Albert R. Roberts, Ph.D.
Professor of Criminal Justice, Rutgers University
Kimberly Bender, M.S.W.
Doctoral Candidate, University of Texas at Austin
Juvenile Assessment Centers
Components of Risk and Need Assessment
Tools for Mental Health and Associated Risk Assessment
Conclusion and Suggestions for Further Research
IN THIS AGE OF accountability and performance-based measures, criminal justice professionals are being increasingly required by state and federal agencies to demonstrate the reliability and validity of their assessment instruments including brief symptom inventories, diagnostic tools, and violence risk assessment measures. Risk assessment tools assist institutional classification boards as well as parole boards to: 1) determine an initial security rating and placement into a particular facility and program(s); 2) develop a rehabilitation treatment plan; 3) assess eligibility for early release; and 4) determine the type of supervision needed while on parole. This article first describes how the juvenile justice system assesses youths’ risks and needs through Juvenile Assessment Centers, then explores common components of assessment in the juvenile justice system, and concludes with an examination of the most commonly used risk and mental health assessment tools and the evidence that supports their use.
All experienced probation officers, juvenile counselors, and forensic clinicians should have skills in risk assessment. Clinical assessment knowledge and skills provide the foundation for clinical judgments, applied research, and evidence-based practice. Within the juvenile justice system, prediction can be operationally defined as an assessment of future lawbreaking for the juvenile offenders who are officially processed through the system.
There are two primary types of prediction: clinical and actuarial. Clinical predictions are made by trained juvenile justice and forensic specialists after they have examined an individual’s criminal and psychosocial history, and the results from psychosocial scales and inventories. Actuarial prediction methods are based on known properties, parameters and statistical formulas applied to identical sets of data (e.g. demographic data, criminal history). Because of the two authors’ backgrounds in forensic mental health and social work, we focus on clinical judgments and the most commonly used assessment scales for measuring mental health status, psychosocial functioning, and future criminality.
There is no single scale or assessment tool that can predict future mental health status or criminality with 100 percent certainty. Behavior, abilities, peer influences, family factors and deviant behavior patterns are not static. They often change with age and different experiences. Empirical evidence from classic longitudinal studies indicate that violent juveniles are strongly influenced by male siblings of similar ages, delinquent gangs, and small groups of delinquent friends (Farrington & Loeber, 2000; Farrington & West, 1990). Therefore, it is critically important to use multiple assessment tools with clients at different points in the juvenile justice process.
Clinical prediction is based on per captions and judgments in which the juvenile justice professional and/or mental health clinician uses different data sources, such as clinical diagnosis, ratings and scores on psychosocial risk assessment scales, interviews, psychotherapy records, and criminal history data to make judgments about the offender’s placement in institutional or community-based treatment programs, progress, and discharge from probation.
The early roots of prediction in juvenile justice can be traced to the establishment of the first juvenile court mental health clinic in Chicago in 1909 (Roberts, 2004), and the rapid growth and development of over 600 child guidance clinics by the late 1950s connected to juvenile courts throughout the United States (Roberts, 2004). In these clinics forensic psychiatrists and social workers collaborated on behalf of troubled juveniles.
By the late 1990s, most state juvenile correctional agencies included formal and informal dangerousness and risk of further violence and re-offending in their intake classification and assessment centers. The goal of risk assessments is twofold: 1) to predict the probability that a juvenile offender will re-offend; 2) to predict which youths are at high risk of exhibiting violence in the institution or residential treatment facility, or upon release to parole supervision in the community. In general, classification decisions are made based on forecasts regarding which treatment/rehabilitation program is likely to be effective in changing the behavior patterns of specific types of juveniles, generally viewed as either property-related offenders or violent offenders adjudicated for offenses against persons.
One of the most overlooked areas of juvenile justice is the assessment and treatment of juvenile offenders with mental health disorders, especially co-morbid psychiatric disorders. Research indicates that at least two-thirds of juvenile detainees have one or more mental health disorders in addition to their juvenile offenses. Incarcerated juveniles suffering from impulsiveness, hopelessness and depression are at an increased risk of suicide ideation, suicide attempts, and death (Rapp-Palicchi & Roberts, 2004). The death rate from suicide is 4.6 times higher in juvenile detention centers than in the general population (Sheras, 2000). Therefore, it is imperative that experienced mental health professionals be hired by juvenile justice agencies so that they can conduct extensive assessments at the pre-adjudicatory, incarceration, and community release stages (Rapp-Palicchi & Roberts, 2004). At the present time, most large juvenile probation departments do have a few probation officers with expertise in forensic mental health assessment and treatment. However, the time is now ripe for the National Juvenile Detention Association (NJDA), as well as the American Correctional Association (ACA), and state and county correctional administrators to follow the lead of the American Probation and Parole Association (APPA) in giving priority to setting standards, and encouraging their members to hire and train staff in all aspects of juvenile assessment and treatment.
Juvenile Assessment Centers
A promising advancement of juvenile assessment is the innovative development of centralized, single point of entry, intake juvenile assessment centers. These assessment centers are based on a general model for bringing together a variety of community agencies to one centralized location in which all justice system-involved youth can receive thorough assessment. Juvenile justice, law enforcement, school truancy, diversion programs, and other human service agencies are centrally located, allowing for efficient and comprehensive assessment of youths’ risks and service needs (Dembo, Schmeidler, & Walters (2004).
Key elements of juvenile assessment centers (JACs) include 1) A single, 24-hour, centralized point of contact for all youth in contact or at risk of contact with the juvenile justice system, 2) Screenings and comprehensive assessments of youths’ circumstances and service needs, 3) Management Information Systems that centralize information to avoid repetition and assure appropriate treatment, and 4) Case management services that integrate information in order to recommend appropriate referrals and follow up on youth after they are referred (Dembo, Schmeidler, & Walters, 2004).
Juvenile assessment centers got their start in the early 1990s in Florida and quickly gained the attention of the Florida legislature, which was struggling with prison overcrowding. With a growing budget due to special appropriations, JACs quickly spread to several counties across Florida and were eventually established in other states, including Colorado and Kansas. Investing further in assessment centers with an initiative in 1996, the OJJDP allocated funds to two assessment centers, in Denver, CO and Lee County, FL, designated as planning sites to develop more assessment centers. Additional funds supported improving services at two designated enhancement sites in Jefferson County, CO and Orlando, FL (Dembo, Schmeidler, & Walters, 2004).
JACs vary considerably by location, due in large part to access resources and the unique needs of the communities they are serving. For example, many Florida JACs work closely with nearby juvenile addiction receiving facilities to provide detoxification, assessment, and stabilization for youth with substance abuse problems. JACs differ in the range of services they provide, from those with only juvenile justice agencies to those such as the Hillsborough County, Florida JAC that provides an array of services, including booking, supervision, detention center screening, diversion, and truancy programming at one site. JACS located in urban settings tend to have longer hours, process many youth, and thus conduct more thorough assessments off-site (Dembo, Schmeidler, & Walters, 2004).
Despite these differences, JACs share common benefits to the juvenile justice system. They provide a centralized site for legally required mandates to be carried out more efficiently, saving time locating youth, completing multiple screenings, and providing information to courts for decision making. Integrating information in one information system allows for better-informed decisions regarding need for services and necessary level of supervision. Accessing all system-involved youth, JACs create a prime opportunity for prevention and early intervention. Finally, on a macro-level, information from JACs informs the community of broader juvenile justice trends and needs for new services (Dembo, Schmeidler, & Walters, 2004).
Dembo et al. (2004) note an ongoing struggle for funding experienced by many JACs. Consistent funding at the federal and state levels is needed in order to provide decent salaries to well-trained staff, thereby reducing staff turn-over and improving quality of service. Additional funds would also allow JACs to maintain their original goals of comprehensively responding to youths’ multifaceted needs, preventing JACs from skimming services and becoming mere processing centers.
With necessary funding and support, the future utility of JACs is broad and influential. JACs have the potential to play a major role in developing empirical knowledge in the future. With large sample sizes, JACs’ information systems could easily gather data on youths’ characteristics, service needs, and outcomes in different treatment programs, providing juvenile justice research with difficult to obtain information. This information can then be used to inform program development and service provisions to juvenile offenders.
JACs can also provide much-needed solutions to assessment, referral, and service delivery in the future. By integrating information among many agencies, JACs can help to identify youth who slip through the system by failing to follow through on treatment recommendations. Furthermore, providing objective measures of substance use through urinalysis screening is an invaluable service offered through JACs and has implications for validating youths’ self reports of substance use and subsequent appropriate treatment placements. Finally, JACs ensure investment in prevention efforts, keeping youth from further developing delinquency careers; these prevention efforts inversely relate to the number of youth requiring long-term incarceration that is expensive and fairly ineffective (Dembo, Schmeidler, & Walters, 2004).
Components of Risk and Need Assessment
The central goals of youth assessment in juvenile justice are: 1) the safety of the community by preventing re-offending; and 2) youth rehabilitation and clinical treatment. In other words, mental health assessments seek to identify both risk and treatment needs. Assessments must be comprehensive and cover several domains. Comprehensiveness includes assessing a youth’s offense history, family/environmental factors, education/employment history, peer relationships, and psychosocial functioning.
Assessing psychosocial functioning is particularly important as the juvenile offender population has elevated rates of mental health and substance use disorders (Teplin, Abram, McClelland, Duclan, & Mericle, 2003). Furthermore, these psychosocial factors (i.e. personality characteristics, behaviors, affect, attitudes, beliefs and interpersonal constructs) predict youths’ infractions while incarcerated and their behaviors once they are released into the community (Cauffman, 2004; Hathaway & Moncachesi, 2003).
Mental health status is often under-assessed and consequently under-treated in the juvenile justice system. This is because of a lack of resources and trained staff, as well as a punishment mentality. Teplin et al. (2005) report that only 15.4 percent of detained adolescents who needed mental health treatment received treatment in the detention center; it is estimated that as many as 13,000 detained youth with major mental health disorders go untreated (Teplin et al., 2005). Effective mental health assessment and treatment are critical for achieving effective juvenile justice.
In 2002 the Consensus Conference convened, composed of more than 20 researchers with expertise in mental health assessment and juvenile justice, with the aim of developing recommendations for mental health assessment in the juvenile justice system. The Consensus Conference brought together nationally recognized experts in the areas of mental health, juvenile justice, and child welfare service systems. It was guided by data from a national survey of current mental health assessment practices conducted by the Center for the Promotion of Mental Health Assessment in Juvenile Justice. Directed by Gail Wasserman at Columbia University’s Department of Child Psychiatry, the Center’s national survey provided information on the current practices and needs of juvenile justice systems across the nation. From these findings, the Consensus Conference was then able to create recommendations for standardizing mental health assessment practices on a national level. The Consensus Conference recommended that these four types of assessments should be conducted:
1) Emergent risk needs should be assessed immediately upon arrival at a secure facility; 2) A comprehensive mental health assessment should be conducted on all youths at the facility to identify those needing more thorough mental health assessments; 3) Prior to community re-entry, all youth should be assessed to facilitate transition and referral to community mental health services; and 4) continued re-assessments should take place after the youths have returned to the community, to prevent re-offenses.
In the past two decades, several measures have been developed to assess juvenile offenders’ mental health and associated risks (Grisso, 2005). These measurement instruments aim to be accurate, reliable, and thorough while being fairly quick and inexpensive to administer.
Tools for Mental Health and Associated Risk Assessment
There are currently several well-validated assessment measures used to predict the likelihood of re-offending upon release, mental health treatment needs, and danger towards self (suicide ideation and suicide attempts) and others, based on the presence or absence of substance abuse, suicide ideation, personality traits, thought disturbance, and depression-anxiety. Below we describe several of the most common assessment tools used in juvenile justice research and practice. Several scales are actuarial in nature while others integrate actuarial assessment with supplemental clinical judgment. Instruments are categorized according to their utility as brief screening tools, comprehensive assessment instruments, or risk assessments predicting recidivism or dangerousness in the future. Descriptions are intended to give a brief overview and should not be considered full reviews. For more detailed information on each of these instruments, readers are directed to Grisso, Vincent, and Seagrave’s (2005) Mental health Screening and Assessment in Juvenile Justice or to literature by each scale’s developer.
Brief Screening Tools
Brief screening tools are instruments that can be administered very quickly (usually in 30 minutes or less) and help staff to identify youth who may be of immediate risk to self or others. Furthermore, the screenings should help staff identify youth in need of more comprehensive mental health assessment. These instruments should be easily administered by front-line staff with little specialized training, allowing for quick and inexpensive use. Brief screening tools should not be used to inform treatment plans; instead their utility is in identifying those youth in need of emergency mental health services or those who need more comprehensive assessment that can then inform treatment needs. Table 1 describes the strengths and limitations of three commonly used brief screening tools.
MAYSI-2. The Massachusetts Youth Screening Instrument—Version 2 (MAYSI-2) was developed by Grisso and Barnum (2003) as a self-report measure to identify youth entering the juvenile justice system with thoughts, feels, or behaviors indicative of mental health problems. The MAYSI-2 can be administered by pencil-paper or by CD-ROM and consists of 52 yes-no questions asking whether each item is true for the youth. Seven subscales are assessed, including alcohol/drug use, angry-irritable, depressed-anxious, somatic complaints, suicide ideation, thought disturbance, and traumatic experiences. This objective measure includes cut-off scores from a normative juvenile justice sample that can be used as indicators of clinical significance (Grisso & Quinlan, 2005). Research evaluating the reliability of the MAYSI-2 reports internal consistency ranging from .61 to .86 (Grisso et al., 2001) and support for test-retest reliability on most subscales (Cauffman, 2004). Similar positive findings were found in studies of validity comparing the MAYSI-2 to other standardized scales (Espelage et al., 2003) and to the DSM-IV (Wasserman et al., 2004). Of interest were several studies that found the MAYSI-2 to predict future behaviors such as institutional maladjustment, sentence length, and necessary intervention for suicide risk and assaultive behavior (Cauffman, 2004; Stewart & Trupin, 2003). Cauffman and MacIntosh (2006) recently found different properties on some subscales, in particular the alcohol/drug use, anger-irritability and suicide ideation subscales, across ethnic and gender groups. Further research should continue to examine the extent to which these subscales are valid measures for female and ethnic minority youth.
POSIT. The Problem-Oriented Screening Instrument for Teenagers (POSIT) was developed by Rahdert (1991) as a self-report brief screening to identify troubled youths’ problems in psychosocial functioning requiring further assessment. The POSIT, available by pencil-paper or by CD-ROM, consists of a self-administered questionnaire with 139 yes-no questions and assesses 10 functional areas, including substance use/abuse, physical health, mental health, family relations, peer relations, educational status, vocational status, social skills, leisure/recreation, and aggressive behavior/delinquency (Dembo & Anderson, 2005). Youths’ total scores in each problem area can be compared to empirically-based cut-off scores allowing for a classification of low-, medium-, or high-risk for that problem area. While the POSIT is objectively scored, collateral information is recommended to validate youths’ responses. Research evaluating the reliability of the POSIT indicates internal consistency exceeded .70 and test-retest reliability significantly better than chance (Knight, Goodman, Pulerwitz, & DuRant, 2001). Hall, Richardson, Spears, & Rembert (1998) found high construct validity for the POSIT. Preliminary research indicate the POSIT is useful in classifying youth by predicting return to the juvenile justice system (Dembo, Turner, et al., 1996).
CAFAS. The Child and Adolescent Functional Assessment Scale (CAFAS) was developed by Hodges (2000a) to assess youths’ everyday psychosocial functioning across school, home, community, and work settings. Different ratings (parents, teachers, youth) of youth’s behaviors are obtained across 10 subscales (school/work, home, community, behavior toward others, moods/emotions, self-harmful behavior, substance use, thinking, material needs, and family/ social). Questions are asked for each subscale that identify severe, moderate, mild, or no impairment and also include questions that indicate strengths or protective behaviors exhibited by the youth. Raters score the CAFAS after collecting information based on both their own observations and a family of instruments that assess the youth’s and their caregiver’s perspective on everyday functioning. Studies report positive results for test-rest reliability, validity, and ability to predict level of service needs (Hodges & Wong, 1997).
Comprehensive Assessment Instruments
Before forensic mental health specialists, correctional counselors, and probation officers can recommend a treatment plan, comprehensive risk assessment data must be collected. In comparison to brief assessment tools, comprehensive assessment instruments more thoroughly assess several domains of youths’ mental health, personality, and psychosocial characteristics. These assessments often involve longer, more intensive interviews and several also collect collateral information from other settings in the youth’s life (i.e., teachers, parents, or chart information). Comprehensive assessment instruments help to clarify mental health needs, can inform treatment planning, and are most often conducted by professionals or require more involved training. Table 2 describes the strengths and limitations of three commonly used comprehensive assessment instruments.
DISC. The Diagnostic Interview Schedule for Children: Present State Voice Version (Voice DISC) was developed by Shafer, Fisher, Lucas, Dulcan, & Scwab-Stone (2000) to assess mental health problems and provides a diagnosis by evaluating how youth meet DSM-IV criteria. Now self-administered on the computer, the assessment employs a unique pattern of questions based on respondents’ answers to previous questions, assessing the degree to which they meet criteria for more than 30 diagnoses (Wasserman, McReynolds, Fisher, & Lucas, 2005). Subscales include anxiety, mood, disruptive behavior, substance use, and miscellaneous (eating disorders, tic disorders, etc.). After the assessment tool determines a youth meets diagnostic criteria, further questions inquire about the severity and frequency of these problems in an attempt to understand impairment. However, youth may be limited in their ability to recognize the consequences of their own behaviors, and it is suggested that clinicians use collateral information to determine impairment. DISC reports include a list of those diagnoses for which the youth met criteria, impairment and symptom scores, and a list of “clinically significant symptoms.” Acceptable reliability for most diagnoses and good test-retest reliability have been reported (Shaffer et al. 2000). Moderate to poor correlation with clinician diagnosis has been found (Aronen, Noam, & Weinstein, 1993); however, independent clinical diagnosis is known to be fairly subjective and unreliable.
MMPI-A. The Minnesota Multi-phasic Personality Inventory—Adolescent (MMPI) was developed by Archer (1997) and adapted for adolescents by Butcher et al. (1992) and is the most widely used personality assessment (Archer & Baker; 2005). The MMPI consists of 478 items with validity scales (e.g. defensiveness, tendency to exaggerate, response consistency), clinical scales (e.g. psychopathology such as depression, anxiety, schizophrenia, antisocial behaviors), content scales (e.g. externalizing behaviors, anger, low self-esteem), and supplementary scales (immaturity, repression). Raw scores are converted to t-scores and are compared to normative scores, resulting in classification of youth who are clinically elevated, marginally elevated, or typically adolescent. Early research found scale 4 (psychopathic deviate) especially helpful in predicting delinquency (Hathaway & Monachesi, 1963). This is confirmed in later studies that found scale 4 (psychopathic deviate), scale 8 (schizophrenia) and scale 9 (Hypomania) predictive of higher rates of delinquency (Archer, Bolinskey, Morton, & Farris, 2003). Over 100 studies have examined aspects of the MMPI-A, and the instrument is known for its good reliability and validity ; a good resource is a review by Forbey (2003).
MACI. The Millon Adolescent Clinical Inventory (MACI) developed by Millon (1993) as a short assessment that provides clinical information on a variety of psychological problems, including psychopathology, peer difficulties, family problems, and confusion about self (Salekin, Leistico, Schrum, Mullins, 2005). It also assesses a balance of externalizing/delinquency risk factors as well as suicidal tendency and risk towards self. Based on the DSM-IV, the MACI includes: 3 validity scales (disclosure, desirability, debasement), a reliability scale, 7 clinical syndrome scales (eating dysfunction, substance abuse, delinquent predisposition, impulsive propensity, anxious feelings, depressive affect, suicidal tendency), 12 personality scales (introversive, inhibited, doleful, submissive, dramatizing, egotistic, unruly, forceful, conforming, oppositional, self-demeaning, borderline tendencies), and 8 expressed concern scales (identity confusion, self-devaluation, body disapproval, sexual discomfort, peer insecurity, social intensity, family discord, child abuse). Base-rate scores are calculated and are interpreted by mental health professionals, who first examine validity and reliability before identifying problem scales with elevated base-rate scores. The MACI is shown to have good internal and test-retest reliability and concurrent and predictive validity (Millon, 1993). A recent study by Taylor, Skubic-Kemper, Loney, and Kistner (2006) extends support for using the MACI as a tool for classifying subtypes of serious juvenile offenders. Furthermore the MACI has been shown useful in assessing clinical change from intake to discharge in inpatient settings (Piersma, Pantle, Smith, Boes, & Kubiak, 1993) and is predictive of recidivism (Salekin, Ziegler, Larrea, Anthon, and Bennet (2003)).
Risk for Recidivism and Dangerousness Assessment Tools
Research has identified several factors that put youth at risk for future violence or recidivism. While no definitive list of factors has been developed, research has shown that there are common pathways to recidivism that can be predicted with some accuracy. These factors have been used to compose assessment tools that measure youths’ risk of re-offending once released into the community. These instruments help juvenile justice centers make decisions to protect the community and identify need for further services such as case management. Risk for recidivism assessments often involve collecting collateral information from parents or chart materials in addition to interviewing, and is time-intensive. Thus, these assessments require specialized training or a professional degree to administer and score. Assessments of future risk behavior include varying degrees of clinical judgment to interpret the results and make decisions. Table 3 describes the strengths and limitations of three commonly used assessments of future risk.
YLS/CMI . The Youth Level of Service/Case Management Inventory (YLS/ CMI; Hoge, Andrews, & Leschied, 2002) is designed to predict juvenile offender recidivism as well as case management needs, making it especially useful in planning for transitions out of the juvenile justice system. The YLS/CMI assesses the offender as high or low risk, assesses need by targeting services due to risk factors, and assesses responsivity or reaction to interventions. The YLS/CMI is composed of six sections: 1. Assessment of risk and needs (42-item checklist assessing prior/current offenses, family circumstances, education/employment, peer associations, substance abuse, leisure, personality/behavior, and attitude); 2. Summary of risk/need factors (comparing scales to normative ranges); 3. Other needs/special circumstances (situational information such as parental drug use or behavioral records that add information specific to youth); 4. Professional override feature (asks clinician to use clinical judgment considering all relevant information to rate youth’s risk level); 5. Contact level (intensive services should be recommended for high risk youth); and 6. Case management plan (specific goals and objectives for reaching goals). Due to the complexity and knowledge it requires, the YLS/CMI is completed by a trained professional and purposely incorporates a degree of clinical judgment to supplement the objective portions of the assessment. Adequate internal consistency (Rowe, 2002) and inter-rater reliability have been found in empirical studies (Schmid, Hoge, & Robertson, 2002), except for the leisure/recreation subscale, which has a wide range of interrater reliability (.05-.92). Several subscales of the YLS/CMI have been correlated with other externalizing measures (Rowe, 2002). Ability to predict new charges, new convictions and serious offense charges have been consistently demonstrated with males and more inconsistently for girls (Rowe, 2002; Schmidt et al., 2002).
SAVRY . The Structured Assessment of Violence Risk in Youth (SAVRY; Bartel, Borum, & Forth, 2000) involves professional judgment based on systematic appraisal of the degree to which youth demonstrate risk factors for future violence. The appraisal involves assessment of 6 protective factors (prosocial involvement, strong social support, strong attachments and bonds, positive attitude toward intervention and authority, strong commitment to school, resilient personality traits). The instrument also assesses 24 risk factors including: historical (history of violence, of nonviolent offending, early initiation of violence, history of self harm, childhood exposure to maltreatment, parental criminality, early caregiver disruption, poor school achievement), individual (negative attitudes, risk taking/impulsivity, substance use difficulties, anger management problems, low empathy/remorse, ADHD difficulties, poor compliance, low interest/commitment to school) and social/environmental (peer delinquency, peer rejection, stress/poor coping, poor parental management, lack of personal/social support, community disorganization) domains. Information should be gathered by the examiner through interviews with the youth, review of records, and observation (Borum, Bartel, Forth, 2005). Numerical ratings are not the goal of this assessment; identifying empirically validated risk factors specific to each youth is the goal. Thus clinicians are faced with reviewing the identified risk factors and making a clinical judgment about a youth’s overall risk. Inter-rater reliability is moderate to high (.81) (Catchpole & Gretton, 2003) and studies show support for concurrent validity as compared to the YLS/CMI and PCL:YV (Catchpole & Gretton, 2003). Moderate yet significant correlations were found between the SAVRY and measures of violence and aggression (McEachran, 2001; Gretton & Abramowitz (2002). Additionally, those youth characterized as low risk had violent recidivism rates (6 percent) much lower than those characterized as moderate (14 percent) or high risk (40 percent) (Catchpole & Gretton, 2003).
PCL:YV . The Hare Psychopathy Checklist: Youth Version (PCL:YV; Forth, Kosson, & Hare, 1990) uses multiple sources of information across interpersonal, affective and behavior domains to identify symptoms predictive of serious psychopathy in adolescents. The examiner uses information from an intensive interview with the youth, collateral sources, and review of the chart to rate the youth according to a 20-item checklist including: impression management, grandiose sense of worth, stimulation seeking, pathological lying, manipulation of personal gain, lack of remorse, shallow affect, lack of empathy, parasitic orientation, poor anger control, impersonal sexual behavior, early behavior problems, lacks goals, impulsivity, irresponsibility, failure to accept responsibility, unstable interpersonal relationships, serious criminal behavior, serious violence of conditional release, and criminal versatility. Total scores provide the number of psychopathic features observed for each youth but do not result in cut off or classification. However raters can compare youth’s scores to percentile scores based on institutional, probation, and community samples. After extensive training required to administer the PCL:YV, inter-rater reliability scores are generally high (.90-.96) and internal consistency is adequate (.85-.94, Forth et al., 2003). Moderate correlations with reports of delinquency, externalizing symptoms and aggression are reported, while the PCL: YV (as intended) does not correlate with measures of internalizing disorders (Cambell et al., 2004). Recent studies report the PCL:YV significantly predicted both violent and nonviolent recidivism (Corrado, Vincent, Hart, & Cohen, 2004) as well as clean urine screens and participation in treatment (O’Neill et al., 2003a). Eden, Buffington, Colwell, Johnson, & Johnson (2002) further support the ability of the PCL: YV to predict disciplinary infractions in their sample juvenile sex offenders. However, Spain, Douglas, Poythress, & Epstein (2004) found negative results, with no relationship evident between the PCL:YV and treatment progress.
Conclusion and Suggestions for Further Research
Several Brief Assessment Screens have been developed, and research supports their ability to identify youth with emergent risk, and screen for youth who should receive more comprehensive mental health assessments. A study by Wasserman et al. (2004) confirms that such brief tools as the MAYSI-2 are useful for identifying youth who have a possible mental health problem so that they can be further evaluated by such tools as the DISC to identify a more specific diagnosis. These brief assessment tools should be utilized for these purposes and they are most effective when administered promptly upon the youth’s arrival at a secure setting.
Several of the assessments described in Table 2 address the aim of tapping different domains of functioning in the assessment. Some tools utilize multiple sources of data such as files and collateral sources of information in addition to self report, while others do not. This reflects the constant struggle to provide thorough assessment with limited time/expense resources. Considering the importance of accurately assessing and treating offender mental health problems, we conclude that assessments that tap into a range of information sources are worth the time and effort.
It is unclear from the current empirical evidence how effective the above reviewed assessment tools are in re-assessing risk once the youth has been reintegrated into the community. Further research is needed to clarify whether current assessment tools are useful for post-incarceration re-assessment or whether other assessments, which take into consideration the importance of environmental transition, should be developed for that purpose.
It is important to mention that, while great progress has been made in beginning to understand and assess juvenile offender mental health risk, there is much work to be done in testing the ability of these assessment tools to generalize beyond the population for which they were developed. Recent research by Wasserman, McReynolds, Ko, Katz, and Schwank (2005) examining the prevalence of psychiatric disorders among youths at probation intake, reported that violent female offenders were up to five times more likely to report anxiety disorders than their male counterparts. Furthermore, of youth with conduct disorder, girls seemed to be more likely than boys to have complex diagnoses due to elevated rates of co-occurring internalizing disorders. Research also shows that ethnic minority youth are overrepresented in the juvenile justice system, yet few mental health risk assessment tools have been tested across gender or ethnic groups (Devine, Coolbaugh, & Jenkins, 1998). Further research is needed to evaluate assessment instruments with female offenders and ethnic minority offenders, as research suggests adaptations may need to be made to accurately assess the needs of these vulnerable groups (Cauffman et al., 2006).
One additional facet of risk assessment appears particularly lacking in the field of juvenile justice. Risk assessments of juvenile offenders need to identify those youths likely to re-offend into adulthood, and who are likely to be the chronic career criminals. Several classic studies have documented the pattern of desistance of delinquent behavior in young adulthood (Farrington & West, 1977; Gottfredson & Hirschi, 1990; Elliott, et al., 1983). Gottfredson and Hirshi’s (1990) key finding was that most juveniles discontinue their delinquent acts in early adulthood. Farrington and West’s classic study indicated that only 22.6 percent of their research subjects had subsequent convictions as adults. Elliott and associates found that only 2 to 3 out of every 10 adjudicated violent juveniles were arrested for violent crimes in adulthood. Also noteworthy is the fact that while the majority of delinquent youth do not seem to present a long-term risk of re-offending in adulthood, there is a small group of 5 to 6 percent of different cohorts that chronically persist in crime into adulthood and are responsible for a high volume of multiple offenses. Wolfgang, Figlio and Sellin (1972) reported in their birth cohort study that approximately 6 percent of their subjects were responsible for over 50 percent of the official crimes by the cohort. These chronic offenders were likely to have poor school grades and achievement, low IQ test scores, be of non-white racial background, low socioeconomic status, and school dropouts. Farrington (1985) found similar results of chronic offending by a small percentage of offenders—6 percent committing 49 percent of the criminal offenses.
Risk assessment instruments to date have not been well tested in their ability to differentiate those youth who will chronically offend into adulthood from those who are temporary adolescent offenders. Perhaps Hare’s Psychopathy checklist comes closest to beginning to identify this type of particularly serious long-term offender. However, it is clear that, while short-term recidivism is important to assess, much work is needed to expand current mental health assessment tools to better identify the potential long-term chronic offenders.