How to Protect your Client from Bad Risk Assessments Thomas A. Powell, Ph.D. John C. Holt, Ph.D. Vermont Forensic Assessment, Shelburne, Vermont June 12, 2008 (C) 2008 Vermont Forensic Assessment All rights reserved PLLC Outline How things are supposed to work Static and dynamic measure of any type of criminal recidivism and offender needs Level of Service Inventory - Revised Assessment perspective Risk principle Need principle Responsivity principle (C) 2008 Vermont Forensic Assessment All rights reserved Outline Static actuarial measures of sexual recidivism risk STATIC-99 RRASOR VASOR Dynamic measures of sexual recidivism risk SONAR STABLE/ACUTE Treatment Progress Scale (C) 2008 Vermont Forensic Assessment All rights reserved Outline Additional topics of interest The use of multiple risk instruments: Static 99, RRASOR and VASOR as policy “Denial” - a misleading predicator of recidivism. Conflict of interest by assessors and treatment providers. The DOC case staffing: Salem Witch Trials redux? (C) 2008 Vermont Forensic Assessment All rights reserved QuickTime™ and a TIFF (Uncompressed) decompressor are needed to see this picture. (C) 2008 Vermont Forensic Assessment All rights reserved How Things are Supposed to Work PSI preparation - Policy 342.01 (rev. 4/9/07) http://doc.vermont.gov/about/policies/rpd/fin al_commissioner-signed_psi_3.22.07.pdf Prescriptive format Defines LSI-R as “an objective, quantifiable instrument that provides a consistent and valid method of predicting risk to re-offend and a reliable means of measuring offedner change over time …” (C) 2008 Vermont Forensic Assessment All rights reserved How Things are Supposed to Work RRASOR: “a four-item actuarial risk measure used to aid in assessing sexual recidivism risk …” Static-99: “… a 10-item actuarial risk measure is used in a similar manner as the RRASOR.” VASOR: “a risk assessment scale … designed to assist PO’s in making placement and supervision decisions.” (C) 2008 Vermont Forensic Assessment All rights reserved How Things are Supposed to Work “A PSI provides the sentencing judge … with relevant information on which to base a sentencing decision. It also reflects a professional judgment of and assessment by the DOC regarding risk management of the defendant.” “The PSI investigator can use information collected during the PSI, particularly during the defendant interview, to complete necessary risk assessment tools, e.g. the LSI-R.” (C) 2008 Vermont Forensic Assessment All rights reserved How Things are Supposed to Work “Provide a narrative explanation of the numerical results of assessment tools (e.g. LSI-R), including an examination of key risk areas and a summary of any court-ordered clinical assessments.” “Briefly summarize the major points of the report …” “The report will conclude with a recommendation … based upon the defendant’s offense, background and assessed risk.” (C) 2008 Vermont Forensic Assessment All rights reserved How Things are Supposed to Work “Note: Only if requested by the Court, the PSI investigator will include a recommended minimum and maximum term of incarceration or a minimum and maximum term of suspended sentence.” (emphasis not added by us) (C) 2008 Vermont Forensic Assessment All rights reserved General Risk Assessment Level of Service Inventory - Revised Criminal History Education/Employment Financial Family/Marital Accommodation (the influence of address) Leisure/Recreation (spare time use) Companions (friends and associates) Alcohol/Drugs (mostly looking at past year) Emotional/Personal (psychological status) Attitudes/Orientation (antisocial/prosocial views) (C) 2008 Vermont Forensic Assessment All rights reserved ASSESSMENT PERSPECTIVE Prediction 2 Environmental Contingencies Assessment 1 Behavior Psychological Predisposition Time 1: Primary objective is to understand factors linked to future behavior 2: Ultimate assessment objective is to estimate future behavior Simourd, D.J. (1999, May), On a general forensic evaluation model: A practical framework for offender assessment. Paper presented at the annual meeting of the Canadian Psychological Association, Halifax, N.S. Risk Principle 1. 2. 3. Reserve intensive levels of corrections supervision and treatment for higher risk cases. Do not use these levels of supervision and treatment for lower risk cases. Use of valid and reliable risk assessment tools gives the criminal justice system a triage capability. (C) 2008 Vermont Forensic Assessment All rights reserved Need Principle 1. 2. 3. 4. There are many needs and potential treatment targets identified in corrections populations. Some are related to criminal risk; some are not. Focus on those needs which are “criminogenic” (ie. causally related to crime) such as social affiliations, drugs and criminal attitudes. When criminogenic needs are addressed, the risk of recidivism declines. (C) 2008 Vermont Forensic Assessment All rights reserved Responsivity Principle 1. 2. 3. 4. When providing treatment services for forensic or corrections populations, use methods which are matched to the learning style of the population. Cognitive-behavioral, skill-building, and relapse preventing interventions fit this model. Non-directive, non-specific “talk therapy” methods do not. The use of appropriate intervention models lowers risk; other models have either no effect on recidivism or they may actually raise it. (C) 2008 Vermont Forensic Assessment All rights reserved Sex Offense Specific Measures Static Actuarial Measures RRASOR Static-99 VASOR Dynamic/Treatment Measures SONAR STABLE/ACUTE Sex Offender Treatment Needs and Progress Scale (C) 2008 Vermont Forensic Assessment All rights reserved RRASOR Rapid risk assessment of sex offender recidivism ww2.ps-sp.gc.ca/publications/corrections/ 199704_e.pdf Four item scale All four are incorporated into the STATIC-99 Total score 0-6 0 1,2 3,4 5,6 Low Moderate-Low Moderate-High High (C) 2008 Vermont Forensic Assessment All rights reserved RRASOR - Rapid risk assessment of sex offender recidivism STATIC-99 Ten item scale Total score 0-12 0,1 2,3 4,5 6+ Low Moderate-Low Moderate-High High (C) 2008 Vermont Forensic Assessment All rights reserved STATIC-99 Widely used in criminal, forensic and correctional settings. Instrumental in most sexually violent predator (SVP) determinations. Not as easy to score as it seems. Many scoring rules to understand. (C) 2008 Vermont Forensic Assessment All rights reserved STATIC-99 The Static-99 scoring manual is 80 pages long. http://ww2.pssp.gc.ca/publications/corrections/pdf/Static99-coding-Rules_e.pdf Or google “Static-99” (C) 2008 Vermont Forensic Assessment All rights reserved STATIC-99 Issues Like most actuarial risk assessment instruments, it is only +/- 70% accurate. False positives - over-predicting risk False negatives - under-predicting risk Over-reliance on unchanging static risk factors to the exclusion of changing dynamic risk factors is a significant problem in Vermont (and elsewhere). (C) 2008 Vermont Forensic Assessment All rights reserved STATIC-99 - age and recidivism STATIC-99 - age and recidivism “There were few sexual offenders of any type in the advanced age category (>50 years !!) … and their recidivism rates were generally low.” • decreasing sex drive • fewer opportunities Hanson, K.W. (2002). Recidivism and age. Journal of Interpersonal Violence, 17 (10), 1046-1062. STATIC-99 - time at risk Listserv communication from Hanson and Harris (3/10/08) “Recently, (we) found follow-up recidivism rates to be below that which would be expected given those found in the original STATIC-99 samples of origin. This has led to some question as to whether there has been a shift in the nature of sexual recidivism and whether the estimates generated from the samples of origin are still valid.” VASOR Vermont assessment of sex offender risk Weighs 19 variables 13 on the Reoffense Risk Scale 6 on the Violence Scale According to the manual, scores are plotted on an overall risk chart Reoffense Risk Scale x Violence Scale (C) 2008 Vermont Forensic Assessment All rights reserved VASOR Prior sex offenses Prior adult convictions Court order violations Force used during current arrest Relationship to victim Male victim History of exhibitionism Deviant sexual fixation Alcohol of drug use (past five years) Address changes (past year) Time employed/ in school Past treatment success Amenability to treatment Prior convictions for violent crimes Prior conviction for crime involving deadly weapon Force used during current offense Sexual intrusiveness of current offense Physical harm to victim Victim <5, >55, or handicapped (C) 2008 Vermont Forensic Assessment All rights reserved VASOR - Daubert criteria issues No peer-reviewed studies on its reliability and validity Manual relies on studies presented in poster format Manual clearly marks the VASOR as a “research version” Not cited in major reviews of other actuarial scales Used rarely outside of Vermont (C) 2008 Vermont Forensic Assessment All rights reserved VASOR - Validity issue According to Georgia Cumming, new cases are being added to the VASOR database, and within “a couple of weeks” there will be new cut points for describing the risk level of cases, based on their scores. (C) 2008 Vermont Forensic Assessment All rights reserved Dynamic Measures of Sexual Recidivism Risk SONAR: Sex offender need assessment rating Stable 2007 and Acute 2007 Treatment Progress Scale (Vermont-specific) (C) 2008 Vermont Forensic Assessment All rights reserved SONAR - sex offender needs assessment rating rates dynamic factors related to recidivism improves the predictive accuracy and utility of actuarial (historical factors) instruments by considering changing circumstances in the offender’s life (ie. dynamic factors) http://ww2.pssp.gc.ca/publications/corrections/200001b_e .asp (C) 2008 Vermont Forensic Assessment All rights reserved Stable and Acute - 2007 Stable factors Significant social influences Capacity for relationship stability Emotional identification with children Hostility toward women General social rejection Lack of concern for others Impulsive Poor problem solving skills (C) 2008 Vermont Forensic Assessment All rights reserved Stable and Acute - 2007 Stable factors (cont.) Negative emotionality Sex drive preoccupation Sex as a coping mechanism Deviant sexual preference Deviant sexual interests in possible remission Cooperation with supervision (C) 2008 Vermont Forensic Assessment All rights reserved Stable and Acute - 2007 Acute factors Victim access Hostility Sexual preoccupation Rejection of supervision Emotional collapse Collapse of social supports Substance abuse (C) 2008 Vermont Forensic Assessment All rights reserved Stable and Acute - 2007 The most recent version incorporates the Static-99 variables and the most riskrelevant dynamic variables to describe both risk level and supervision intensity needs for a client. This represents best evidence-based practice at this time. (C) 2008 Vermont Forensic Assessment All rights reserved Vermont Treatment Progress Scale http://www.csom.org/pubs/SexOffTreatScale.pdf McGrath & Cumming (2003) Bob McGrath, MA, licensed psychologist, clinical director, Vermont Treatment Program for Sexual Aggressors Georgia Cumming, administrative director, Vermont Treatment Program for Sexual Aggressors (C) 2008 Vermont Forensic Assessment All rights reserved Vermont Treatment Progress Scale Sexual deviancy Criminality Self-regulation Treatment and supervision cooperation Lifestyle stability Social supports (C) 2008 Vermont Forensic Assessment All rights reserved Multiple Static Assessments DOC uses three static actuarial scales RRASOR, STATIC-99, VASOR It is unnecessary to use more than one static actuarial assessment “...combining scales did not provide a statistically significant or consistent advantage over the single best actuarial risk scale in predicting either serious or sexual recidivism.” Seto (2005) (C) 2008 Vermont Forensic Assessment All rights reserved Multiple Static Assessments So why does DOC policy require the use the Static-99, RRASOR and VASOR on all sex offenders? The HIGHEST of the three ratings is used Low Medium low Medium Medium high High Fishing for the highest score? (C) 2008 Vermont Forensic Assessment All rights reserved Multiple Static Assessments If there are going to be multiple instruments, why use three static factor assessments? Why exclude dynamic variables? Sonar Stable/Acute 2007 Vermont Treatment Progress Scale Evidence-based practice vs. not (C) 2008 Vermont Forensic Assessment All rights reserved Denial “Denial” vs. minimization Weak predicator of recidivism and treatment compliance Hanson and Morton-Bourgeon (2005) article p. 1158 (C) 2008 Vermont Forensic Assessment All rights reserved Conflict of Interest DOC relies on a network of providers who provide both forensic assessments and clinical interventions. “Irreconcilable conflict between therapeutic and forensic roles” (Greenberg & Shuman, 1997). “Failure to maintain adequate separation of roles has a number of counter-therapeutic consequences in sex offender management” (Covell & Wheeler, 2006). (C) 2008 Vermont Forensic Assessment All rights reserved Conflict of Interest National professional organizations specifically highlight the importance of avoiding dual roles American Psychological Association American Academy of Psychiatry & the Law Association for the Treatment of Sexual Abusers (ATSA) (C) 2008 Vermont Forensic Assessment All rights reserved Conflict of Interest “Members recognize that there may be potential conflicts of interest when they provide both evaluation and treatment services to the same person. When it is necessary to fulfill both functions (for example, in rural settings or institutions), members take reasonable steps to manage and resolve any conflicts in the best interest of the client and the community.” ATSA practice standards and guidelines (2005) (C) 2008 Vermont Forensic Assessment All rights reserved DOC Case Staffings 2 to 10 P.O.’s gathered with a supervisor to discuss a case. Victims’ advocates, others, included? Referred to, in policy, as a “peer review process” No record or transcript. (C) 2008 Vermont Forensic Assessment All rights reserved DOC Case Staffings “This is an opportunity for the PSI investigator to present the information collected to fellow staff members for comment and advice on further areas of investigation, additional needed information, and a sentencing recommendation decision.” (C) 2008 Vermont Forensic Assessment All rights reserved DOC Case Staffings This is not considered best-practice for improved decision-making. Nationally, reviews typically involve small panels of higher-trained professionals who critically consider the evidence and determine whether to override lower level decisions. In Vermont there is a strong tendency toward peer group-think. (C) 2008 Vermont Forensic Assessment All rights reserved DOC Case Staffings “I have heard of a very similar process occurring in Boston relative to their civil commitment process. (It) was thought to be so bad that a local TV station did an "expose" on the process and how political and poor the decision making was - despite needing to emphasize during he program that the network was not trying to protect sex offenders by making the expose. They instead pointed out that there were obvious errors in decisions in both directions - holding people who should not be held, and letting others go who should not have been let go.” Listserv colleague, commenting on Vt case staffing process (C) 2008 Vermont Forensic Assessment All rights reserved How to Protect your Client from Bad Risk Assessments Presented by Vermont Forensic Assessment, PLLC Thomas A. Powell, Ph.D. & John C. Holt, Ph.D. 4281 Shelburne Road Shelburne, VT 05482 802-985-2412 www.vtfa.com (C) 2008 Vermont Forensic Assessment All rights reserved