1. What are evidenced-base risk and needs assessment practices?

Transcription

1 Identifying and Measuring What s Important in Sex Offender Treatment and Supervision: How to Use a Treatment Needs and Progress Scale Connecting for Children s Justice 8 th Annual Conference Nashville, TN November 20-22, 2011 Robert J. McGrath, M.A., Clinical Director Vermont Treatment Program for Sexual Abusers Workshop Outline 1. What are evidence-based assessment practices? 2. What is the SOTIPS? 3. How does the SOTIPS perform? 4. How do I score the SOTIPS? 5. How can the SOTIPS be used in practice? 1. What are evidenced-base risk and needs assessment practices? Association for the Treatment of Sexual Abusers (ATSA) Practice Standards and Guidelines (2005) evaluate client s risk and clinical needs before beginning treatment periodically re-evaluate client s treatment progress Principles of Effective Services RNR 1. Risk Principle Who to treat? - Assess offenders risk to reoffend. - Focus treatment on offenders at moderate risk or higher. 2. Need Principle What to treat? - Assess offenders treatment needs. - Focus treatment on offenders criminogenic needs. 3. Responsivity Principle How to treat? - Assess offenders learning styles. - Match services to offenders learning styles. Programs that adhere to RNR principles have lower rates of sexual recidivism than those that do not. 23 sex offender treatment outcome studies (Hanson et al., 2009) Odds Ratio Adhere to none Adhere to 1 principle Adhere to 2 principles Adhere to all 3 principles 1

2 Implementing RNR principles requires conducting risk and needs assessments. Predicting things is difficult, especially when they re in the future. Yogi Berra We know which risk factors are most closely linked to sexual offending. (Mann, Hanson & Thornton, 2010; Hanson & Bussiere, 1999) Types of Risk Factors Static Stable Acute Static Risk Factors Prior sex convictions Prior criminal convictions Prior violent convictions Early onset Younger age Any male victims Any unrelated victims Any stranger victims Never married Dynamic Risk Factors Sexual preoccupation Deviant sexual interests Offense supportive attitudes Poor adult attachments Lifestyle impulsivity Resistance to supervision Poor problem solving Grievance and hostility Negative social influences Heart Attack prior heart attack cholesterol stressful event family history weight drug reaction Sexual Reoffending prior sex offense sexual interests victim access prior other crimes sexual attitudes supports collapse unrelated victims impulsivity intoxication But how do we weight and add up these risk factors? Ignore them and use our intuition? Use unstructured professional judgment? Use a structured approach? What sex offender risk assessment methods work best? Hanson, R.K., & Morton-Bourgon, K.E. (2009.)The accuracy of recidivism risk assessments for sexual offenders: A metaanalysis of 118 prediction studies. Psychological Assessment, 21, Examined studies from 1972 to samples from 16 countries 45,398 sexual offenders 2

3 Which methods work best? Hanson & Morton-Bourgon (2009) Meta-analysis of 118 studies Methods Effect Size d (95% CI) k Structured - Empirical actuarial moderate.67 ( ) 81 - Mechanical moderate.66 ( ) 29 Unstructured - Unstructured small.42 ( ) 11 What is the actuarial approach? Insurance companies use to set group rates Sex offender risk instruments each have set risk factors. Each risk factor has a certain weight. Weights are added and yield a total score. Total score is associated with a % rate of reoffense. Note. These percentages vary across jurisdictions, time and populations but higher scores are consistently associated with higher reoffense rates & lower scores with lower rates. What structured static risk instruments do programs use? Programs for Adult Males McGrath et al. (2010). Safer Society Survey. Static-99R, Static-2002R, MnSOST-3, & VASOR-R What is the predictive accuracy of structured static risk instruments? (Hanson & Morton-Bourgon, 2009) d (95% CI) N (k) Static ( ) 20,010 (63) Static ( ) 3,330 (8) MnSOST-R.76 ( ) 4,672 (12) SVR-20 Totals.68 ( ) 1,699 (10) Risk Matrix ( ) 2,755 (10) What structured dynamic risk instruments do programs use? Programs for Adult Males McGrath et al. (2010). Safer Society Survey Use of SVR-SO not surveyed. What combined static and dynamic assessment schemes add incrementally to predicting sexual recidivism? Stable 2007 and Acute 2007 and Static-99 Hanson, Harris, Scott, & Helmus (2007) SRA (or SARN) and Static-99 or RM2000 Thornton (2002) SVR-SO and Static-99 Olver, Wong, Nicholaichuk, & Gordon (2007) SOTIPS and Static-99R or VASOR-R McGrath, Lasher, & Cumming (2011) 3

4 SOTIPS Sexual Offender Treatment Intervention and Progress Scale 2. What is the SOTIPS? Purpose: Assess dynamic risk, treatment need and progress among adult male sex offenders; community or residential Description: 16 dynamic risk factors scored on a 4 point scale every 6 months Used With: Static-99R or VASOR Qualifications: Mental health professionals; probation and parole officers Availability: From authors; no cost SOTIPS Development SOTIPS 2000 NIJ grant Conducted literature review, consulted an expert panel (W. Ballantyne, S. Huot, R. Goldman, and R. Packard), and began drafting the scale Field tested and wrote scoring criteria for the 22-item Sex Offender Treatment Needs and Progress Scale. Released first version of the manual Conducted initial predictive validity study NIJ grant Revised and renamed scale the SOTIPS; Reduced scale to 16 items; Developed models to integrate it with Static-99R and VASOR-R Released new scoring manual. 16 Dynamic Risk Items 1. Offense Responsibility 9. Treatment Cooperation 2. Sexual Interests 10. Supervision Cooperation 3. Sexual Attitudes 11. Emotion Management 4. Sexual Behavior 12. Problem Solving 5. Risk Management 13. Impulsivity 6. Criminal Attitudes 14. Employment 7. Criminal Behavior 15. Residence 8. Stage of Change 16. Social Influences SOTIPS McGrath, R. J., Lasher, M. P., & Cumming, G. F. (2011). The Sex Offender Treatment Intervention and Progress Scale (SOTIPS) Manual. Middlebury, VT: Author. McGrath, R. J., Lasher, M. P., & Cumming, G. F. (2011). A model of static and dynamic risk assessment. Final grant report to the National Institute of Justice. Award Number 2008-DD-BX McGrath, R. J., Lasher, M. P., & Cumming, G. F. (2011). The Sex Offender Treatment Intervention and Progress Scale (SOTIPS): Psychometric properties and incremental predictive validity with the Static-99R. Manuscript under review. SOTIPS & Vermont DOC Treatment Programs (2001 to 2011) Prison Programs High Intensity 24 month program Moderate Intensity 12 month program Low Intensity 6 month program Community Programs Probation, parole, or furlough 24 month program St. Albans Burlington Morrisville Middlebury Randolph Rutland Bennington Montpelier Newport St. Johnsbury White River Springfield Brattleboro 4

5 Study Design 3. How does the SOTIPS perform? Start Community VTPSA Treatment (N = 759) Follow-up Period Months 3 Years 1 Year Time 1 Time 2 Time 3 Time 1 Time 2 Recidivism (new charges) 1. Sexual 2. Violent (sex and violent) 3. Any criminal 4. Return to prison Time 3 Participants adult male sex offenders Placed in community in VT 2001 to 2007 Age at release 34.2 Static-99R score 2.5 White 96.4% Education, 12 or more years 73.3% Community release status Probation 80.2% Parole or furlough 19.8% Offender type Non-contact 7.8% Extra-familial child molester 59.6% Incest 14.6% Rapist 18.1% Reliability (320 cases scored by 17 treatment providers & 24 POs) ICC 1 ICC 2 Total score F1 Sexual deviance F2 Criminality F3 Social stability and supports Standard error of measurement (68% CI) Cronbach s alpha.87 Gutman split-half reliability.87 Sexual Recidivists vs. Non-recidivists SOTIPS score 32 vs vs vs. 531 Area under the curve of the receiver operating characteristic Probability that a randomly selected recidivist will have a higher score than will a randomly selected nonrecidivist. AUC values range from 0 to 1, with.5 representing chance-level prediction and 1 representing perfect prediction. Predictive Validity AUC Statistic 5

6 Sexual Recidivism AUC s for 3-Year Follow-up Sexual Recidivism AUC s for 3-Year Follow-up Months in treatment Combined Static-99R and SOTIPS AUCs at 1- and 3-year Follow-up Type of recidivism Time 1 Time 2 Time Sexual.74** /.70***.89*** /.76***.73* /.72*** Violent.75*** /.70***.78*** /.71***.69** /.69*** Any.67*** /.69***.74*** /.74***.68*** /.70*** Return to prison.77*** /.74***.78*** /.76***.77*** /.74*** Fits predictions to a S-Shaped curve Logistic Regression Can model predictions with multiple scales Growth pattern found in many fields of study Used for predictions in Static Norms GEE Risk Levels and Recidivism Rates Generalized Estimating Equations Sexual Recidivism Rates at 3 Years by Combined Static-99R and SOTIPS GEE Risk Levels Models correlated data in repeated or multi-level designs Provided a single regression model for the SOTIPS at Times 1, 2 and 3 combined with the Static-99R For recidivism risk, used a binary logistic model For survival, used an interval censored survival model Time 1 Time 2 SOTIPS Risk Level Static-99R Risk Level Time 3 Low Moderate-low Moderate-high High 6

7 Static-99R Relative Risk Levels Low (0 to 10) SOTIPS Moderate (11 to 20) Low (-3 to 1) low low Moderate-Low (2 to 3) low Moderate-High (4 to 5) High (6 to 12 High (21 to 48) high moderatelow moderatelow moderatehigh moderatelow moderatehigh moderatehigh high high Sexual Recidivism Rates GEE Risk Levels 1- and 3-Year Follow-up Periods Low Moderate-low Moderate-high High Percent of Offenders in Each Static99-R and SOTIPS Combined Risk Group How does this translate into practice? Herbert Static-99R = Mod-Low Early 60 s Stranger Male Victims Prior non-contact SOTIPS = High Poor risk management Attitudes and Interest supportive of offending Poor self-regulation Mel Static-99R = Mod-Low Early 30 s One known female victim No long-term relationship SOTIPS = Low Stable lifestyle Cooperates w/ supervision and treatment Manages risk well Overall Relative Risk Moderate-high Overall Relative Risk Low Limitations Construction sample may exaggerate the SOTIPS and recidivism relationship Participants characteristics predominantly white and from a rural state Sexual recidivism base rate 5% over three years Sub-group analyses sex recidivism not predicted for adult-victim-only (20%) Application Need for replication and local norms 4. How do I score the SOTIPS? 7

8 Who Scores the SOTIPS? Single Scorers Clinician Probation/parole officer Correctional caseworker Multiple scorers Two co-therapists Therapist and probation officer Therapist and offender during group Qualifications Expertise in the area of sex offenders Understand principles of assessment When do I score the SOTIPS? Initial during first month of treatment/supervision During treatment/supervision as often as every 6 months End of treatment/supervision What information can I use to score the SOTIPS? Interview Observation Offender self-reports Collateral reports probation and parole officers family and support persons Polygraph Phallometric and VTM testing Anchors for scoring SOTIPS Items Scoring Window: 6-month window or current status 0 minimal or no need for improvement 1 some need for improvement 2 considerable need for improvement 3 very considerable need for improvement How is it Scored? Using the SOTIPS Sliders: If unsure how to score some items (0 or 1; 1 or 2, 2 or 3),score about half higher and half lower Individual items Structured method of identifying and describing specific intervention targets and progress Total Score Method of identifying overall level of treatment and supervision need Combines with static risk measures to improve predictive validity Absent local norms, focus on relative risk 8

9 1. Sex Offense Responsibility See SOTIPS Scoring Manual 1. I might have done it, but I don t remember. I was in an alcoholic blackout. 2. I knew it was wrong but I went ahead and did it anyway. I know I have trouble controlling myself. 3. It was 100% my fault because I am the adult. Even though he physically forced victim, he says she was cooperative and she enjoy it as much as I did. She never said no. She used to hug me in a sexy way. I never would have forced her to do anything. 4. It was 60/40 - mostly my fault. I knew her age was on the edge of being legal, but she told me she was 17. I had no idea she was only Sexual Behavior 1. He is suspected by staff of forcing his wife to have sex. No other recent behavior problems. 2. He reports masturbating 2-3 times a day. 3. Staff found three adult oriented XXX DVD s at his home and possession of pornography is against his probation conditions. This is his first violation for this type of behavior 4. He was caught with child pornography on his computer. 3. Sexual Attitudes 1. A year ago, he picked up an intoxicated woman in a bar and raped her. He has avoided going to bars, but reports some struggles managing thoughts about going to bars to pick up women. 2. Seeing children on mainstream T.V. shows sometimes triggers fleeting sexual thoughts. He said he catches himself and successfully uses self-talk interventions to avoid or extricate himself from these situations. 3. He says there are two types of women good girls and whores. Whores are just out to use men. They deserve what they get (e.g. abused). 4. He is able to describe how he challenges thoughts that supported his sex offending. However, he says he has lapsed 6 times over the last 6 months and masturbated to orgasm while thinking about children. 4. Sexual Interests 1. He molested two unrelated 10-year-old boys. He has had only three sexual experiences with adult females all unsatisfying. He says his sexual interests are now exclusively towards adult females. 2. He molested three of his daughters when they were between the ages of 13 and 15. He had an active sex life with his wife for several years. On PPG: arousal is adult females = 65%, adolescent females = 40%; all other stimuli less than 20%. 3. He has a history of multiple short and long-term sexual relationships with adult females and one conviction for a coercive date rape type offense. He said all his sexual interests concern consensual sexual activity. 5. Sexual Risk Management 1. Seeing children on mainstream T.V. shows sometimes triggers fleeting sexual thoughts. He said he catches himself and successfully uses self-talk interventions to avoid or extricate himself from these situations. His wife confirms this. 2. He is able to describe his risk factors and appropriate intervention strategies in detail. However, during his recent compliance polygraph exam, he admitted to using sexual fantasies of minor females a few times a week when he masturbates. 3. He has lived with his parents for a few weeks and is frightened to tell his mom about his sex offending risk factors, even though he knows it s a good idea and is a probation requirement. He is willing to do so. 9

10 6. Criminal and Rule-Breaking Behavior 7. Criminal and Rule-Breaking Attitudes 1. He justifies getting two traffic tickets for speeding in the last six months because he was going to be late for work. 2. He is compliant with probation and treatment rules. 3. He was charged with domestic assault against his wife about 10 months ago. No problems evident since that time 4. He failed an breathalyzer test during a home visit by his probation officer (.03 BAC). 1. He does not have a non-sexual criminal history and he follows program and probation rules. 2. He often complains that his probation conditions are unfair so he should not have to follow them. He consistently supports the antisocial statements of other group members. He sees nothing wrong with and is not open to examining any of these views. 3. Treatment notes and his home work indicate that he is able to recognize and describe how he challenges the thoughts that support his rule-breaking behavior. However, he was charged with and justifies driving without a license twice in the last six months. 8. Stage of Change 1. He has has been very stable for over 12 months. He recognizes and manages his risks and has developed a healthy social support system. His new lifestyle is incompatible with sex offending. 2. He has been actively working the program for 9 months, has no disciplinary problems, and is getting close to being moved into a monthly aftercare group. 3. He vacillates between admitting he has a sexual offending problem for which he needs treatment and saying that if he simply stays away from the wrong people he will never reoffend. 4. He said he does not have a problem with sexual offending and will never reoffend, but is willing to do the program if that is what is required to get out. 9. Cooperation with Treatment 1. He started the program a few weeks ago and says he is too anxious to talk in group. Staff are attempting to work with him at being more comfortable participating. 2. He always seems to have an excuse about why he is late for group. He is late by about about one-third of the time. He has been given a written warning indicating concerns about this behavior. 3. During his initial evaluation meetings, he was very cooperative with the interviewer. However, he denies some aspects of the behaviors for which he was convicted. 4. He is prepared for group with his written home work assignments about 85% of the time. 10. Cooperation with Supervision 1. His probation officer has asked him to come in for meetings more often because of concerns about his abusing alcohol. 2. He misses about one out of every 10 of his probation meetings but is cooperative following his other conditions. 3. He failed to follow his release conditions about three months ago while he was in the community awaiting sentencing. 4. He was returned to prison on a revocation of his conditional release. 11. Emotion Management 1. Anger was his predominant precursive emotional state when he sexually offended. He is sensitive to criticism. He often lashes out verbally in anger when he perceives he is being put down. He is typically able to talk himself down only after staff intervene. 2. Seeking emotional closeness was a motivating factor in his sexual offending. When under stress he will sometimes isolate himself in his apartment but he has done this less during the last 6 months. He increasingly will seek out friends from AA to talk with during stressful times in his life. 10

11 12. Problem Solving 13. Impulsivity 1. He has trouble learning from his past mistakes and repeats the same problems over and over. He has difficulty taking feedback from others and examining how he could have handled situations differently. 2. He sets goals for himself, follows through, and generally achieves his goals. 3. He occasionally gets flustered when his normal scheduled of activities is changed, such as when his probation appointment is cancelled. But, he can talk through his concerns with staff and adjust relatively quickly. 1. He often can t sit still in group and frequently blurts out irrelevant comments. Therapists have moderate success redirecting him in group. 2. He is very self controlled, rigid, and deliberate in his actions. 3. He has been diagnosed as having adult ADHD but takes medication and functions with only infrequent problems. 4. He often spends money on things irresponsibly and then regrets the resulting problems he has with his credit cards. He also recently said he panicked when a police officer tried to pull him over. He tried to outrun in the officer and was later apprehended. 14. Employment 15. Residence 1. He has changed jobs three times in the last 6 months.. 2. He refuses to work. He says the jobs that are open, such as cleaning toilets or sweeping floors, are beneath him. 3. He is a full-time college student and works part time to help pay his bills. He enjoys and does well at school and is making ends meet. 4. He is retired and spends his time working around his house and with friends who are a prosocial influence. 1. He has no permanent residence. He crashes at several friends apartments for a few weeks at a time. 2. He has lived with his wife in the same apartment for 3 years. 3. He lives in an apartment in a high crime neighborhood, has had lots of problems with neighbors, and wants to move, but cannot afford to do so. 16. Social Influences 1. He lives with his parents but has refused to tell them about his sexual offending history and probation conditions. They continue to believe he is innocent. The friends with whom he associates are not involved in substance abuse or any criminal activity. 2. He hangs out exclusively with friends who have extensive criminal histories and continue to get into trouble with the law. 5. How can the SOTIPS be used in practice? 3. He and his wife have lots of friends, all of whom are prosocial. His wife has been in a significant others group, knows about his offending patterns, and is judged to be a good support person by staff. 11

12 Vermont Prison Programs Treatment Providers Assessment Schedule Intake Mid-treatment End of Treatment Vermont Community Programs Treatment Providers Assessment Schedule Intake January July January July SOTIPS SOTIPS SOTIPS Static-99R scored in group VASOR-R with client Before release, providers give copies to: Probation and parole officer Community treatment provider QI and research staff We use results for treatment and release planning. SOTIPS SOTIPS SOTIPS SOTIPS SOTIPS Static-99R VASOR-R Providers give copies to: Probation and parole officer QI and research staff We use results for treatment, supervision, and case planning. Summary Best practice involves assessing risk and needs Integrated static and dynamic assessment schemes show considerable promise In the Present Study SOTIPS added incremental predictive accuracy to the Static-99R and VASOR-R In practice, we have used the SOTIPS to: Provide a structure for assessing risk and needs Inform initial treatment and supervision placement decisions Reassess and recalibrate services at regular intervals Measure treatment progress Inform release and treatment completion decisions Robert McGrath rmcgrath@sover.net 12