Need for change…
Cabinet Office, 2006; Department for
Education and Skills, 2007: Need for
change in approach to At Risk Teens….
The Effectiveness of Multi Bailey & Scott, 2008: high quality ‘model’
Systemic Therapy (MST) services must be nationally implemented.
preventive community services
high quality interventions
Peter Fonagy good effect sizes
P.Fonagy@UCL.AC.UK better trained and highly motivated and
Research Department of Clinical, skilled teams.
Educational & Health Psychology,
University College London
MST: Potential Antisocial Problems in Adolescence
Government commitment: tackle social Serious / repeated antisocial behaviour
exclusion by tertiary prevention realtively common during childhood and
Reduce out-of-home placement for high risk adolescence
youth Significant, costly long-term consequences
Review of ASPD interventions (Utting, Monteiro, (Farrington, 1995).
& Ghate, 2007)
Lifetime prevalence of conduct disorder
DCSF, YJB and the Cabinet Office: (CD) in the UK and US approx 10%
MST is a promising intervention for reducing
the risk of antisocial behaviour, including 12% among males
substance misuse, 7.1% among females
(Maughan, Rowe, Messer, Goodman, &
offending and Meltzer, 2004; Nock, Kazdin, Hiripi, & Kessler,
conviction rates 2006, 2007)
time spent in custodial institutions.
Conduct Disorder Antisocial Personality Disorder
2/3 of those with diagnosable CD have ASPD linked to CD by definition.
severe problems: Onset in childhood (under 10 years) is especially
29% pervasive CD, average of 8 symptoms bad and associated with:
including aggression polysymptomatic and violent adult ASPD
Axis I GAD, drug dependence
29% endorse on average 6 symptoms i.e.
Axis II paranoid, schizoid and avoidant PD disorders
theft, property oriented offences, BUT not
(Goldstein, Grant, Ruan, Smith, & Saha, 2006).
physical violence
In the UK:
3% primarily aggressive (Nock et al., 2006).
approx. 200,000 offenders aged 10 to 17
(Home Office, 2003)
12% of young men involved with the criminal
justice system by their late teens
(Farrington, 1995; Home Office, 2003).
1
Some Statistics Rehabilitation?
Offences committed by young people Rehabilitative approach to juvenile justice
resulting in a disposal: embedded in Western culture Malleability
Over 300,000 disposals most commonly for ASPD Most prevalent PD in men (1.5% and
robbery, burglary 6.8%)
Females: up 39% from 2002/03 to 2005/06 ASPD precursors diagnosable in childhood:
………on into ASPD in adult life where no
Males: 7% rise (YJB, 2007). treatment appears strikingly effective
Increase in young people in custody in
2005/06 (e.g. Cote, Vaillancourt, LeBlanc, Nagin, & Tremblay, 2006; Fontaine et
o 84% accommodated in young offender institutions al., 2008; Kim-Cohen, Caspi, Moffitt, Harrington, & Milne, 2003;
Loeber, Burke, & Lahey, 2002; Loeber et al., 2005; Moffit, Caspi,
o 8% each in secure children’s homes and secure training Rutter, & Silva, 2001; Perdikouri, Rathbone, Huband, & Duggan,
centres. 2007; Robins, Tipp, & Przybeck, 1991; Simonoff et al., 2004;
Sourander et al., 2006).
Tackling ASPD Needs of Adolescents With Anti-
Social Problems
Childhood and adolescence: pressure (Chitsabesan et al., 2006) between 1/3
from social agencies and/or families for and 1/2 of sample had:
externalising problems in adolescents to Health needs
be tackled. Educational needs
Work needs
Intervention in adolescence appears to be Social relationship needs
the most promising strategy to minimize commonly not recognized by agencies
the financial and psychological costs of
ASPD. But
Narrow focus on forensic outcomes.
Aetiology and treatment Family Risk Factors
Parenting styles (Patterson, 2002; Snyder &
Stoolmiller, 2002)
Low IQpoor understanding of social Intervention appropriate to developmental stage
situationslow empathy (Alexander, Pugh, Parsons, & Sexton, 2000; Myers et
Social skills training al., 2000; Scott, Spender, Doolan, Jacobs, & Aspland,
2001).
Restorative justice
Deviant peer associations /peer rejection
Impulsivity/emotional over- (Gifford-Smith, Dodge, Dishion, & McCord,
reactivityanger management 2005), leads to “peer contagion”.
Parenting styles Parental criminality transmitting values
Excessive use of corporal punishment
Domestic violence (Bell, 1995; Margolin, 1995)
2
MST intellectual basis MST: Methodology
One therapist,
Cumulative risk potentially a range of techniques
Marital and family therapies, Parent training,
Meta-protocl for addressing different kinds
Behavioural and cognitive approaches,
of risk and protective factors supportive therapy, case management (may
3 or more risk factors massively increase involve liaison with outside agencies).
chance of anti-social behaviour risks 9 treatment principles govern delivery
need to be tackled at same time Systemic strengths
Responsible behaviour
Should monitor how and how effectively Targeting sequences of behaviour in multiple systems
MST tackles risks responsible for maintaining behavioural problems
continuous evaluation from multiple perspectives.
MST: Strengths RCT Evidence for MST
Intensive commitment of therapists to a family’s A number of good-quality RCTs suggest MST is the
problems most effective treatment for delinquent adolescents:
Small caseloads and attention to quality control reduces recidivism
Ongoing reporting of outcome by family members improves individual and family pathology
Most severe psychosocial and psychiatric problems
met head-on (Borduin, 1999; Henggeler, Cunningham, Pickrel, Schoenwald, &
Brondino, 1996; Henggeler, Melton, & Smith, 1992; Henggeler,
Clear link between hypothesised pathogenic and Melton, Smith, Schoenwald, & Hanley, 1993; Henggeler et al.,
treatment mechanisms 1986).
Intensive but time-limited therapeutic format More effective than individual treatment even for quite troubled and
disorganised families (Borduin et al., 1995).
Generically but well-trained practitioners Reduces attrition rates (Henggeler, Pickrel, Brondino, & Crouch,
1996).
Favourable Initial Reviews: 4-year follow-up,
(Borduin, 1999; Fonagy, Target, Cottrell, Phillips, & Kurtz, 2002; Roth & recidivism in MST recipients significantly reduced:
Fonagy, 1996).
22.1% compared to 71.4% in recipients of individual therapy
But recent reviews more critical (Littell, 2005; Littell, 2006; Littell, Popa, & Arrests in MST group were for less serious crimes than in the
Forsythe, 2005). individual therapy group (Borduin, 1999)
Borduin et al 1995: 10 year follow-up (Schaeffer & Borduin et al 1995: 10 year follow-up (Schaeffer &
Borduin, 2005). Borduin, 2005): Arrest for violent crime
Survival functions for multisystemic therapy (MST) and
Survival functions for multisystemic therapy
(MST) and individual therapy (IT) groups on individual therapy (IT) groups on time to first violent arrest
time to any first arrest following treatment. following treatment. Completers and dropouts are
Completers and dropouts are combined in combined in each group.
each group.
3
Borduin et al 1995: 10 year follow-up (Schaeffer & Borduin et al 1995: 10 year follow-up (Schaeffer & Borduin, 2005):
Arrest for Drugs
Borduin, 2005): Arrest for non violent crime
Survival functions for multisystemic therapy (MST) and individual therapy (IT)
Survival functions for multisystemic therapy (MST) and groups on time to first drug-related arrest following treatment. Completers and
individual therapy (IT) groups on time to first nonviolent dropouts are combined in each group.
arrest following treatment. Completers and dropouts are
combined in each group.
Ontario Study: (Leschied & Cunningham, 2002) Ontario Study: (Leschied & Cunningham, 2002)
Convictions during follow-up Convictions during follow-up
Survival Curve of Convictions During 3-year Follow-up,
All Sites Combined
MST Group Control Group
SIX MONTHS
At least one conviction 28.5% 33.3%
ONE YEAR
At least one conviction 49.3% 44.0%
TWO YEARS
At least one conviction 62.6% 57.9%
THREE YEARS
At least one conviction 68.2% 66.5%
Systematic Review Incarceration and Conviction
Partial Meta-analyses do not N (MST,
Standard
Effect
Size
95% CI P<
(Overall
incorporate recent investigations Care) Effect)
misleading power calculations? Incarceration/conviction end of 335, 316 0.51 (0.23, 1.16) 0.11
treatment
(Henggeler et al., 1992)
(Henggeler et al., 1997)
we undertook a further systematic (Leschied & Cunningham, 2002)
review Incarceration/conviction 1.7 year
follow-up
82, 73 0.75 (0.52, 1.07) 0.11
Effect sizes obtained from MST (Henggeler et al., 1997)
compared to MAU as a control. Incarceration (days/weeks)
(Henggeler et al., 1992)
325, 298 -0.30 (-0.71,
0.11)
0.15
(Henggeler et al., 1997)
Studies included : (Leschied & Cunningham, 2002)
(Leschied & Cunningham, 2002; Borduin & Schaeffer, 2001; Borduin et al.,
1995; Henggeler, Halliday-Boykins, & Cunningham, 2006; Henggeler,
Melton, Brondino, Scherer, & Hanley, 1997; Henggeler et al., 1992;
Henggeler, Pickrel, & Brondino, 1999; Ogden & Hagen, 2006; Rowland
et al., 2005; Timmons-Mitchell, Bender, & Kishna, 2006).
4
Review Forest Plot: Re-arrests
N (MST, Effect 95% CI P<
Incareration/conviction Standard
Care)
Size (Overall
Effect)
Rearrested 12-18 month follow-up 393, 372 0.70 (0.45, 1.09) 0.12
(Borduin et al., 1995)
(Henggeler et al., 1992)
(Leschied & Cunningham, 2002)
(Timmons-Mitchell et al., 2006)
Rearrested 8-14 year follow-up 116, 105 0.53 (0.31, 0.90) 0.02
(Borduin et al., 1995)
(Borduin & Schaeffer, 2001)
Number of arrests short term 354, 325 -0.39 (-0.81, 0.06
follow-up 0.02)
(Borduin et al., 1995)
(Henggeler et al., 1992)
(Henggeler et al., 1997)
(Henggeler et al., 1999)
(Henggeler et al., 2006)
(Rowland et al., 2005)
(Timmons-Mitchell et al., 2006)
Number of arrests 4 year follow-up 43, 37 -0.33 (-0.77, 0.14
(Henggeler et al., 1999) 0.11)
Re-arrests Number of arrests
CBCL and TRF Externalizing Faces-III, Psychiatric Symptoms
N (MST,
Standard
Effect
Size
95% CI P<
(Overall
Internalizing
Care) Effect)
N (MST, Effect 95% CI P<
CBCL: Parent Reports (2 years 43, 26 -0.51 (-1.01, - 0.04 Standard Size (Overall
after pre-assessment) 0.02) Care) Effect)
CBCL 89-Item Problem Scale
FACES-III Adaptability 61, 35 -0.34 (-0.76, 0.08) 0.11
(Ogden & Halliday-Boykins, 2004)
(Ogden & Halliday-Boykins, 2004)
CBCL Externalizing 43, 26 -0.17 (-0.66, 0.32) 0.49
FACES-III Cohesion 61, 35 -0.08 (-0.49, 0.34) 0.71
(Ogden & Halliday-Boykins, 2004)
(Ogden & Halliday-Boykins, 2004)
CBCL Internalizing 43, 26 -0.69 (-1.19, - 0.007
General Psychiatric Symptoms 185, 152 -0.14 (-0.56, 0.27) 0.50
(Ogden & Halliday-Boykins, 2004) 0.19)
(SCL/BSCL)
CBCL: Teachers Reports (2 years 43, 26 -1.10 (-1.62, - < 0.0001 (Borduin et al., 1995)
after pre-assessment) 0.58) (Henggeler et al., 1992)
TRF 89-Item Problem Scale (Henggeler et al., 1997)
(Ogden & Halliday-Boykins, 2004)
Psychiatric Symptoms: TAS 4 year 43, 37 0.16 (-0.28, 0.60) 0.47
TRF Externalizing 43, 26 -1.09 (-1.61, - < 0.0001 follow-up: Externalizing Scale
(Ogden & Halliday-Boykins, 2004) 0.57) (Henggeler et al., 1999)
TRF Internalizing 43, 26 -1.14 (-1.67, - < 0.0001 Internalizing Scale 43, 37 0.11 (-0.33, 0.55) 0.61
(Ogden & Halliday-Boykins, 2004) 0.62) (Henggeler et al., 1999)
5
Self-reported delinquency Self-reported delinquency
N (MST, Effect 95% CI P<
Standard Size (Overall
Care) Effect)
Self Reported Delinquency end of 227, 214 -0.18 (-0.42, 0.07) 0.15
treatment
(Henggeler et al., 1992)
(Henggeler et al., 1997)
(Henggeler et al., 1999)
(Henggeler et al., 2006)
(Rowland et al., 2005)
Self Reported Delinquency 6 58, 60 0.05 (-0.31, 0.41) 0.77
month follow-up
(Henggeler et al., 1999)
Self Reported Delinquency 2 year 43, 26 -0.26 (-0.75, 0.23) 0.30
follow-up
(Ogden & Halliday-Boykins, 2004)
Self Reported Delinquency 4 year 43, 37 -0.33 (-0.77, 0.11) 0.14
follow-up
(Henggeler et al., 1999)
Missouri Peer Relations Inventory and
Revised Behavior Problem Checklist
Results of Meta-Analysis
N (MST, Effect 95% CI P< Near-significant reductions in the number of arrests
Standard Size (Overall post-treatment, re-arrests and incarceration/conviction
Care) Effect)
MPRI: Peer Bonding 185, 152 -0.38 (-1.16, 0.40) 0.34
(Borduin et al., 1995)
Teacher-rated effect size of reduction of problem
(Henggeler et al., 1992) behaviours is particularly impressive for both
(Henggeler et al., 1997) internalising and externalising behaviours (p = .0001).
MPRI: Maturity 185, 152 0.04 (-0.18, 0.25) 0.75
(Borduin et al., 1995)
(Henggeler et al., 1992)
May influence individual wellbeing as well as behaviour.
(Henggeler et al., 1997)
MPRI: Peer Aggression 173, 149 -0.13 (-0.35, 0.09) 0.24 Other treatments
(Borduin et al., 1995)
(Henggeler et al., 1992)
Re-arrest rates in young women supported in
(Henggeler et al., 1997) multidimensional foster care significantly decrease
Revised Behavior Problem 173, 149 -0.50 (-1.42, 0.42) 0.29
(p =.02),
Checklist (RBPC) Behaviour problems treated with social skills
(Borduin et al., 1995) training (p =.003), or parent training (p =.000001),
(Henggeler et al., 1992)
(Henggeler et al., 1997)
and or functional family therapy (p =.001).
MST Transportability study MST Transportability study
(Schoenwald, et al. 2003, 2008) (Schoenwald, et al. 2003, 2008)
45 sites across 12 states and Canada Primary referral sources
1979 youth and caregivers 44% juvenile justice
Age:14.0 (SD = 2.35) 23% social services
17% mental health agencies
65% male and 60% Caucasian, 20%
African American Referral reasons
47% status offenses
50% resided with their mother and 16%
with both parents and 17% in special living 47% criminal offenses
arrangements (e.g. foster family) 31% substance use problems
30%) school suspensions or expulsions
6
MST Transportability study
Primary Therapists
(Schoenwald, et al. 2003, 2008)
Caregiver age: 40.8 years (SD = 8.48) 429 Primary therapists (the therapist
Caregivers’ education: treating the family for the entire treatment
66% completed high school episode or the majority of the episode)
34% some years of college (33.9%) 75% of therapists were female
Income: Education of therapists
50% less then $20k 61% master’s degrees
Demographically similar to samples in 32% a bachelor’s degree
randomized trials of MST 3% doctoral degree
MST Transportability study MST Transportability study
(Schoenwald, et al. 2008) (Schoenwald, et al. 2008)
Psychological Climate Questionnaire Variance accounted for
Fairness, rolefairness, role clarity, role overload, role Youth/family level: 43%
conflict, cooperation, growth and advancement, job
satisfaction, emotional exhaustion, personal Therapist level: 4%
accomplishment, and depersonalization Therapist adherence: 1-2%
Organisational Structure Organization level: 5%
Participation in Decision-making (8 items; from Hage
and Aiken 1967);
Variance in adherence
Hierarchy of Authority (4 items; from Hall,1963); Provider organization: 3%
Procedural and Rule Specification (3 items from Hall Therapist: 5%
1963) Residual variance: 92%
Outcome: CBCL-Ext.
MST Transportability study
(Schoenwald, et al. 2008)
Climate predictors of outcome
Growth and advancement (p<.000)
Structure and outcome
Participation in decision making (p<.04)
Relative perception of participation (compared
to others in same organisation (p <.008)
Hierarchical considerations not major
determinants of authority (p<.001)
7
Children in residential care
60,000 young people in residential care in UK
62% are aged between 11 and 17 (DCSF,
2006).
Recent Matching Needs and Services (MNS)
audit (London Borough of Islington, 2006) of
children in care
2/3 had a record of offending behaviour or
severe behaviour difficulties at home
79% have emotional and mental health
difficulties.
Only 28% have their needs addressed fully
by out of home care.
Massive psychological risks associated with
residential care (Meltzer et al., 2003)
MST and Out of Home Placement MST and Out of Home Placement
MST trials looking at family preservation
and the avoidance of psychiatric hospital The reduction in end of treatment long
admission have yielded extremely strong term hospitalisation or incarceration
effects (Henggeler et al., 1999) estimated on the basis of 3 US studies
RR(random) =0.51 (95% CI .94 to .19; p carries:
=.002).
RR of .34 (95% CI: .20, .56) at treatment
Norwegian study:
termination
relative risk of out of home placements associated
with MST 2 years after randomisation was .59 (95% and .75 (95% CI: .52, 1.07) at 18 months
CI: .32, 1.06) follow-up.
almost half (48%) of children in the MAU condition
being taken into care.
The MAU incarceration rate was about 50%.
General Conclusions Will MST be effective in the UK?
There is reasonable evidence to Equipoise
suggest that MST is better than MAU, US, Canadian and Norwegian trials
Inconsistent findings
No treatments that we have identified in
MST superior to MAU in some trials / some measures
our systematic review are better than …. less so in others.
MST. UK: far broader group of young people than US
However, questions remain about the No research on young people currently receiving
different services via different agencies
applicability of MST to the particular set
of complex clinical problems which Ongoing Brandon Centre trial with Camden and
Haringey Youth Offending Teams; limitations in terms of
characterise the UK context. Statistical power
Exclusive focus on chronic delinquency.
8
Issues for investigation Issues for investigation
(4) Do the 9 MST principles discriminate MST
(1) Inflation of effect sizes owing to lack of intent- from standard care,
to-treat analyses
Many principles shared by UK MAU
(2) Developer’s involvement and commitment to
(5) Does the adherence measure (TAM) actually
the success of the intervention in efficacy trials
evaluates adherence to MST principles,
transporting treatments to real-world that
as opposed to measuring simply good
require such intensive supervision;
therapeutic practice
(3) Developer involvement in ensuring adherence
(6) Significance of fidelity to manual
to the 9 MST treatment principles,
Is this related to outcome?
Hence effect size difference between efficacy
(d=.81) and effectiveness studies (d=.26); Therapist fidelity limited by extent to which a
family “permits” a therapist to stick to
manual;
Issues for further Investigation Further Issues of Transportability
(7) Limited evidence for transportability
One major trial (Ontario study) shows no treatment Application in the UK context is broader then
effects
Different approaches to youth justice, i.e. no the targeted sample of previous trials
convictions under 15 in Norway In the UK, the following issues require
systematic monitoring:
(8) Generalisability issues:
large differences between comparison groups across (a) clinical competences of therapists to practice MST:
studies curriculum differences for psychologists and social
level of organisation of MAU services workers in the UK and the US;
(b) stronger evidence base of services as usual in the
(9) High arrest rates at 18 months follow-up, even UK than those used in earlier clinical trials (e.g.
in successful trials such as the Hawaiian RCT individual psychotherapy)
MST 66.7% and MAU 86.7% (c) MST therapists who were not involved in its
although highly significant drop in the likelihood of re- development
arrest in wityh MST.
Further Issues of Transportability Final Conclusion
Major concerns
(d) contextual issues
US results may not transfer
differences in national standards with regard to
sentencing policy and practice; Out of home placement may be the best
Any major UK trial must includes as much detail outcome in some instances (i.e dangerous
as possible about the types of probationary home environment), but…..
and other services made in MAU. MST is the most likely intervention which,
(e) the need for person centred statistical if applied at an appropriate time, could
analyses empower families to provide adequate
to identify typical trajectories of response (e.g. care for young people at risk of
callous and unemotional traits).
incarceration, hospitalisation or other
forms of out of home placements.
9