1

The Transportability of MST to Sweden: A two year follow-up of a Randomized Controlled

Trial of Conduct Disordered Youth

Cecilia Andrée LöfholmKnut SundellKjell HanssonTina Olsson

Halmstad

Social welfare states (low rate of unemployment, poverty, drug use, violence…)

Perceived lack of effectiveness in interventions for young people with behaviour problems

Increased interest in prevention and in evidence-based programs

Researchers and professionals turned to the USA for inspiration

The Scandinavian context

2

Multisystemic Multisystemic therapytherapy

Adolescents with serious clinical problemsAdolescents with serious clinical problems

Intensive familyIntensive family-- and communityand community--based based treatmenttreatment

Services dependent on the family’s needs

Services available 24 / 7

Treatment 4-6 months

Licensed by MST services Inc.Licensed by MST services Inc.Weekly expert consultations

Treatment adherence measured

THE STUDY

Started in 2004Started in 2004

MST MST vsvs TreatmentTreatment--AsAs--Usual (50/50)Usual (50/50)

Randomized controlled trial Randomized controlled trial

27 local authorities / 6 MST27 local authorities / 6 MST--teamsteams

IntentionIntention--toto--treat (randomization intact)treat (randomization intact)

Explicit inclusion criteria Explicit inclusion criteria

MultiagentMultiagent / / multimethodmultimethod assessment assessment

Study of Study of ””real lifereal life”” (effectiveness)(effectiveness)

No economical or other interests in MSTNo economical or other interests in MST

3

Participants

Conduct disordered youthConduct disordered youth

12 12 –– 17 years17 years

Exclusion criteriaExclusion criteria::

No motivated guardianNo motivated guardian

Sexual offending Sexual offending

Autism, acute psychosis, risk of suicide Autism, acute psychosis, risk of suicide

The presence of the youth in the home posed The presence of the youth in the home posed a serious risk to the youth or to the familya serious risk to the youth or to the family

MST

Treatment As Usual

PrePre--measuremeasure

7 7 monthmonth

followfollow--upup

24 24 monthmonth

followfollow--upup60 60 monthmonth

followfollow--upup

RandoRando--

misationmisationRecruitRecruit--

mentment

TAU

TAU

MST

MST

168 156(93%)

149(96%)

147(94%)

4

Child Behavior Checklist / Youth Self-Report (Achenbach, 1991a; 1991b)

Sense of Coherence (Antonovsky, 1987)

Self-Report Delinquency Scale (Elliott et al., 1983)

Alcohol Use Disorder Identification Test (Babor et al., 1992)

Drug Use Disorder Identification Test (Berman et al., 2005)

Pittsburgh Youth Study (Keenan et al., 995)

Social Competence with Peers Questionnaire (Spence, 1995)

Social Skills Ratings System (Gresham & Elliott, 1990)

Symptoms Checklist-90 (Derogatis & Cleary, 1997)

Parenting skills (Stattin,2004)

School attendance

Social services (case file review)

Police reports

Measures

77 randomized to TAU

64 started

- 20 counseling

- 16 family therapy

- 12 mentoring

- 8 out-of-home care

- 4 ART

- 4 Other services

79 randomized to MST

75 started MST

55 completed

77 randomized to TAU

64 started

48 completed

95%70% 86%62%

5

0

0,5

1

1,5

2

0 7 24

MST

TAU

Antisocial behaviors (SRD)

Arrested during the last 6 months(parent reports)

0

20

40

60

80

100

0 7 24

%

MST

TAU

6

0

2

4

6

8

10

0 7 24

MST

TAU

0

2

4

6

8

10

0 7 24

Drugs

(Dudit score)Alcohol

(Audit score)

Substance abuse (change scores)

0

20

40

60

80

0 7 24

MST

Trad

0

20

40

60

80

0 7 24

Self-reported

(YSR)

Parents

(CBCL)

Mental Health Problems

7

0

0,5

1

1,5

2

0 7 24

MST

Trad

Mothers’ mental health (SCL-90)

Days in out-of-home care (Cumulative %)

0

20

40

60

80

100

0 7 24

%

MST

TAU

8

Days with services (Cumulative)

0

100

200

300

400

500

600

700

0 7 24

days

MST

TAU

Average Costs (SEK) per Youth

0

100 000

200 000

300 000

400 000

500 000

600 000

MST TAU

MST

Other

Placement

MST Break even = annual work load of 40 families

(instead of 27)

9

IN SUM

Youth decreased their problem behavior and improved social skills and family relations

None of these improvements were statistically different between the groups

This outcome is contrary to results from the U.S. / Norway, but similar to one from Canada

WHY DID NOT MST OUTPERFORM TAU?

1 Different populationYouth symptomatology in Sweden was high

2 Swedish MST less potent than in the USAPoor fidelity

MST program standards were compromised (e.g., low average

caseloads, use of interpreter)

3 Swedish TAU more potent than TAU in the USAFew iatrogenic interventions (e.g., residential care)

Proactive family service orientation

4 Sociodemografic context supports SwedenLow prevalence of illicit drug consumption

Low prevalence of delinquency

Low rate of poverty, teenage pregnancy et cetera

10

MST treatment fidelity

Therapist Adherence MeasureTherapist Adherence Measure

(Parents interviewed monthly about the therapy)

Not related to outcome measures

SWSW USUS

4,0(.61)

4,4

(.49)

Low High

Mental health (CBCL) change T-scores in evaluations of MST (6-12 month follow-up)

0

4

8

12

Sundell et al

(in press)

Ogden &

Halliday-

Boykins

(2004)

Henggeler et

al (2006)

Henggeler et

al (2006) -

cont man

Rowland et al

(2005) -

external

Rowland et al

(2005) -

internal

MST TAU

11

Cannabis Cannabis useuse ((lifelife--timetime) )

boys boys agedaged 15 15 (ter (ter BogtBogt et al, 2006)et al, 2006)

0

10

20

30

40

50

60

70

80

Schwizerland

Canada

Greenland

USA

UK

Spain

Tjeckia

France

Ukraine

Slovenia

Belgium

Holland

Germany

Irland

Italy

Danm

ark

Portugal

Polen

Estland

Russia

Croatia

Hungary

Lettland

Austria

Lithuania

Finland

Malta

Israel

Greece

Sweden

Makedonia

%

Cannabis Cannabis useuse ((lifelife--timetime) ) girlsgirls

agedaged 15 15 (ter (ter BogtBogt et al, 2006)et al, 2006)

0

10

20

30

40

50

60

70

80

Greenland

Canada

Schwitzerland UK

Spain

USA

Tjeckia

France

Slovenia

Holland

Belgium

Danm

ark

Germany

Italy

Ukraine

Irland

Portugal

Croatia

Austria

Polen

Estlonia

Hungary

Finland

Russia

Lethuania

Sweden

Litauen

Israel

Malta

Greece

Makedonia

%

THE TRANSPORTABILITY OF EVIDENCE-

BASED INTERVENTIONS TO SCANDINAVIA

More effective as TAU

The Incredible Years

Functional family therapy

Multi-dimensional treatment foster care

Parent management training (PMTO)

Multisystemic Therapy (Norway)

Marte meo

Equally effective as TAU

MST (Sweden)

Strengthening families programme

Unplugged (school-based drug education program)

Transportability of US Evidence Based Transportability of US Evidence Based Programs for Reducing HomelessnessPrograms for Reducing Homelessness

Campbell Collaboration Colloquium, Oslo, May 18-20 2009,

Session 2 Tuesday 13:15-14:45

Sten Anttila

Institute for Evidence Based Social Work Practice (IMS)

National Board of Health and Welfare, Sweden

sten.anttila@socialstyrelsen.se

Interpretation of systematic reviews and Interpretation of systematic reviews and development of national guidelinesdevelopment of national guidelines

The problem

Prediction: sampling

Implementation: intervention and local context

Possible strategies

Effect size: control group and estimating effects of local usual care

Content: intervention and comparing with models in local practice

Predicting local effects Predicting local effects –– a sampling a sampling problemproblem

RCT subjects – not a random sample

Representative sample – USA, Sweden etc?

Meta-analysis subjects – not a random sample

Representative sample – USA, Sweden etc?

Treatment group

Possibly representative with manuals?

Control group

Probably not representative as usual care?

Implementation and delivery Implementation and delivery –– a mismatch a mismatch problemproblem

Intervention

Critical ingredients

proportion and intensity

sequence and duration

Mismatch with local practice

Practitioners expertise:

knowledge, skills …

Client preferences and actions:

values, options …

Clinical state and circumstances:

resources, traditions, routines, policy, legislation …

EBP model

Focus behavior of control groups and Focus behavior of control groups and intervention groupsintervention groups

Proportion of past time in stable housing - changes from baseline

unweigted means from 4-5 RCT:s in USA

0,00

0,10

0,20

0,30

0,40

0,50

0,60

Baseline 6 months 12 months 18 months

Mean %

of

tim

e in s

table

housin

g

Intervention 1 mean

improvement

Intervention 2 mean

improvement

Usual care mean

improvement

Total mean improvement

Similar development

among homeless people

in Swedish UC

compared to US UC

supports transferability

of results.

Study register based

cohorts.

Housing movements in USA and Sweden: Housing movements in USA and Sweden: different time frame, approximately similar different time frame, approximately similar categoriescategories

NYC, USA, n=46

0,15

0,22

0,48

0,15

0,00

0,10

0,20

0,30

0,40

0,50

0,60

Homeless Marginal

housing

Institution Stable

housing

Perc

ent of pers

ons 3

years

5 cities, Sweden, n=210

0,20

0,04

0,14

0,53

0,00

0,10

0,20

0,30

0,40

0,50

0,60

Homeless Marginal

housing

Institution Stable

housing

Perc

ent of pers

ons 5

years

Housing movements in USA and Sweden: Housing movements in USA and Sweden: similar time frame, unknown similarity of similar time frame, unknown similarity of categoriescategories

Housed - Housing First & UC in NYC (Tsemberis & Eisenberg 2000)

and UC in Sweden, 5 cities (Gerner & Blid 2002)

0

0,88

0

0,47

0

0,71

0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

1

baseline 5 year follow-up

Pro

port

ion o

f housed p

ers

ons

USA, Pathw ays to housing

in NYC

USA, UC in NYC

Sw eden, UC in 5 cities

If Swedish UC is below

US intervention and

close to US UC,

transferability of

improved results with

new intervention is

supported.

Example of intervention: Critical Time Example of intervention: Critical Time Intervention (CTI)Intervention (CTI)

Goal

- to prevent recurrent homelessness and other adverse outcomes

- to enhance continuity of care during transition from institutional to community living

Components

- strengthening the individual's long-term ties to services, family, friends

- providing emotional and practical support during transition

Skilled clinicians or case managers with

- knowledge about substance abuse, motivational interviewing, assessment of symptom severity, and system resources.

- ongoing supervision and consultation, training and support services, clear evaluation tools to support quality assurance

Critical Time Intervention (CTI): Critical Time Intervention (CTI): ingredientsingredients

Implementation Material: Model description, fidelity scale, manual, (ex of treatment plans, ex of activity logs)

Resources Caseload: 10-15 clients per case manager

1-3 months Accommodation: home visits, accompany patients to appointments, meet with care givers, substitute care givers when necessary, give support and advice to patient and caregiver, mediate conflicts between patient and caregiver, help negotiate ground rules for relationships

4-7 months Tryout: observe trial of ground rules, help negotiate ground rules as necessary

8-9 months Termination: reaffirm ground rules, hold parties and meetings to symbolize transfer of care

Transferability and implementationTransferability and implementation

Present local practice and intervention

approximately similar goals

include most critical ingredients

follow basic sequence

Local implementation of intervention

is possible

Model description, manual and fidelity scale

support clinicians and protect pratice during budget process

may improve outcome

DiscussionDiscussion

Prediction

Interpreting effect sizes

assess similarity of control group and control group behavior

Implementation

Identifying possible mismatch problems

assess similarity of local practice with intervention