002).

Recent meta-analyses (Feldstein & Miller,2007; Miller, Woodson, Howell, & Shields, 2009)examining SASSI’s use among primarily nonof-fender populations, have also focused on the psy-chometric properties (i.e., reliability and validity)of this inventory. The general finding of thesemeta-analyses is that the direct scales and the cor-rections scale are the most reliable and valid partsof the SASSI but the indirect scales have poormeasurement properties. Feldstein and Miller(2007, p. 40) conclude “We found no evidence

52 M.L. Hiller et al.

Tab

le4

.2M

appi

ngof

scre

enin

gan

das

sess

men

tins

trum

ents

toth

eR

-N-R

fram

ewor

k

Inst

rum

ent

Scre

enor

asse

ssm

ent

“Big

8”R

isk/

Nee

ds(A

ndre

ws

&D

owde

n,20

07)

Ant

isoc

ial

beha

vior

Ant

isoc

ial

pers

onal

ityA

ntis

ocia

lco

gniti

ons

Ant

isoc

ial

asso

ciat

esFa

mily

/mar

ital

stat

usSc

hool

/wor

kL

eisu

reac

tiviti

esSu

bsta

nce

abus

eO

ther

risk

s/re

spon

sivi

ty

LSI

-RA

sses

smen

tY

esY

esY

esY

esY

esY

esY

esY

esY

es

WR

NSc

reen

Yes

No

No

No

No

Yes

No

Yes

No

ASI

Ass

essm

ent

Yes

No

No

No

Yes

Yes

No

Yes

Yes

SASS

ISc

reen

Yes

No

No

No

Yes

No

No

Yes

No

TC

UD

SSc

reen

No

No

No

No

No

No

No

Yes

Yes

MA

STSc

reen

No

No

No

No

No

No

No

Yes

No

DA

STSc

reen

No

No

No

No

No

No

No

Yes

No

IPA

SSA

sses

smen

tY

esN

oN

oN

oN

oY

esN

oY

esY

es

CO

DSI

Scre

enN

oN

oN

oN

oN

oN

oN

oY

esY

es

CE

STA

sses

smen

tN

oY

esY

esY

esN

oN

oN

oN

oY

es

TC

UC

TS

Ass

essm

ent

No

Yes

Yes

No

No

No

No

No

No

CA

IA

sses

smen

tN

oY

esY

esN

oN

oY

esN

oY

esY

es

4 Screening and Assessment: An Evidence-Based Process for the Management and Care of Adult. . . 53

to support claims that the indirect scales of theSASSI offer a unique or additive advantage tocorrectly detecting current substance use disor-ders.” In addition, Feldstein and Miller (2007)indicate an average false positive rate of 38%.The authors of both meta-analyses suggest thatit may be more cost effective to use other screen-ing instruments (like the MAST, see below) thatare “shorter, have equal or better reliability, andare in the public domain”. The SASSI is not inthe public domain, is typically completed as aself-administered instrument in less than an hour,and requires specialized training to hand score.Computerized scoring programs are available forpurchase.

Texas Christian University Drug Screen(TCUDS). Although, when compared to theASI, SASSI, MAST, and DAST, the TCU DrugScreen is a relative newcomer to the field, it isimplemented in some large correctional systems,including the Texas Department of CriminalJustice and the Pennsylvania Department ofCorrections (Knight, Simpson, & Hiller, 2002).This screen includes a face-valid measure withitems that closely map to the Diagnostic andStatistical Manual IV (American PsychiatricAssociation, 1994) as well as additional ques-tions for assessing the extent to which theindividual has used 11 substances (includ-ing alcohol, marijuana, hallucinogens, crack,cocaine, heroin, heroin and cocaine mixed, streetmethadone, other opiates, methamphetamine,and tranquilizers) (Knight, Simpson, & Morey,2002). As such, the majority of this instru-ment measures substance abuse as a criminogenicrisk/need. However, two supplemental questionsask respondents to indicate whether they haveever been in substance abuse treatment and theextent to which they feel they currently needtreatment. These items fit more closely withspecific responsivity in the R-N-R model (seeTable 4.2).

Perhaps related to its relative newness, theTCUDS has been subjected to less empiricalinvestigation than the other measures coveredin this section, and most of what has been doneis not published in the peer-reviewed literature(e.g., Hiller & Narevic, 2005; Knight, Simpson,

& Hiller, 2002; Knight, Simpson, & Morey,2002). One notable exception is a study con-ducted by Peters and colleagues (2000). Thisstudy administered eight different substanceabuse screening instruments to 400 inmates,including the drug use scales from the ASI,the Alcohol Dependence Scale, DAST, MAST,SASSI-2, TCUDS, and the Simple ScreeningInstrument. The clinical assessment “goldstandard” against which the performance ofthese screens was judged was provided by thesubstance use disorders module of the StructuredClinical Interview for the DSM-IV. Findingsshowed that the ADS, SSI, and TCUDS werethe most effective for identifying substance usedisorders. The TCUDS had the highest sensitiv-ity (.85) and overall accuracy (.82) among thescreening instruments examined, and also hadgood specificity (.78). Studies also indicate thatthe TCUDS has very good test-retest reliabilityamong offenders (.89–.95; Knight, Simpson, &Morey, 2002; Peters et al., 2000). That is, theTCUDS is a valid, reliable instrument that is atleast as accurate as, and in some cases more accu-rate than, other commonly available measures.It is typically self-administered, requires fewerthan 10 minutes to complete, and is publicallyavailable at no charge.

Michigan Alcohol Screening Test (MAST).One of the first brief screening questionnairesever developed (Selzer, 1971), the MAST is a25-item scale that is used to identify individualswho may have an alcohol use disorder. A shorterform, the SMAST, includes 13 of the original25 items from the full version of the question-naire (Selzer, Vinokur, & van Rooijen, 1975).Because it measures only problematic alcoholuse, it is only relevant to the substance abusecriminogenic risk factor in the R-N-R model.Several meta-analyses have confirmed that it hasgood measurement properties, representing botha reliable and a valid test of problem alcoholuse (Shields, Howell, Potter, & Weiss, 2007;Storgaard, Nielsen, & Gluud, 1994; Teitelbaum &Mullen, 2000). However, like most screenersand assessments, the MAST does not performuniformly well across all samples relevant tocriminal justice planning. For example, Shields

54 M.L. Hiller et al.

and colleagues (2007) note that samples withhigher proportions of females yield lower levelsof reliability, while clinical samples yield higherreliability than do nonclinical ones. Teitelbaumand Mullen (2000) noted that the validity of thescores on the MAST is also influenced by gen-der and type of clinical problem. Specifically,in studies with proportionally higher numbersof women, validity estimates were higher thanin samples with fewer women. Also, samplesfrom alcoholism treatment programs yield scoreswith higher levels of validity than those frompsychiatric settings. Both studies conclude thatcharacteristics of the clients to whom, as well asthe setting in which, the MAST is administeredshould be considered prior to actual use of theinstrument. One concern in using the MAST withoffenders is that it tends to have greater sensitiv-ity than specificity (Peters et al., 2000), and thusis more likely than other screeners to misidentify,as alcohol abusers, offenders who do not havealcohol use disorders. Importantly, a review ofthe literature provided in these meta-analyses andother reviews included few references to stud-ies of the MAST in criminal justice settings (thevast majority of which were offenders arrested forDriving While Intoxicated). This suggests morestudy is needed to determine the extent to whichdifferent types of criminal justice samples andsettings affect the performance of the MAST. TheMAST is publicly available, self-administered,and hand-scored easily by staff. It typically takesfewer than 10 minutes to complete.

Drug Abuse Screening Test (DAST). TheDAST was first published in 1982 as a 28-itembrief screening instrument for identifying thosewho may have drug problems. It was devel-oped by rewording questions from the MAST torefer to drugs instead of alcohol (Skinner, 1982).Since it was introduced, two shorter versions havebeen published, including one that comprises asubset of 20 of the original items and an evenshorter one that includes 10 of the original items.Although it is copyrighted, releases for not-for-profit research and clinical applications are usu-ally granted (Yudko, Lozhkina, & Fouts, 2007).As shown in Table 4.2, it measures only one ofthe “Big 8” criminogenic needs. Research on it

with criminal justice samples has been limited,with only one study identified in an extensivereview of this instrument, in which it was usedwith women in jail or on probation (Saltstone,Halliwell, & Hayslip, 1994). Each version of theDAST was observed to have good measurementproperties and was shown to produce reliable andvalid measures of problem drug use (Yudko et al.,2007); but, as with the MAST, its reliability andvalidity estimates vary across different groups.With only limited research on its applicationto the criminal justice system, more empiricalwork is needed to determine whether the type ofoffender and/or the specific setting (e.g., prisonvs. community corrections) affects the reliabil-ity and validity of its scores. Like the MAST,the DAST is publicly available, self-administeredand easily hand-scored by staff. It typically takesfewer than 10 minutes to complete.

Screening and Assessment InstrumentsDeveloped During CJDATS

Seeking to call national attention to the need touse evidence-based practices in substance abusetreatment, NIDA, in 2002, funded nine researchcenters under a cooperative agreement known asCriminal Justice Drug Abuse Treatment Studies(CJDATS; Fletcher, 2003). CJDATS conductedmultisite research to improve drug treatment ser-vices for drug-involved offenders and developand test system-level drug treatment models, witha goal of building new evidence-based interven-tions for criminal justice populations. One par-ticular focus of CJDATS was the developmentof new screening and assessment instruments, toimprove substance abuse treatment for offend-ers with drug and alcohol programs. Severalinstruments were developed and tested duringthis project, showed promise, and were founddeserving of additional research (Simpson &Knight, 2007). All of these instruments, dis-cussed below, are in the public domain, are self-administered, and require no specialized trainingto use. Because they are so new, estimates of theamount of time to complete are not yet availablefor discussion.

4 Screening and Assessment: An Evidence-Based Process for the Management and Care of Adult. . . 55

Inmate Prerelease Assessment (IPASS). TheIPASS is a prerelease assessment designedand tested under CJDATS (Farabee, Knight,Garner, & Calhoun, 2007) with a focus on iden-tifying the risk and needs of prisoners in drugabuse treatment nearing parole. Its primary intentis to identify the types of additional treatmentneeded by the offender, upon release, as well aseach offender’s amenability for continuing treat-ment in the community. Specifically, it consists offour major parts including a general backgroundrisk index (GBRI), Texas Christian UniversityDrug Use Scale (TCUDS, described above),the Client Evaluation of Treatment (CET), andthe Counselor Evaluation of Client (CEC). TheGBRI, based on the Salient Factor Score, a well-validated 2G risk assessment (Hoffman, 1983),provides an actuarial measure for determining anoffender’s risk for recidivism through analysis oftheir preincarceration criminal behavior, includ-ing arrest and incarceration history, revocationhistory, and their age at first arrest. As alreadynoted, the TCUDS is a brief assessment based onthe Diagnostic and Statistical Manual (DSM) cri-teria for substance dependence. The CET is com-posed of several items that measure an offender’sperceptions of his/her in-prison treatment expe-rience, and the CEC is the primary counselor’sappraisal of the offender. Each of these fourdifferent assessments are scored as compositemeasures, with the combined sum of the GBRIand TCUDS (representing risk and need for addi-tional treatment) ranging from 0 to 18 and thecombined score of the CET and CES (represent-ing amenability to additional treatment) rangingfrom 0 to 18. A final index is created by sub-tracting the scores on the GBRI/TCUDS fromthe CET/CEC which can range from –18 to 18.Within the R-N-R framework, IPASS specificallyaddresses antisocial behavior and drug abuseas criminogenic risk factors and taps specificresponsivity, through an offender’s self-reportand the primary counselor’s report of his/heramenability to treatment. Analyses reported byFarabee and colleagues (2007) show the instru-ment has sound measurement properties, havinghigh levels of validity and reliability. AlthoughIPASS is promising, additional research is needed

to replicate these findings with other samples,to establish a larger body of research on thisinstrument and determine whether it has a broadevidence base supporting its use with offendersamples.

Co-occurring Disorders Screening Instru-ments (CODSIs). Sacks and colleagues (2007a)discuss the development of two brief screen-ing instruments for use in the criminal justicesystem. These instruments, the Co-occurringDisorders Screening Instrument – MentalDisorder (CODSI-MD) and the Co-occurringDisorders Screening Instrument – Severe MentalDisorder (CODSI-SMD), were created by tak-ing six items (CODSI-MD) and three items(CODSI-SMD) from three commonly usedmental health screening instruments that had thehighest correlation with diagnoses provided bythe Structured Clinical Interview for the DSM-IV(SCID-IV) and combining them with TCUDS(described above) (Sacks et al., 2007a). Eachmeasures one of the “Big 8” criminogenic needs,substance abuse, and also measures mentalillness, considered a noncriminogenic need in theR-N-R model, which can affect an individual’sresponsiveness to treatment. Psychometric analy-ses show that the CODSI-MD and CODSI-SMDhave good reliability and are valid for use withcriminal justice populations. In addition, whencompared with other standardized mental healthassessments, the CODSI-SMD showed thehighest overall accuracy rate and required theleast amount of time to administer among theinstruments studied (Duncan et al., 2008; Sackset al., 2007a, 2007b).

TCU Criminal Thinking Styles Scales (TCUCTS). Based on Glenn Walters’s (1995) work anddeveloped in collaboration with the Bureau ofPrisons with funding from the National Instituteon Corrections, the TCU CTS is a brief self-report assessment (36 items, estimated to takebetween 5 and 10 minutes to complete) thatassesses dynamic risk factors including antiso-cial attitudes and cognitions (Knight, Simpson, &Morey, 2002). It contains six scales, repre-senting entitlement, justification, personal irre-sponsibility, power orientation, cold heartedness,

56 M.L. Hiller et al.

and criminal rationalization (Knight, Garner,Simpson, Morey, & Flynn, 2006). As shown inTable 4.2, these six scales fit squarely withinthe antisocial personality and antisocial cogni-tions areas of the “Big 8” criminogenic risks andneeds. This instrument is designed to be admin-istered at repeated intervals during an offender’stenure in a treatment program to assess changesin these dynamic criminogenic factors. To date,however, there is no published peer-reviewedliterature on this specific intended application.However, psychometric data are available froma cross-sectional application of the TCU CTSto a large sample of offenders in 26 residentialprograms during CJDATS. Findings reported byKnight et al. (2006) shows the measure is reliable,but, to date, has not been examined for validity.Although the TCU CTS shows promise, addi-tional research is needed to determine whetherit also is a valid instrument, especially forthe monitoring of offender progress duringtreatment.

TCU Client Evaluation of Self and Treatment(TCU CEST). Compared to the TCU CTS, arelatively large amount of research has beenaccomplished with the TCU Client Evaluationof Self and Treatment. Earlier versions of thisinstrument (referred to at different times as theTCU Motivational Scales, the TCU Self-RatingForm, and the TCU Evaluation of Self andTreatment) were developed for application inoutpatient community-based drug treatment pro-grams, and numerous studies showed it was reli-able and valid in this setting (see Joe, Broome,Rowan-Szal, & Simpson, 2002; Simpson & Joe,1993; Simpson, Joe, Rowan-Szal, & Greener,1995). It was adapted and renamed the ResidentEvaluation of Self and Treatment (REST; Hiller,Knight, Leukefeld et al., 2002; Hiller, Knight,Rao, & Simpson, 2002; Hiller, Knight, Saum, &Simpson, 2006; Welsh, 2006; Welsh & McGrain,2008) for use with drug-involved offenders incorrectional settings. Subsequent to this, it wasrenamed the TCU Client Evaluation of Selfand Treatment (TCU-CEST) and was tested dur-ing CJDATS (Garner, Knight, Flynn, Morey, &Simpson, 2007; Saum et al., 2007; Staton-Tindall

et al., 2007). The TCU-CEST contains 130 self-administered statements on which the respon-dents rate the extent to which they agree ordisagree (1 = strongly disagree, 5 = stronglyagree) with each. A total of 15 scales in threemajor domains (i.e., treatment motivation, psy-chosocial functioning, and treatment engage-ment) are scored, and this questionnaire mayalso be combined with the TCU CTS, describedabove (see Garner et al., 2007). Subscales forthe treatment motivation domain include desirefor help, treatment readiness, treatment needs,and pressures for treatment. The psychosocialfunctioning domain includes anxiety, depression,self-esteem, decision making, hostility, and risktaking. Finally, the treatment engagement domainincludes treatment participation, treatment sat-isfaction, counseling rapport, peer support, andsocial support. When considered within the con-text of the R-N-R framework, the scales from theTCU CEST map to some of the “Big 8” crimino-genic needs. For example the hostility, risk tak-ing, and decision making scales measure antiso-cial cognitions and attitudes (see Table 4.2) whilepeer support reflects prosocial peer networks.The remaining scales examine noncriminogenicneeds (e.g., anxiety, depression) as well as spe-cific responsivity (e.g., desire for help, treatmentreadiness, treatment participation). The reliabil-ity and validity of this scale (as well as previousversions) are good, and it shows a great deal ofpromise for measuring changes in criminogenicrisk and needs as well as in other needs andresponsivity factors. The recommended use forthis instrument is as a monitoring tool adminis-tered repeatedly throughout an offender’s tenurein treatment to assess change in these areas andinform the revision of treatment plans (Garneret al., 2007).

Client Assessment Inventory (CAI). Like theTCU CEST, the CAI was first developed forapplication to community-based substance abusetreatment programs and was modified and testedwithin CJDATS for use with criminal justicesamples (Kressel, De Leon, Palij, & Rubin,2000; Sacks, McKendrick, & Kressel, 2007).Based on the theoretical work of De Leon(2000) on therapeutic communities (TC), the

4 Screening and Assessment: An Evidence-Based Process for the Management and Care of Adult. . . 57

CAI is a 103-item self-administered question-naire that includes 14 scales in four generaldomains (i.e., developmental, socialization, psy-chological, and program participation). Scalesfor the developmental domain include matu-rity, responsibility, and values. For the social-ization dimension, scales include drug/criminallifestyle, maintaining images, work attitude, andsocial skills. The psychological domain com-prises scales for cognitive skills, emotional skills,and self-esteem/self-efficacy. Finally, the pro-gram participation dimension includes philos-ophy/understands program rules, engagement,attachment/investment, and role model. Withinthe context of R-N-R, the CAI measures anti-social personality and cognitions, drug/criminallifestyle, and employment as criminogenic needs(see Table 4.2). Specific responsivity also is mea-sured by several scales (e.g., emotional skills,self-esteem/self-efficacy). Scores on it have beenshown to be reliable among offenders in correc-tional substance abuse treatment (TC and non-TC; Sacks et al., 2007), but to date no validitydata have been published. Like the TCU CEST,its intended use is to monitor treatment progress,which requires repeated administration through-out an offender’s time in treatment.

ConclusionsWide variation exists in screening and assess-ment practices implemented within the crim-inal justice system, particularly for thoseoffenders with substance use and other relateddisorders (Taxman, Cropsey et al., 2007).Standardized instruments are not used in manycriminal justice settings, and when used areoften deficient with respect to reliability andvalidity. As indicated in this review, most sub-stance abuse screening and assessment instru-ments have not been extensively validatedfor use with offenders. Thus, it is unclearwhether the observed reliability and valid-ity of these instruments are generalizable tocriminal justice settings. Many settings do notemploy an integrated set of screening andassessment instruments to address the widerange of psychosocial needs among offenders(Belenko, 2006). Also of concern is that most

settings do not routinely compile informationregarding offenders’ risk levels.

Even when the level of offender risk andneed is assessed, there are challenges in devel-oping evidence-based approaches to translatethis information into triage/placement deci-sions that affect the level of rehabilitative ser-vices and supervision provided. This referraland service delivery gap is particularly prob-lematic during reentry to probation or parole.Supervision priorities favor security and mon-itoring over rehabilitation, community super-vision officers are not trained to make servicereferrals, and supervision requirements makeit difficult for offenders to access services(Belenko, 2006; Marlowe, 2003; Taxman,Young, & Byrne, 2004).

The Risk-Needs-Responsivity (R-N-R)model provides a useful framework to addressthese deficiencies in offender screening andassessment, and has proven to be effectivein the United States and internationally inguiding assessment and treatment amongoffenders who have substance abuse and otherpsychosocial problems (Bonta & Andrews,2007; Taxman & Marlowe, 2006). At the coreof this model, the risk principle asserts thatcriminal behavior can be reliably predictedthrough assessment of key static and dynamicvariables and that treatment interventionsshould focus on the dynamic variablespresented by those assessed as higher-riskoffenders. The R-N-R model identifies“criminogenic” needs empirically linked torecidivism and its psychosocial antecedents(e.g., antisocial beliefs, substance use dis-orders, self-control/management, criminalpeers) as the cornerstone for both assessmentand subsequent treatment interventions. Thismodel is ideally suited to guide the devel-opment and implementation of screening,assessment, and triage/service matchingwithin the criminal justice system, and also tohelp organize offender treatment, supervision,and community reentry (Dowden & Andrews,2004).

At least five evidence-based screening andassessment principles flow from the R-N-R

58 M.L. Hiller et al.

model. First, assuming reduced recidivism isthe most important outcome, each offender’srisk should be assessed. Second, more inten-sive services should be provided to offenderswith the highest assessed risk. Third, screen-ing and assessment instruments should bedesigned to calibrate offender risks and needs.Screens should lead to more extensive assess-ment and not be used as primary means formaking treatment referral decisions. Fourth,assessments would optimally address multi-ple areas of criminogenic risk and need (e.g.,criminal history, substance abuse, criminalpeers, maladaptive belief systems). Finally,assessment based on the R-N-R model is anongoing process and results should be used ateach sequential point in the criminal justicesystem to inform institutional and commu-nity placement, involvement in rehabilitativeservices, and supervision approaches.

There are five decades of research indicat-ing that objective, actuarial assessment is a keycomponent of any human services interven-tion (Grove & Meehl, 1996). Moreover, thereis evidence that offender programs that userisk and needs assessment information pro-duce greater reductions in recidivism than pro-grams that do not (Lowenkamp et al., 2006).Thus, increased attention to the psychometricadequacy of commonly used substance abuseassessment instruments is critical for bothresearchers and criminal justice practitioners,and should be a priority agenda.

A number of screening and assessmentinstruments have been developed and arereviewed here, that can be integrated to exam-ine the level of offenders’ risk and needs. Foroffender populations, these include evidence-based screens for substance abuse (TCUDS)and co-occurring disorders (CODSI), sub-stance abuse assessment instruments (ASI),prerelease instruments (IPASS), and special-ized instruments to evaluate offenders’ risklevel (LSI-R). As indicated previously, addi-tional research will help determine the psy-chometric properties of instruments designedto evaluate offender risk and needs, and todetermine how the results of screening and

assessment can best be integrated to formulatedecisions related to community placement,treatment, and supervision. Improvements inassessment instruments and the use of theseassessments for appropriate supervision andtreatment planning for offenders are needed toenhance both public safety and public health.

References

Alterman, A. I., Bovasso, G. B., Cacciola, J. S., &McDermott, P. A. (1994). A comparison of the pre-dictive validity of four sets of baseline ASI sum-mary indices. Psychology of Addictive Behavior, 15,159–162.

Alterman, A. I., Mulvaney, F. D., Cacciola, J. S., Cnaan,A., McDermott, P. A., & Brown, L. S. (2001). Thevalidity of the interviewer severity ratings in groupsof ASI interviewers with varying training. Addiction,96, 1297–1305.

American Psychiatric Association. (1994). Diagnosticand statistical manual of mental disorders (4Th Ed.).Washington, DC: Author.

Andrews, D. A., & Bonta, J. (1995). The LSI-R: Thelevel of service inventory-revised. Toronto, ON: Multi-Health Systems, Inc.

Andrews, D. A., & Bonta, J. (2006). The psychologyof criminal conduct (4th ed.). Newark, NJ: LexisNexis/Matthew Bender.

Andrews, D. A., Bonta, J., & Hoge, R. D. (1990).Classification for effective rehabilitation: Redis-covering psychology. Criminal Justice and Behavior,17, 19–52.

Andrews, D. A., Bonta, J., & Wormith, S. J. (2004). Thelevel of service/Case management inventory (LS/CMI).Toronto, ON: Multi-Health Systems.

Andrews, D. A., Bonta, J., & Wormith, S. J. (2006).The recent past and near future of risk and/or needassessment. Crime and Delinquency, 52, 7–27.

Andrews, D. A., & Dowden, C. (2007). The risk-need-responsivity model of assessment and human ser-vice in prevention and corrections: Crime-preventionjurisprudence. Canadian Journal of Criminology andCriminal Justice, 49(4), 439–464.

Andrews, D. A., Kiessling, J. J., Mickus, S., & Robinson,D. (1986). The construct validity of interview basedrisk assessment in corrections. Canadian Journal ofBehavioral Science, 18, 460–470.

Baird, C. (2009). A question of evidence: A critique of riskassessment models used in the justice system. Oakland,CA: National Council on Crime and Delinquency.

Baird, C., Heinz, R., & Bemus, B. (1979). The Wisconsincase classification/staff deployment project. Madison,WI: Wisconsin Division of Corrections.

Baird, C., Heinz, R., & Bemus, B. (1981). The Wisconsincase classification staff deployment project: Two year

4 Screening and Assessment: An Evidence-Based Process for the Management and Care of Adult. . . 59

follow-up report. Madison, WI: Wisconsin Division ofCorrections.

Belenko, S. (2006). Assessing released inmates forsubstance-abuse related service needs. Crime andDelinquency, 52, 94–113.

Bonta, J. (2002). Offender risk assessment: Guidelines forselection and use. Criminal Justice and Behavior, 29,355–379.

Bonta, J., & Andrews, D. A. (2007). Risk-need-responsivity model for offender assessment and reha-bilitation. Ottawa, ON: Public Safety Canada.

Bonta, J., & Motiuk, L. L. (1987). The diversion ofincarcerated offenders to correctional halfway houses.Journal of Research in Crime and Delinquency, 24,302–323.

Bonta, J., & Motiuk, L. L. (1990). Classification to cor-rectional halfway houses: A quasi-experimental evalu-ation. Criminology, 28, 497–506.

Brennan, T., Dieterich, W., & Ehret, B. (2009). Evaluatingthe predictive validity of the COMPAS risk and needsassessment system. Criminal Justice and Behavior, 36,21–40.

Broome, K. M., Flynn, P. M., Knight, D. K., & Simpson,D. D. (2007). Program structure, staff perceptions, andclient engagement in treatment. Journal of SubstanceAbuse Treatment, 33, 149–158.

Broome, K. M., Knight, D. K., Knight, K., Hiller, M. L., &Simpson, D. D. (1997). Peer, family, and motivationalinfluences on drug treatment process and recidivismfor probationers. Journal of Clinical Psychology, 53,387–397.

Clawson, E., Bogue, B., & Joplin, L. (2005).Implementing evidence-based practices in cor-rections. Washington, DC: National Institute onCorrections.

Clear, T. R., & Gallagher, K. W. (1983). Screeningdevices in probation and parole: Management prob-lems. Evaluation Review, 7, 217–234.

Coulson, G., Ilacqua, G., Nutbrown, V., Giulekas, D., &Cudjoe, F. (1996). Predictive validity of the LSI forincarcerated female offenders. Criminal Justice andBehavior, 23, 427–439.

De Leon, G. (2000). The therapeutic community: Theory,model & method. New York: Springer.

Dowden, C., & Andrews, D. A. (2000). Effectivecorrectional treatment and violent re-offending: Ameta-analysis. Canadian Journal of Criminology, 42,449–467.

Dowden, C., & Andrews, D. A. (2004). The importanceof staff practices in delivering effective correctionaltreatment: A meta-analysis of core correctional prac-tices. International Journal of Offender Therapy andComparative Criminology, 48, 203–214.

Duncan, A., Sacks, S., Melnick, G., Cleland, C. M.,Pearson, F. S., & Coen, C. (2008). Performanceof the CJDATS Co-Occurring Disorders ScreeningInstruments (CODSIs) among minority offenders.Behavioral Sciences and the Law, 26, 351–368.

Eisenberg, M., Bryl, J., & Fabelo, T. (2009). Validationof the Wisconsin department of corrections risk

assessment instrument. New York: Council of StateGovernments Justice Center.

Farabee, D., Knight, K., Garner, B. R., & Calhoun, S.(2007). The inmate prerelease assessment for reen-try planning. Criminal Justice and Behavior, 34,1188–1197.

Fass, T. L., Heilbrun, K., DeMatteo, D., & Fretz, R.(2008). The LSI-R and the COMPAS: Validationdata on two risk-needs tools. Criminal Justice andBehavior, 35, 1095–1108.

Feldstein, S. W., & Miller, W. R. (2007). Does subtlescreening for substance abuse work? A review of theSubstance Abuse Subtle Screening Inventory (SASSI).Addiction, 102, 41–50.

Fletcher, B. W. (2003). The national criminal justicedrug abuse treatment studies (CJ-DATS). OffenderSubstance Abuse Report, 3, 1–5.

Garner, B. R., Knight, K., Flynn, P. M., Morey, J. T., &Simpson, D. D. (2007). Measuring offender attributesand engagement in treatment using the client eval-uation of self and treatment. Criminal Justice andBehavior, 34, 1113–1130.

Gendreau, P., Goggin, C., & Smith, P. (2002). Is the PCL-R really the “unparalleled” measure of offender risk?A lesson in knowledge cumulation. Criminal Justiceand Behavior, 29, 397–426.

Grove, W. M., & Meehl, P. E. (1996). Comparative effi-ciency of informal (subjective, impressionistic) andformal (mechanical, algorithmic) prediction proce-dures: The clinical-statistical controversy. Psychology,Public Policy and Law, 2, 293–323.

Hanlon, T. E., O’Grady, K. E., & Bateman, R. W. (2000).Using the Addiction Severity Index to predict treat-ment outcome among substance abusing parolees.Journal of Offender Rehabilitation, 31, 67–79.

Hiller, M. L., Knight, K., Leukefeld, C., & Simpson,D. D. (2002). Motivation as a predictor of treatmentengagement in mandated residential substance abusetreatment. Criminal Justice and Behavior, 29, 56–75.

Hiller, M. L., Knight, K., Rao, S. R., & Simpson, D.D. (2002). Assessing and evaluating mandated correc-tional substance abuse treatment. In C. G. Leukefeld,F. Tims, & D. Farabee (Eds.), Treatment of drugoffenders: Policies and issues (pp. 41–56). New York:Springer.

Hiller, M. L., Knight, K., Saum, C. A., & Simpson,D. D. (2006). Social functioning, treatment dropout,and recidivism of probationers mandated to a mod-ified therapeutic community. Criminal Justice andBehavior, 33, 738–759.

Hiller, M. L., & Narevic, E. (2005). Evaluating brief sub-stance abuse screening instruments for drug-involvedoffenders. Offender Substance Abuse Report, 5(3),35–37, 48.

Hoffman, P. B. (1983). Screening for risk: A revisedsalient factor score. Journal of Criminal Justice, 11,539–547.

Hubbard, D. J., Travis, L. F., & Latessa, E. J. (2001). Caseclassification in community corrections: A national

60 M.L. Hiller et al.

survey of the state of the art. Washington, DC:National Institute of Justice.

Joe, G. W., Broome, K. M., Rowan-Szal, G. A., &Simpson, D. D. (2002). Measuring patient attributesand engagement in treatment. Journal of SubstanceAbuse Treatment, 22, 183–196.

Kelly, C. E., & Welsh, W. N. (2008). The predictive valid-ity of the level of service inventory-revised for drug-involved offenders. Criminal Justice and Behavior, 35,819–831.

Knight, K., Garner, B., Simpson, D. D., Morey, J. T., &Flynn, P. M. (2006). An assessment of criminal think-ing. Crime and Delinquency, 52, 159–177.

Knight, K., Simpson, D. D., & Hiller, M. L. (1999). Three-year reincarceration outcomes for in-prison therapeu-tic community treatment in Texas. Prison Journal,79(3), 337–351.

Knight, K., Simpson, D. D., & Hiller, M. L. (2002).Screening and referral for substance-abuse treatmentin the criminal justice system. In C. G. Leukefeld,F. Tims, & D. Farabee (Eds.), Treatment of drugoffenders: Policies and issues (pp. 259–272, 373–376).New York: Springer.

Knight, K., Simpson, D. D., & Morey, J. T. (2002). TCU-NIC cooperative agreement: Final report. Fort Worth,TX: Texas Christian University, Institute of BehavioralResearch.

Kressel, D., De Leon, G., Palij, M., & Rubin, G.(2000). Measuring client clinical progress in thera-peutic community treatment: The therapeutic com-munity client assessment inventory, client assessmentsummary, and staff assessment. Journal of SubstanceAbuse Treatment, 19, 267–272.

Latessa, E. J., & Holsinger, A. (1998). The importanceof evaluating correctional programs: Assessing out-come and quality. Corrections Management Quarterly,2, 22–29.

Latessa, E. J., & Lowenkamp, C. (2006). What works inreducing recidivism? University of St. Thomas LawJournal, 3, 521–535.

Leonhard, C., Mulvey, K., Gastfriend, D. R., & Shwartz,M. (2000). The Addiction Severity Index: A fieldstudy of internal consistency and reliability. Journal ofSubstance Abuse Treatment, 18, 129–135.

Lowenkamp, C., Hodsinger, S., & Latessa, E. (2001).Risk/need assessment, offender classification, and therole of child abuse. Criminal Justice and Behavior, 28,543–563.

Lowenkamp, C. T., & Latessa, E. J. (2002). Evaluationof Ohio’s community based correctional facilities andhalfway house programs. Cincinnati, OH: Universityof Cincinnati.

Lowenkamp, C. T., Latessa, E. J., & Smith, P. (2006).Does correctional program quality really matter? Theimpact of adhering to the principles of effective inter-vention. Criminology & Public Policy, 5, 575–594.

Loza, W., & Simourd, D. J. (1994). Psychometric eval-uation of the Level of Supervision Inventory (LSI)among male Canadian federal offenders. CriminalJustice and Behavior, 21, 468–480.

Mäkelä, K. (2004). Studies of the reliability and validity ofthe Addiction Severity Index. Addiction, 99, 398–410.

Marlowe, D. B. (2003, August). Integrating substanceabuse treatment and criminal justice supervision.Science and practice perspectives. Rockville, MD:National Institute on Drug Abuse.

McLellan, A. T., Kushner, H., Metzger, D., Peters, R.,Smith, I., Grissom, G., et al. (1992). The fifth editionof the Addiction Severity Index. Journal of SubstanceAbuse Treatment, 9, 199–213.

McLellan, A. T., Luborsky, L., Cacciola, J., & Griffith,J. E. (1985). New data from the Addiction SeverityIndex: Reliability and validity in three centers. Journalof Nervous and Mental Disease, 173, 412–423.

McLellan, A. T., Luborsky, L., Woody, G. E., &O’Brien, C. P. (1980). An improved diagnostic eval-uation instrument for substance abuse patient, theAddiction Severity Index. Journal of Nervous andMental Disease, 168, 26–33.

Miller, C. S., Woodson, J., Howell, R. T., & Shields, A.L. (2009). Assessing the reliability of scores producedby the substance abuse subtle screening inventory.Substance Use and Misuse, 44, 1090–1100.

Miller, F. G. (1985). The substance abuse subtle screeninginventory manual. Bloomington, IN: Glenn A. Miller.

Miller, F. G., & Lazowski, L. E. (1999). The adult SASSI-3manual. Springville, IN: SASSI Institute.

National Institute on Drug Abuse. (2006). Principles ofdrug abuse treatment for criminal justice populations.Rockville, MD: Author.

Ogloff, J. R. P., & Davis, M. R. (2004). Advances inoffender assessment and rehabilitation: Contributionsof the risk-needs-responsivity approach. Psychology,Crime, and Law, 10, 229–242.

Peters, R. H., Bartoi, M. G., & Sherman, P. B. (2008).Screening and assessment of co-occurring disorders inthe justice system. Delmar, NY: The National GAINSCenter.

Peters, R. H., Greenbaum, P. E., & Edens, J. F. (1998).Prevalence of DSM-IV substance abuse and depen-dence disorders among prison inmates. AmericanJournal of Drug and Alcohol Abuse, 24, 573–587.

Peters, R. H., Greenbaum, P. E., Steinberg, M. L.,Carter, C. R., Ortiz, M. M., Fry, B., et al. (2000).Effectiveness of screening instruments in detectingsubstance abuse disorders among prisoners. Journal ofSubstance Abuse Treatment, 18, 349–358.

Peters, R. H., & Wexler, H. K. (Eds.). (2005). Substanceabuse treatment for adults in the criminal justice sys-tem. Treatment improvement protocol (TIP) series 44.Rockville, MD: Substance Abuse and Mental HealthServices Administration, Center for Substance AbuseTreatment.

Rikoon, S. H., Cacciola, J. S., Carise, D., Alterman,A., & McLellan, A. T. (2006). Predicting DSM-IVdependence diagnoses with the Addiction SeverityIndex composite scores. Journal of Substance AbuseTreatment, 31, 17–24.

4 Screening and Assessment: An Evidence-Based Process for the Management and Care of Adult. . . 61

Roberts, E. A., Contois, M. W., Willis, J. C., Worthington,M. R., & Knight, K. (2007). Assessing offender needsand performance for planning and monitoring criminaljustice drug treatment. Criminal Justice and Behavior,34, 1179–1187.

Sacks, J. Y., McKendrick, K., & Kressel, D. (2007).Measuring offender progress in treatment using theClient Assessment Inventory. Criminal Justice andBehavior, 34, 1131–1142.

Sacks, S., Melnick, G., Coen, C., Banks, S., Friedmann, P.D., Grella, C., et al. (2007a). CJDATS Co-occurringdisorders screening instrument for mental disorders(CODSI-MD): A pilot study. The Prison Journal, 87,86–110.

Sacks, S., Melnick, G., Coen, C., Banks, S., Friedmann, P.D., Grella, C., et al. (2007b). CJDATS Co-occurringdisorders screening instrument for mental disorders:A validation study. Criminal Justice and Behavior, 34,1198–1215.

Saltstone, R., Halliwell, S., & Hayslip, M. A. (1994).A multivariate evaluation of the Michigan AlcoholScreening Test and the Drug Abuse Screening Test ina female offender population. Addictive Behaviors, 19,455–462.

Saum, C. A., O’Connell, D. J., Martin, S. S., Hiller, M. L.,Bacon, G. A., & Simpson, D. D. (2007). Tempest in aTC: Changing treatment providers for in-prison ther-apeutic communities. Criminal Justice and Behavior,37, 1168–1178.

Selzer, M. L. (1971). The Michigan Alcoholism ScreeningTest: The quest for a new diagnostic instrument.American Journal of Psychiatry, 127, 1653–1658.

Selzer, M. L., Vinokur, A., & van Rooijen, L. (1975).A self-administered Short Michigan AlcoholismScreening Test (SMAST). Journal of Studies onAlcohol, 36, 117–126.

Shields, A. L., Howell, R. T., Potter, J. S., & Weiss, R. D.(2007). The Michigan Alcoholism Screening Test andits shortened form: A meta-analytic inquiry into scorereliability. Substance Use and Misuse, 42, 1783–1800.

Shields, I. W., & Simourd, D. J. (1991). Predicting preda-tory behavior in a population of young offenders.Criminal Justice and Behavior, 18, 180–194.

Simourd, D. (2004). Use of dynamic risk/need assessmentinstruments among long-term incarcerated offenders.Criminal Justice and Behavior, 31, 306–323.

Simourd, D. (2006, May). Validation of risk/needs assess-ments in the Pennsylvania Department of Corrections.Interim report. Collingswood, NJ: Volunteers ofAmerica-Delaware Valley.

Simourd, D. J., & Malcolm, P. B. (1998). Reliabilityand validity of the Level of Service Inventory-Revisedamong federally incarcerated offenders. Journal ofInterpersonal Violence, 13, 261–274.

Simpson, D. D., & Joe, G. W. (1993). Motivation as apredictor of early dropout from drug abuse treatment.Psychotherapy, 30, 357–368.

Simpson, D. D., Joe, G. W., & Broome, K. M. (2002).A national 5-year follow-up of treatment outcomes for

cocaine dependence. Archives of General Psychiatry,59, 538–544.

Simpson, D. D., Joe, G. W., Rowan-Szal, G. A., &Greener, J. M. (1995). Client engagement andchange during substance abuse treatment. Journal ofSubstance Abuse Treatment, 7, 117–134.

Simpson, D. D., & Knight, K. (2007). Offender needs andfunctioning assessments from a national cooperativeresearch program. Criminal Justice and Behavior, 34,1105–1112.

Skinner, H. A. (1982). The Drug Abuse Screening Test.Addictive Behaviors, 7, 363–371.

Staton-Tindall, M., Garner, B. R., Morey, J. T., Leukefeld,C., Krietemeyer, J., Saum, C. A., et al. (2007). Genderdifferences in treatment engagement among a sam-ple of incarcerated substance abusers. Criminal Justiceand Behavior, 34, 1143–1156.

Storgaard, H., Nielsen, S. D., & Gluud, C. (1994).The validity of the Michigan Alcohol Screening Test(MAST). Alcohol and Alcoholism, 29, 493–502.

Swartz, J. A. (1998). Adapting and using the SubstanceAbuse Subtle Screening Inventory-3 with criminal jus-tice offenders: Preliminary results. Criminal Justiceand Behavior, 25, 344–365.

Taxman, F. S., Cropsey, K. L., Young, D. W., & Wexler, H.(2007). Screening, assessment, and referral practicesin adult correctional settings: A national perspective.Criminal Justice and Behavior, 34, 1216–1234.

Taxman, F. S., & Marlowe, D. B. (2006). Risk, needs,responsivity: In action or inaction? Crime andDelinquency, 52(1), 3–6.

Taxman, F. S., Perdoni, M., Young, D., Belenko, S., &Hiller, M. (in press). Twenty years of drug treatmentcourts: The current state of drug courts. In D. K.Shaffer (Ed.). Drug Courts and the Criminal JusticeSystem. Boulder, CO: Lynne Rienner Publishers.

Taxman, F. S., Thanner, M., & Weisburd, D. (2006). Risk,need, and responsivity (RNR): It all depends. Crimeand Delinquency, 52, 28–51.

Taxman, F., Young, D., & Byrne, J. (2004). Transformingoffender reentry into public safety: Lessons fromOJP’s Reentry Partnership Initiative. Justice Policyand Research, 5, 101–128.

Taxman, F. S., Young, D., Wiersema, B., Rhodes,A., & Mitchell, S. (2007). The National criminaljustice treatment practices survey: Methods and pro-cedures. Journal of Substance Abuse Treatment, 32,225–238.

Teitelbaum, L., & Mullen, B. (2000). The validityof MAST in psychiatric settings: A Meta-analyticintegration. Journal of Studies on Alcohol, 61,254–261.

Thanner, M., & Taxman, F. S. (2003). Responsivity:The value of providing intensive services to high-riskoffenders. Journal of Substance Abuse Treatment, 24,137–147.

Walters, G. D. (1995). The psychological inventory ofcriminal thinking styles: Part I. Reliability and pre-liminary validity. Criminal Justice and Behavior, 22,307–325.

62 M.L. Hiller et al.

Welsh, W. N. (2006). Evaluation of drug treatmentprograms at the State Correctional Institution(SCI) at Chester: A partnership between thePennsylvania Department of Corrections, GaudenziaInc., and Temple University. (Final Report to theNational Institute of Justice 2002-RT-BX-1002).Retrieved October 9, 2009, from the PADOCWeb site http://www.cor.state.pa.us/stats/lib/stats/ChesterReporttoNIJ-Final-November2006.pdf

Welsh, W. N., & McGrain, P. N. (2008). Predictors of ther-apeutic engagement in prison-based drug treatment.Drug and Alcohol Dependence, 96, 271–280.

Yudko, E., Lozhkina, O., & Fouts, A. (2007). A com-prehensive review of the psychometric properties ofthe Drug Abuse Screening Test. Journal of SubstanceAbuse Treatment, 32, 189–198.