Does Multisystemic Therapy Change Criminogenic Risk Factors? A 10-Year Study Among Norwegian Youths With and Without Offenses

Abstract

The study investigated outcomes of Multisystemic Therapy (MST), for youths with and without recorded offenses. Over 10 years, 4,103 young people in Norway completed MST, and 5.39% of them had a criminal record. Those with previous offenses had higher scores on most dynamic risk factors. This high-risk group is predominantly boys, below the age of criminal responsibility (15 years) and have substance abuse issues. We observed reductions in risk factors from the beginning to the end of MST for both groups, with the group with offenses showing the most significant improvements. Committing a criminal offense at a young age is strongly associated with later recidivism. This study suggests that MST could disrupt cycles of offending by addressing criminological risk factors.

Keywords

recidivism criminogenic risk factors routine outcome monitoring data YLS/CMI juvenile delinquency

Introduction

In Norway, young people aged 15 to 24 accounted for 29% of all criminal charges (Ministry of Justice and Public Security, 2020–2021). Since 2016, registered offenses mainly attributable to violence, physical abuse, and sexual offenses committed by youths under the age of 18 years have increased in Norway (Ministry of Justice and Public Security, 2020–2021). However, punitive measures against young criminal acts are rare in Norway and other Nordic countries, where juveniles with offenses are often treated by Child Welfare Services (CWS).

Committing a criminal offense at a young age is strongly related to recidivism later in life (Bonta & Andrews, 2018; Frase & Roberts, 2019). This pattern also seems to apply to Norwegian youths (Bhuller & Røgeberg, 2022; Ministry of Justice and Public Security, 2020–2021), with about half of those aged 15 to 17 at the time of initial offense being recharged within 5 years (Bhuller & Røgeberg, 2022). Criminal behaviors among youths may turn into more serious behaviors that are difficult to change. Thus, evidence-based interventions in settings where delinquency is treated by the CWS rather than in the criminal justice system are necessary to prevent and stop criminal pathways for youths and their long-term involvement with the criminal justice system (Aase et al., 2020). However, research on treatment for youths with an offense on record in a non-juvenile justice system remains scarce (McCart et al., 2023).

The United Nations and Council of Europe (2021) recommend home-based care interventions. One of the treatments offered to juveniles with offenses is Multisystemic Therapy (MST), which is rooted in Bronfenbrenner’s socio-ecological model. his intervention is designed to change dynamic risk factors in the youth’s setting to prevent juvenile criminal recidivism. Traditionally, most young people with serious behavioral problems in Norway have been placed in residential care institutions. However, in 1999, the government introduced a 5-year plan to tackle youth crime (Ministry of Children and Families, 1999). As part of this plan, MST was rolled out within the Norwegian CWS as an alternative to residential placement. Cultural adaptation of the materials to the Norwegian context was approved by the U.S. developer organization. Currently, 24 operational Norwegian MST teams have been established within the state/regional, specialized CWS and youths are referred from the municipal CWS. An MST team consists of a supervisor and three to four therapists who are accessible 24 hr a day, 7 days a week. The course of treatment typically lasts 3 to 5 months. Every year, about 600 Norwegian families receive MST and more than 12,000 youths and families have been admitted to the treatment.

Criminogenic Risk Factors

According to the Responsivity, Needs, and Risk Model, criminogenic risk factors are evidence-based characteristics or problems that are found to directly relate to recidivism later in life (Bonta & Andrews, 2018). The Youth Level of Service/Case Management Inventory (YLS/CMI) is a risk/need assessment tool designed specifically for juveniles with offenses that is based on the principles of the Responsivity, Needs, and Risk Model, and includes the “central eight” criminogenic risk factors. Risk level is determined by assessing both static and dynamic risk factors that are empirically linked to recidivism.

Static factors are features of a person’s background and history (e.g., age, gender, and current/prior offenses), which have been found to relate to future criminal offenses (Casey, 2016; Clarke et al., 2017). Previous offense history is considered a strong static predictor of juvenile criminal recidivism (Baglivio & Jackowski, 2013; Cottle et al., 2001; Cuervo & Villanueva, 2015; Savage et al., 2013) and has consistently been included as a static risk factor throughout the evolution of risk assessment tools (Miller et al., 2021). Dynamic risk factors (or criminogenic needs) are predictors of criminal behavior that can be modified and changed, and provide incremental predictive values over static risk factors (Clarke et al., 2017; van den Berg et al., 2018). These encompass antisocial personality pattern, pro-criminal attitudes, antisocial associates, problematic family/marital relationships, issues with school/work, lack of pro-social recreational activities, and substance abuse (Heffernan et al., 2019). Reductions in dynamic risk factors are associated with reduced recidivism when addressed through appropriate interventions (Hanson et al., 2009).

Two recent Norwegian studies have evaluated the YLS/CMI and its usefulness among youths receiving MST (Hukkelberg et al., 2023) and living in residential care (Kanestrøm et al., 2024a). The authors concluded that the YLS/CMI is a suitable tool for predicting the likelihood of criminal behavior among at-risk youths in Norway (Hukkelberg, et al., 2023; Kanestrøm et al., 2024a). The instrument appears to capture relevant dynamic risk factors in the youth environment, aligning with the Responsivity, Needs, and Risk Model, which posits that accumulated risk factors increase the likelihood of continued delinquency.

Multisystemic Therapy

MST is a family-based treatment program developed for adolescents between 12 and 18 years of age who are at risk for out-of-home placement due to serious antisocial behavior. The program is based on Bronfenbrenner’s (1979) social-ecological theory and is designed to address empirically validated dynamic risk factors while building protective factors. In MST, the family is the most important arena for intervention, and dynamic risk factors are actively addressed in the work with the family. Dysfunctional parenting is an important dynamic risk factor for juvenile delinquency, and it has been theorized that appropriate interventions targeting caregivers will reduce other dynamic risk factors, such as the influence of antisocial peers and collaboration with teachers and other community resources. Thus, MST is designed to both directly and indirectly address criminogenic dynamic risk factors for recidivism or serious delinquency. The overall treatment goal is to reduce the youth’s problem behavior.

Juvenile delinquency is a major indicator for inclusion in MST, and preventing recidivism is an important treatment target. Several studies have found both short- and long-term reductions in rearrests and convictions after MST treatment (e.g., Borduin et al., 1995; Butler et al., 2011; Henggeler et al., 1992, 1993; Schaeffer & Borduin, 2005; Timmons-Mitchell et al., 2006). A follow-up of a randomized clinical trial found a decrease of 36% in felony arrests, 75% in violent felony arrests, and 33% in days in adult confinement for the MST group 21.9 years after treatment compared to individual therapy (Sawyer & Borduin, 2011). However, Weiss et al. (2013) found no significant reductions in arrests after MST, while Fonagy et al. (2018) found no reductions in arrests for the MST group 18 months after treatment compared to management as usual.

Meta-analyses on MST have also shows a mixed picture. Van der Stouwe et al. (2014) included 23 studies from 1985 until 2012 and found small but significant treatment effects on delinquency (N = 4,066 juveniles, ES = .201 for overall delinquency). Littell et al. (2021) included 23 studies conducted between 1983 and 2020, and concluded that MST reduced out-of-home placements and arrests/convictions in the United States (N = 3,987 families, ES = .27). While analysis of the treatment generally showed a positive impact on self-reported delinquency, the studies included conducted a wide range of MST adaptations for various target groups besides juvenile delinquency, such as youths with psychiatric disorders. As a result, potential effects that are specific to MST for juvenile delinquency may have been blurred.

Previous studies on the effectiveness of MST may not necessarily transfer to the Norwegian or Nordic setting. First, the Norwegian age of criminal responsibility is 15 years, while youths aged 15 to 17 years are the responsibility of the adult criminal justice system but governed by special legislation. Youths under 18 years of age very rarely get convicted, these youth are taken care of by the CWS. In the United States, the age of criminal responsibility varies by state, and juvenile delinquency is typically addressed through sanctions within the judicial system. Second, juvenile delinquency in Norway is often addressed by providing parental support or treatment from the CWS, with MST being one of the few evidence-based interventions targeting this group.

The Present Study

This study examines the impact of MST among Norwegian youths with and without prior or current registered criminal offenses on record when they began treatment. We then examined how dynamic risk factors for criminal behavior changed in these two groups from start to completion of MST. The following are the study’s specific objectives:

(1) To investigate the possible differences, such as number of youths, length of treatment, and characteristics (e.g., age, gender), between groups of youths with and without recorded prior and current offenses

(2) To investigate the differences in the dynamic risk factors within and between the two at-risk groups of youths from intake to MST completion

Because there is a high threshold for conviction of youths in Norway, we hypothesized that the group of youths having a recorded offense would have a higher level of risk based on the dynamic risk factors at intake to MST. This corresponds with other studies from the Norwegian context (Aase et al., 2020; Bhuller & Røgeberg 2022). Based on earlier studies (Hukkelberg, et al., 2022; Fonagy et al., 2018; Littell et al., 2021), we hypothesize that the dynamic risk factors of the two groups would be reduced from pre to post treatment. Youths who have committed an offense at the start of treatment may have more risk factors compared to those without such history, making them potentially more difficult to treat. For example, poorer peer relationships and family issues, and lower engagement in positive activities (Bonta & Andrews, 2018) may make them more resistant to treatment.

Criminal offense history and dynamic risk factors are registered through national routine outcome monitoring (ROM) data collection. Together with dynamic risk factors, prior or current offenses are considered a static risk factor and routinely assessed in the ROM database by the leader of the MST team at intake of treatment. Data were collected over a 10-year period from 2013 to 2022 and include the entire population of young people who completed MST treatment during that period in Norway.

This study presents a contribution to the international literature concerning changes in criminogenic risk factors through MST between the two groups. It offers valuable insights for making empirically-grounded decisions regarding the management of juveniles with offenses in Norway through the CWS.

Methods

Participants

Eligible participants were youths between the ages of 7 and 18 years who were referred to MST treatment by the municipal CWS and who completed the program between 2013 and 2022. The current study included 4,103 youths (63% boys). The sample mean age was 14.4 (SD = 1.6); only eight participants were between 7 and 8 years old, and five were 18 years old. The mean number of treatment days was 143 (SD = 39).

Measures

Characteristics of Youth

The following baseline characteristics were collected by the youth’s MST-team: does not attend school, uses violence, and substance abuse. All non-risk outcomes were scored 1 (else 0). In addition to characteristics such as gender, age, immigrant status, foster home, and days of MST treatment.

Youth Level of Service/Case Management Inventory

The YLS/CMI-Part I (Hoge & Andrews, 2011) assesses risks and identifies treatment targets among juveniles with offenses (12–17 years). The scale content is organized around the “central eight” criminogenic risks, and items are designed to reflect the unique needs of youths (Olver et al., 2012). The measure includes 42 items scored on a dichotomous scale (present = 1/not present = 0) and summarized across the eight subscales. In the present study, as part of the ROM for MST in Norway, the MST team leader completed the measure based on the referral from the CWS, as well as on the initial assessment with the family at the beginning of treatment and family and therapist information at the end of treatment.

The YLS is also used for research purposes (e.g., Hukkeberg et al., 2022, 2023). Psychometric support for the YLS/CMI in terms of reliability and validity is presented in the YLS/CMI manual (Hoge & Andrews, 2011) and supported by other studies (e.g., Onifade et al., 2008; Schmidt et al., 2005). The psychometric properties of the YLS/CMI-Part I have also been investigated in the Norwegian context (Hukkelberg, et al., 2022), where the instrument has been shown to have satisfactory concurrent and predictive validity. Reliability of the total and dynamic YLS/CMI scale was somewhat low at the time of admission to treatment, but became considerably higher at the end of treatment. This suggests that the YLS/CMI subscales overlapped more at the latter time point.

The youths were divided into two groups based on the static risk factor prior and current offenses and dispositions: (a) with recorded prior and current offenses (value = 2–5) and (b) without recorded prior and current offenses (value = 0–1). The value 0 indicated that the youth had not interacted with the justice system. A score of 1 could indicate that the youth had interacted with the justice system; however, it could also indicate a failure to comply (i.e., a breach of conditions related to a criminal or community sentence or to out-of-home placement by the CWS, such as failure to appear or comply or being unlawfully at large). A score of 2 to 5 indicated that the youth had committed one or more offenses on record that led to grave societal reactions, such as custody, convictions, and/or probations (or, if below the age of criminal responsibility, committed acts that would otherwise have qualified for a serious judicial response).

The following seven dynamic risk factors were used to measure change from intake to posttest:

Family circumstances/parenting measures inadequate supervision, difficulty in controlling behavior, inappropriate discipline, and poor youth–parent relationship (value = 0–6).

Education/employment measures disruptive classroom/schoolyard behavior, problems with peers or teachers, low achievement, and truancy (value = 0–7).

Peer relations measures the presence of delinquent friends or acquaintances, and lack of positive ones (value = 0–4).

Substance abuse measures occasional and chronic drug and alcohol use and associated negative life consequences or offenses (value = 0–5).

Leisure/recreation measures limited organized activities, poor use of leisure time, and lack of personal interests (value = 0–3).

Personality/behavior measures inflated self-esteem, verbal and physical aggression, short attention span, poor frustration tolerance, and inadequate guilt feelings (value = 0–7).

Attitudes/orientation measures antisocial/pro-criminal attitudes, avoidance or rejection of help, defiance of authorities, and lack of concern for others (value = 0–5).

Procedure

This study utilized a pre-post design and ROM data from Norway’s national MST continuous quality improvement system. MST treatments that began on or after January 1, 2013 and were completed before August 31, 2022 were included. Data on the YLS/CMI were collected at baseline and at the end of treatment from the youths’ MST teams.

Recruitment of families to MST takes place through the municipal CWS. Referrals to these services can be made by family members, schools, or by professionals or other adults who are in contact with the youth and/or family. The inclusion criteria for MST are youths between 12 and 17 years of age at risk of severe system consequences due to serious externalizing, antisocial, and/or delinquent behaviors. Exclusionary criteria are youths who live independently; engage in sex offenses in the absence of other antisocial behavior; have moderate to severe autism; are actively homicidal, suicidal, or psychotic; have psychiatric problems that are the primary reason leading to referral or who have severe and serious psychiatric problems; or for whom an intellectual disability is the only influence or the most powerful, direct contributor to the youth’s referral behaviors.

All included cases were categorized as having completed MST treatment. Cases were scored as “completed” if the family was able to receive the full MST treatment without interruptions due to lack of engagement, out-of-home placement, or factors unrelated to the progress of treatment, such as illness/death or family relocation.

Ethics

The data materials used in this study were collected for quality assurance purposes and anonymized after data collection concluded. For data collected before July 20, 2018, processing was approved by the Norwegian Data Protection Authority. After the European Union’s General Data Protection Regulation (GDPR) came into force, processing was based on informed consent. This informed consent gave explicit information on the right to opt out at any time.

The study was submitted to the Norwegian Regional Committee for Medical and Health Research Ethics (REC), which concluded that the project was considered quality assurance under Norwegian law and thus not subject to Norwegian law on research ethics and medical research.

Data Analysis

Baseline descriptive statistics of the included study variables were calculated. Categorical variables were presented using the number of participants, and percentages and differences were investigated using the chi-square test. Continuous variables were presented by means and standard deviations (SDs), and independent t tests were conducted. Pre- and post-group differences were examined using a paired sample t test. Cohen’s d was used to consider effect size, indicating the standardized difference between two means, and can be calculated as the difference between the means divided by the pooled SD (d = 0.20 is a small effect size, d = 0.50 is medium, and d = 0.80 is large; Cohen, 1992). Subsequent independent t tests were conducted, with posttests on the dynamic risk factors as dependent variables and youth group as independent variables. The significance threshold was set to .01. Of 4,103 youths, only 0.3% had missing data. All analyses were conducted using SPSS v. 29 and JASP v. 0.17.1.

Results

Sample Characteristics and Dynamic Risk Factors at Intake

In the total sample of youths (N = 4,103), 5.39% (n = 221) had recorded prior and current offenses. Table 1 shows that the number of treatment days was not significantly higher for youths with registered offenses compared to those without. However, the high standard deviation indicated large variations in both groups (range = 7–343 days of treatment). There were more boys with a somewhat higher mean age in the youth group with offenses than in the other group. About 70% used violence and/or threats, and the rate of such behavior was equally distributed between the groups. Youths with offenses showed a significantly higher rate of substance abuse compared to the others (74% vs. 44%). There was a higher percentage of youths with an immigrant background in the group with an offense.

Table 1.

Sample Characteristics of Youths at Intake to MST.

	Youths with offenses (n = 221)		Youths without offenses (n = 3,882)		p
	n	%	n	%	p
Gender
Girls	45	20.4	1,482	38.3	<.001
Boys	176	79.6	2,400	61.8	<.001
Mean age	M = 5.21	SD = 1.29	M = 14.32	SD = 1.60	<.001
Immigrant	58	26.2	579	14.9	<.001
Days of MST	M = 141.63	SD = 40.16	M = 143.52	SD = 39.49	.48
Foster home	33	14.93	328	8.45	<.001
Not school	162	73.30	2,647	68.19	.03
Use violence	160	72.40	2,862	73.72	.69
Substance abuse	165	74.66	1,716	44.20	<.001

Note. Youths with offenses = scored 2 to 5 on YLS/CMI Prior and Current Offenses. Youths without offenses = scored 0 to 1 on YLS/CMI Prior and Current Offenses.

Table 2 shows the differences regarding dynamic risk factors between offenders and non-offenders at intake to MST. The group with recorded offenses scored significantly higher on five out of the seven dynamic risk factors. Substance abuse showed the largest mean difference of 1.06 (d = 1.03), followed by peer relations (M_diff = 0.92, d = 1.12) and attitudes/orientation toward antisocial/pro-criminal attitudes (M_diff = 0.65, d = 1.97). It follows that, even though the mean difference scores were small, the high effect sizes indicate that their practical significance should not be ignored.

Table 2.

Sample T-Tests on Differences Between the Groups at Intake to MST.

	Youths with offenses	Youths without offenses
	Intake	Intake
YLS/CMI risk factors	M (SD)	M (SD)	df	t	p	Mean diff	d
Family Circ/parenting	3.87 (1.24)	3.72 (1.20)	4,101	−1.65	.09	0.12	1.20
Education/employment	4.17 (1.71)	3.80 (1.61)	4,101	−3.29	.001	0.36	1.61
Peer relations	2.80 (1.03)	1.87 (1.11)	4,101	−11.86	<.001	0.92	1.12
Substance abuse	1.72 (1.46)	0.65 (1.00)	231	−10.68	<.001	1.06	1.03
Leisure/recreation	2.21 (0.81)	1.93 (0.95)	4,101	−4.25	<.001	0.27	0.94
Personality/behavior	4.45 (1.60)	4.21 (1.50)	242	2.11	.03	0.23	1.51
Attitudes/orientation	2.74 (1.20)	2.09 (1.19)	4,101	−7.86	<.001	0.65	1.97

Note. Youths with offenses = scored 2 to 5 on YLS/CMI Prior and Current Offenses. Youths without offenses = scored 0 to 1 on YLS/CMI Prior and Current Offenses. YLS/CMI = Youth level of service/Case management inventory; M = mean; SD = standard deviation; d = Cohens d; p = p-value.

Differences in Dynamic Risk Factors from Intake to Posttest

Table 3 shows changes from intake to posttest on all seven YLS/CMI risk factors for both groups. All change scores were significant, and the effect sizes ranged from d = 0.48 to d = 2.93 (p < .001), which indicates a medium to large change (Cohen, 1992). For the group of youths with recorded prior and/or current offenses, the effect sizes were large, d = 1.40 to d = 2.93 (p < .001). This group showed the highest effect sizes and thus positive change from treatment.

Table 3.

Paired Sample T Tests and Effect Sizes Within Treatment.

	Youths with offenses				Youths without offenses
	Intake	Post-test	d	p	Intake	Post-test	d	p	Change in d-value between youth with and with no offenses
	M (SD)	M (SD)	d	p	M (SD)	M (SD)	d	p	d
YLS/CMI risk factors
Family Circ./parenting	3.87 (1.24)	0.96 (1.22)	2.91	<.001	3.72 (1.20)	1.01 (1.11)	2.71	<.001	0.20
Education/employment	4.17 (1.71)	1.24 (1.27)	2.93	<.001	3.80 (1.61)	1.32 (1.24)	2.47	<.001	0.46
Peer relations	2.80 (1.03)	1.47 (1.05)	1.33	<.001	1.87 (1.11)	0.84 (0.88)	1.02	<.001	0.31
Substance abuse	1.72 (1.46)	0.31 (0.68)	1.40	<.001	0.65 (1.00)	0.17 (0.49)	0.48	<.001	0.92
Leisure/recreation	2.21 (0.81)	1.04 (0.92)	1.17	<.001	1.93 (0.95)	1.01 (0.93)	0.92	<.001	0.25
Personality/behavior	4.45 (1.60)	1.56 (1.34)	2.89	<.001	4.21 (1.50)	1.62 (1.32)	2.59	<.001	0.30
Attitudes/orientation	2.74 (0.08)	0.90 (0.07)	1.84	<.001	2.09 (1.19)	0.71 (0.92)	1.37	<.001	0.47

Youths in the no-offenses group showed effect sizes that ranged from medium to large (d = 0.48 to d = 2.71, p < .001). The pattern of change was the same across the two groups, except for one risk factor, substance abuse. The difference in substance abuse (YLS/CMI) from pre- to posttest between the two groups was d = 0.92. Youths with prior offenses scored significantly higher at intake (M = 1.72), but they also exhibited a larger change after MST compared to the group with no prior offenses (difference in d = 0.92).

Subsequent independent t tests were conducted, with posttests on the dynamic risk factors as dependent variables and youth group as an independent variable. Figures 1 to 3 illustrate the significant differences between the two groups and Table 3 shows the mean values and SDs at posttest. A significant difference was found in three dynamic risk factor scores at posttest (Figures 1 –3). Youths in the no-offenses group scored significantly lower on negative peer relations (t[4,083] = −10.15, p ≤ .001), substance abuse (t[4,083] = −4.11, p < .001), and positive attitudes/orientation toward criminal behavior (t[4,083] = −2.88, p = .004). No other significant differences were found between the youths with and without a recorded offense at posttest.

Figure 1.

Results from intake to post-test in MST on peer relations.

Figure 2.

Results from intake to post-test in MST on substance abuse.

Figure 3.

Results from intake to post-test in MST on positive attitudes and orientation toward criminal behavior.

Discussion

Research has shown that engaging in criminal behavior at a young age is strongly associated with a higher likelihood of reoffending later in life (Bonta & Andrews, 2018; Frase & Roberts, 2019). Antisocial behavior during adolescence directly impacts the long-term well-being and future prospects of young individuals (Colman et al., 2009; Langevin et al., 2022). Thus, evidence-based methods must be utilized to effectively deter such behaviors. According to the Responsivity, Needs, and Risk Model, understanding the dynamic factors associated with these young individuals may address their needs and help them avoid further involvement in criminal activities. Our results showed that, during the last 10 years, 4,103 youths completed MST in Norway. These youths were referred to MST based on serious externalizing, antisocial, and/or delinquent behaviors. The mean age in the group with an offense at intake (5.39%, n = 221) was 15 years, indicating that some were under the Norwegian age of criminal responsibility. Youths with prior/current offenses did not receive significantly more treatment days compared to those with no offenses at intake. It is somewhat surprising that only 221 young people with an offense received MST over a 10-year period.

There were more boys and youths with an immigrant background in the group of youths with recorded offenses. This is in line with research showing that boys are more likely than girls to be involved in delinquency between the ages of 12 and 15 (Smith & McAra, 2004). In addition, register based studies from Nordic countries show that youths of immigrant background were overrepresented among registered juveniles with offenses (Klement, 2020; Salmi et al., 2015). This may be explained by differences in social background, neighborhood resources (Hällsten et al., 2013), delinquent friends (Svensson & Shannon, 2021) or weaker parental monitoring among immigrant parents (Salmi et al., 2015). However, consideration should also be given to the possibility that such difference could be attributed to racial bias among the police and court system (Holmberg & Kyvsgaard, 2003)

Importantly, 70% of the sample in both groups of youths did not go to school, and/or was in a school setting that increased the risk of marginalization. Administrative data from Norway also show that offending youths are overrepresented among those with the lowest performances in primary school measured by school credits, which indicates that they struggle at school (Bhuller & Røgeberg 2022). This is worrying because schools are important developmental arenas for youths, and school dropout can lead to delinquency (Henry et al., 2012), unemployment (Doku et al., 2019), and poorer health later in life (Lansford et al., 2016). In addition, arrests and incarceration may affect future employment opportunities (Bhuller et al., 2019).

The youths with offenses scored significantly higher on six out of the seven dynamic criminogenic risk factors at intake, with mean differences ranging from 0.23 to 1.06. Overall, youths with offenses in this sample also had higher risk levels on dynamic risk factors, which constitute predictors of re-offense (Bonta & Andrews, 2018). This should be addressed to reduce the risk of recidivism. Of note is the large difference between the two groups regarding substance abuse, with the mean difference between the groups being 1.06, suggesting that the offenses group had substantial problems with substance use. One reason for this could be that substance use was a mediating factor for offending, as the influence of drugs or alcohol may lower the threshold for committing offenses. Another possible explanation is that the use of illegal substances for some constituted the actual offense.

We tested for possible differences regarding the dynamic risk factors from intake to completion of MST between the two groups of youths. All the factors measured by the YLS/CMI were significantly reduced from intake to MST completion. This is in line with what we expected based on former studies (Hukkelberg, et al., 2022, 2023; Fonagy et al., 2018). These results correspond with the multisystemic focus of MST, as interventions with caregivers are intentionally developed to address several of the risk and protective factors that are included in the YLS/CMI: positive parenting and family cohesion, peer relationships, substance abuse, school attendance and functioning, and positive activities during the youth’s free time (Henggeler et al., 2009).

One of the largest reductions was found for family circumstances/parenting in both of the two groups of youths (d = 2.91 vs. d = 2.71). This indicates that inadequate supervision, difficulty in controlling behavior, inappropriate discipline, and poor youth–parent relationship decreased after the family participated in MST. Increased family cohesion, positive and consistent parenting, and parental monitoring are among the primary aims of MST (Henggeler et al., 2009). The results indicate that MST succeeded in attaining these aims.

The results indicate that the pattern of change in dynamic risk factors was similar and showed large effect sizes in change for both groups. However, the change in substance abuse for the offenses group was considerably greater than that for the no-offenses group (d = 1.40 vs. d = 0.48). Among juveniles with offenses, previous research indicates that substance use disorder appears to negatively moderate the relationship between risk factors, such as peer relations and antisocial attitudes, and outcomes such as rearrest (Schubert et al., 2011). Youths with substance use disorders have been shown to have a greater number of risk-relevant treatment targets, which should be addressed as an immediate target for change in juveniles with offenses (Guebert & Olver, 2014). In the Norwegian administrative dataset, Bhuller and Røgeberg (2022) found that recidivism was particularly high among people who were punished for drug offenses.

Except for peer relations, the risk factors at the end of MST treatment were comparatively equal in the offenses and non-offenses groups. Although there were significant differences between the two groups regarding substance abuse and attitudes/orientation at the end of treatment (cf. Table 3), the mean differences between these two factors were small (0.17 and 0.19, respectively). This may indicate that, on average, the effect of MST treatment levels off at a certain point regardless of initial risk. It also suggests that MST treatment is more effective for youths presenting higher risk levels at intake, who also represent the defined target group of MST. This may imply that MST is less effective when severity is lower. However, this is not supported in the present analyses, as the risk levels at the end of treatment were just as low or lower for the no-offenses group.

One exception from the pattern was peer relations. The differences in risk level between the two groups were large and significant both at intake and at the end of treatment, and the reductions in risk during treatment were large and significant for both groups. However, unlike the other factors, association with negative peers concluded at an elevated and moderate risk level for the offenses group. Thus, for the initially more criminally burdened group, breaking negative and building positive peer-relations was harder to attain during MST treatment. As negative peer relations are considered the most proximal risk factor for delinquency in the MST model (Henggeler et al., 2009), this may point toward a need to review and, if possible, strengthen the interventions directed at peer relations and thereby prevent further marginalization.

Strength and Limitations

This study’s strength lies in its use of ROM data collected from MST teams over the past 10 years, which represent a national sample of all families that have completed MST. This strengthens external validity, as the data reflect MST in regular clinical practice. The large sample and near absence of missing data (0.3%) added power to the analyses. According to Hodgson et al. (2007), such data have a lower risk of selection bias than data from randomized control trials.

Several limitations of the study should be noted. The study was based on a single group pre-post design. Thus, we cannot exclude the possibility that changes during treatment may be caused by, for example, history, maturation, or regression toward the mean. It should also be noted that, when comparing a small group (youths with prior offenses) with a larger group, several factors may affect the validity and generalizability of the results. Compared to a larger group, a smaller group has less statistical power to reveal statistical differences and may be less representative of the target population. Initial findings also revealed some differences between the two groups, for example, with regard to gender. Moreover, we subtracted effect sizes between a small group (n = 221) and large group (n = 3,882), well aware of the fact that sample size is part of the equation. Larger samples tend to produce smaller standard errors and more precise estimates than small sample sizes. Consequently, the interpretation of these results should be carried out with caution.

Despite strong external validity, we could not estimate reliability for the YLS/CMI subscales as we did not have access to item-level data. This limits the internal validity of the study. Responses from different informants (e.g., youths and therapists) would increase knowledge and strengthen the external validation of the study. The standardization of the YLS/CMI was derived from the Canadian manual (Hoge & Andrews, 2011). It has been documented that the YLS/CMI has proven to be a valuable tool for predicting delinquency in a Norwegian residential setting (Kanestrøm et al., 2024a). However, it is not standardized in Norway. Thus, comparison of the total YLS/CMI score to other studies should be made with caution. Age and gender are also static risk factors that predict criminal behavior. However, such variables cannot be modified and were not included in these analyses. Future research regarding criminogenic risk factors should, however, investigate whether gender and age present important moderators in MST.

Policy Implications

In the context of a national plan aimed at addressing youth crime (Ministry of Children and Families, 1999), and given that MST is one of the few evidence-based treatments for juvenile delinquency, it is paradoxical that only 5.39% of the youths had a recorded prior and/or current offense at intake. This low percentage is incongruous with the number of registered offenses committed by youths in Norway per year (Ministry of Justice and Public Security, 2020–2021). The prevalence of criminogenic factors among these youths suggests that they belong to a marginalized and vulnerable group, there is an overrepresentation of boys with many being below the age of criminal responsibility (15 years) and having substance abuse issues.

Previous research has shown a strong correlation between committing offenses at a young age and later re-offending. It is important to intervene on criminological risk factors to prevent this cycle (Bonta & Andrews, 2018; Frase & Roberts, 2019). This study supports Norway’s restorative rather than punitive policy regarding juvenile offending. Both this study and findings from residential care in Norway (Kanestrøm et al., 2024b) suggest that interventions should be aimed at reducing risk factors according to the principles of the Responsivity, Needs, and Risk Model. MST can effectively reduce dynamic risk factors for youths, especially those with prior offenses, and plays a crucial role in disrupting and preventing criminal pathways for them within the CWS. These findings indicate that increased investment in and expansion of MST within the national plan may lead to more effective interventions for at-risk youth.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Ragnhild Bjørknes

Author Biographies

Ragnhild Bjørknes, PhD, is a Research Professor at The Norwegian Center for Child Behavioral Development. Her research focuses on effective treatment programs in child welfare and mental health services for children and youth with behavior and mental health problems. Her research interests include criminogenic risk factors and recidivism.

Silje Hukkelberg, PhD, is a Senior Researcher at the Norwegian Center for Child Behavioral Development. Her research interests include evaluation of treatment effects in programs targeting child problem behaviors, and psychometric evaluations of instruments used to assess children’s social and emotional behaviors.

Knut Taraldsen is Special Adviser and Psychology Specialist at The Norwegian Center for Child Behavioral Development. He is member of the National Implementation Team for MST in Norway and responsible for the nationwide quality improvement data.

Andreas Teisner Høstmælingen, PhD, is director of research and researcher at The Norwegian Center for Child Behavioral Development. His research focuses on treatment and prevention of mental health issues and behavioral problems within mental health services, child welfare, and schools. He also focuses on theoretical issues regarding evidence-based psychological practice (EBPP).

References

Aase

Lønnum

Sørlie

M.-A.

Hagen

Gustavson

Utgarden

I.-H.

(2020). Children, young people and crime. How to prevent children and young people from entering or continuing with a criminal career? Summary and assessment of effective measures, treatment and organization. Norwegian Institute of Public Health.

Baglivio

M. T.

Jackowski

(2013). Examining the validity of a juvenile offending risk assessment instrument across gender and race/ethnicity. Youth Violence and Juvenile Justice, 11(1), 26–43. https://doi.org/10.1177/1541204012440107

Bhuller

Dahl

G. B.

Løken

K. V.

Mogstad

(2019). Incarceration, recidivism, and employment. Journal of Political Economy, 128(4), 1269–1324. https://doi.org/10.1086/705330

Bhuller

Røgeberg

(2022). Characteristics of children and young people who commit criminal acts and the effects of punishment. Oslo Economics.

Bonta

Andrews

D. A.

(2018). The psychology of criminal conduct (6th ed.). Routledge.

Borduin

C. M.

Mann

B. J.

Cone

L. T.

Henggeler

S. W.

Fucci

B. R.

Blaske

D. M.

Williams

R. A.

(1995). Multisystemic treatment of serious juvenile offenders: Long-term prevention of criminality and violence. Journal of Consulting and Clinical Psychology, 63, 569–578.

Butler

Baruch

Hickley

Fonagy

(2011). A randomized controlled trial of MST and a statutory therapeutic intervention for young offenders. Journal of the American Academy of Child & Adolescent Psychiatry, 50, 1220–1235.

Casey

(2016). Dynamic risk and sexual offending: The conundrum of assessment. Psychology, Crime & Law, 22(1–2), 104–123. https://doi.org/10.1080/1068316x.2015.1111366

Clarke

M. C.

Peterson-Badali

Skilling

T. A.

(2017). The relationship between changes in dynamic risk factors and the predictive validity of risk assessments among youth offenders. Criminal Justice and Behavior, 44(10), 1340–1355. https://doi.org/10.1177/0093854817719915

10.

Cohen

(1992). A power primer. Psychological Bulletin, 112(1), 155–159. https://doi.org/10.1037/0033-2909.112.1.155

11.

Colman

Murray

Abbott

R. A.

Maughan

Kuh

Croudace

T. J.

Jones

P. B.

(2009). Outcomes of conduct problems in adolescence: 40 year follow-up of national cohort. The BMJ, 338, Article 2981. https://doi.org/10.1136/bmj.a2981

12.

Cottle

C. C.

Lee

R. J.

Heilburn

(2001). The prediction of criminal recidivism in juveniles. Criminal Justice and Behavior, 2(8), 367–394. https://doi.org/10.1177/0093854801028003005

13.

Council of Europe. (2021). Contribution by the Council of Europe Secretariat to the UN Committee on the Rights of the Child’s Day of General Discussion on “children’s rights and alternative care”. https://rm.coe.int/coe-contribution-to-uncrc-dgd-on-children-and-alternative-care/1680a3d85f

14.

Cuervo

Villanueva

(2015). Analysis of risk and protective factors for recidivism in Spanish youth offenders. International Journal of Offender Therapy and Comparative Criminology, 59(11), 1149–1165. https://doi.org/10.1177/0306624x14557917

15.

Doku

D. T.

Acacio-Claro

P. J.

Koivusilta

Rimpelä

(2019). Health and socioeconomic circumstances over three generations as predictors of youth unemployment trajectories. European Journal of Public Health, 29(3), 517–523. https://doi.org/10.1093/eurpub/cky242

16.

Fonagy

Butler

Cottrell

Scott

Pilling

Eisler

Fuggle

Kraam

Byford

Wason

Ellison

Simes

Ganguli

Allison

Goodyer

I. M.

(2018). Multisystemic therapy versus management as usual in the treatment of adolescent antisocial behaviour (START): A pragmatic, randomized controlled, superiority trial. Lancet Psychiatry, 5, 119–133. https://doi.org/10.1016/S2215-0366(20)30131-0

17.

Frase

R. S.

Roberts

J. V.

(2019). Prior record and the risk of recidivism. In Roberts

J. V.

Frase

R. S.

(Eds.), Paying for the past: The case against prior record sentence enhancements (pp. 41–59). Oxford University Press. https://doi.org/10.1093/oso/9780190254001.003.0003

18.

Guebert

A. F.

Olver

M. E.

(2014). An examination of criminogenic needs, mental health concerns, and recidivism in a sample of violent young offenders: Implications for risk, need, and responsivity. The International Journal of Forensic Mental Health, 13(4), 295–310. https://doi.org/10.1080/14999013.2014.955220

19.

Hällsten

Szulkin

Sarnecki

(2013). Crime as a Price of Inequality? The Gap in Registered Crime between Childhood Immigrants, Children of Immigrants and Children of Native Swedes. The British Journal of Criminology, 53(3), 456–481, https://doi.org/10.1093/bjc/azt005

20.

Hanson

R. K.

Bourgon

Helmus

Hodgson

(2009). The principles of effective correctional treatment also apply to sexual offenders: A meta-analysis. Criminal Justice and Behavior, 36(9), 865–891. https://doi.org/10.1177/0093854809338545

21.

Heffernan

Wegerhoff

Ward

(2019). Dynamic risk factors: Conceptualization, measurement, and evidence. Aggression and Violent Behavior, 48, 6–16. https://doi.org/10.1016/j.avb.2019.06.004

22.

Henggeler

S. W.

Melton

G. B.

Smith

L. A.

(1992). Family preservation using multisystemic therapy: An effective alternative to incarcerating serious juvenile offenders. Journal of Consulting and Clinical Psychology, 60, 953–961.

23.

Henggeler

S. W.

Melton

G. B.

Smith

L. A.

Schoenwald

S. K.

Hanley

J. H.

(1993). Family preservation using multisystemic treatment: Long-term follow-up to a clinical trial with serious juvenile offenders. Journal of Child and Family Studies, 2, 283–293.

24.

Henggeler

S. W.

Schoenwald

S. K.

Borduin

C. M.

Rowland

M. D.

Cunningham

P. B.

(2009). Multisystemic therapy for antisocial behavior in children and adolescents (2nd ed.). Guilford Press.

25.

Henry

K. L.

Knight

K. E.

Thornberry

T. P.

(2012). School disengagement as a predictor of dropout, delinquency, and problem substance use during adolescence and early adulthood. Journal of Youth and Adolescence, 41(2), 156–166. https://doi.org/10.1007/s10964-011-9665-3

26.

Hodgson

Bushe

Hunter

(2007). Measurement of long-term outcomes in observational and randomized controlled trials. The British Journal of Psychiatry, 50(S50), s7884. https://doi.org/10.1192/bjp.191.50.s78

27.

Hoge

R. D.

Andrews

D. A.

(2011). Youth level of service/case management inventory 2.0: User’s manual. Multi-Health Systems.

28.

Holmberg

Kyvsgaard

(2003). Are immigrants and their descendants discriminated against in the Danish criminal justice system? Journal of Scandinavian Studies in Criminology and Crime Prevention, 4(2), 125–142. https://doi.org/10.1080/14043850310020027

29.

Hukkelberg

S.S.

Ogden

Taraldsen

(2023). Changes in the YLS/CMI risk inventory scores as predictors of MST behavioral treatment goals, Journal of Children’s Services, 18 (2), 133–146. https://doi.org/10.1108/JCS-05-2022-0016

30.

Hukkelberg

Ogden

Thøgersen

D. M.

(2022). Youth Level of Service/Case Management Inventory Assessments as Predictors of Behavioral Change in Multisystemic Therapy and Functional Familiiy Therapy in Norway. Research on Social Work Practice, 1–9. https://doi.org/10.1177/10497315221086641

31.

Kanestrøm

Stallvik

Lydersen

Skokauskas

Kaasbøll

(2024a). Adolescents in therapeutic residential care: Treatment needs and characteristics. European Journal of Social Work, 27, 733–747. https://doi.org/10.1080/13691457.2023.2266779

32.

Kanestrøm

Stallvik

Lydersen

Skokauskas

Hoftun

Nilsen

C. K.

Kaasbøll

(2024b). Criminogenic risk assessment beyond Juvenile justice: Exploring the predictive ability of the Youth Level of Service/Case Management Inventory (YLS/CMI) in Norway. Criminal Justice and Behavior, 51(7), 975–993. https://doi.org/10.1177/00938548241237181

33.

Klement

(2020). Studies of immigrant crime in Denmark. Nordic Journal of Criminology, 21(1), 11–31. https://doi.org/10.1080/2578983X.2019.1702270

34.

Langevin

Caspi

Barnes

J. C.

Brennan

Poulton

Purdy

S. C.

Ramrakha

Tanksley

P. T.

Thorne

P. R.

Wilson

Moffitt

T. E.

(2022). Life-course persistent antisocial behavior and accelerated biological aging in a longitudinal birth cohort. International Journal of Environmental Research and Public Health, 19(21), Article 14402. https://doi.org/10.3390/ijerph192114402

35.

Lansford

J. E.

Dodge

K. A.

Pettit

G. S.

Bates

J. E.

(2016). A public health perspective on school dropout and adult outcomes: A prospective study of risk and protective factors from age 5 to 27 years. Journal of Adolescent Health, 58(6), 652–658. https://doi.org/10.1016/j.jadohealth.2016.01.014

36.

Littell

J. H.

Pigott

T. D.

Nilsen

K. H.

Green

S. J.

Montgomery

O. L. K.

(2021). Multisystemic Therapy® for social, emotional, and behavioural problems in youth age 10 to 17: An updated systematic review and meta-analysis. Campbell Systematic Reviews, 17, CD004797. https://doi.org/10.1002/cl2.1158

37.

McCart

M. R.

Sheidow

A. J.

Jaramillo

(2023). Evidence base update of psychosocial treatments for adolescents with disruptive behavior. Journal of Clinical Child & Adolescent Psychology, 52(4), 447–474.

38.

Miller

W. T.

Campbell

C. A.

Papp

Ruhland

(2021). The contribution of static and dynamic factors to recidivism prediction for Black and White youth offenders. International Journal of Offender Therapy and Comparative Criminology, 66(16), 1779–1795. https://doi.org/10.1177/0306624X211022673

39.

Ministry of Children and Families. (1999). Meld. St. 17 Handlingsplan mot barne- og ungdomskriminalitet. Stm17 (regjeringen.no).

40.

Ministry of Justice and Public Security. (2020–2021). Meld. St. 34 Sammen mot barne-, ungdoms- og gjengkriminalitet. Meld. St. 34 (2020–2021) - regjeringen.no.

41.

Olver

M. E.

Stockdale

K. C.

Wong

S. C. P.

(2012). Short and long-term prediction of recidivism using the Youth Level of Service/Case Management Inventory in a sample of serious young offenders. Law and Human Behavior, 36, 331–344. https://doi.org/10.1037/h0093927

42.

Onifade

Davidson

Campbell

Turke

Malinowski

Turner

(2008). Predicting recidivism in probationers with the Youth Level of Service Case Management Inventory (YLS/CMI). Criminal Justice and Behavior, 35(4), 474–483. https://doi.org/10.1177/0093854807313427

43.

Salmi

Kivivuori

Aaltonen

(2015). Correlates of immigrant youth crime in Finland. European Journal of Criminology, 12(6), 681–699. https://doi.org/10.1177/1477370815587768

44.

Savage

Ellis

S. K.

Kozey

(2013). A selective review of the risk factors for antisocial behavior across the transition to adulthood. Psychology, 4, 1–7. https://doi.org/10.4236/psych.2013.46A2001

45.

Sawyer

A. M.

Borduin

C. M.

(2011). Effects of multisystemic therapy through midlife: A 21.9-year follow-up to a randomized clinical trial with serious and violent juvenile offenders. Journal of Consulting and Clinical Psychology, 79(5), 643–652.

46.

Schaeffer

C. M.

Borduin

C. M.

(2005). Long-term follow-up to a randomized clinical trial of multisystemic therapy with serious and violent juvenile offenders. Journal of Consulting and Clinical Psychology, 73(3), 445–453.

47.

Schmidt

Hoge

R. D.

Gomes

(2005). Reliability and validity analysis of the Youth Level of Service/Case Management Inventory. Criminal Justice and Behavior, 32, 329–344. https://doi.org/10.1177/0093854804274373

48.

Schubert

C. A.

Mulvey

E. P.

Glasheen

(2011). Influence of mental health and substance use problems and criminogenic risk on outcomes in serious juvenile offenders. Journal of the American Academy of Child & Adolescent Psychiatry, 50(9), 925–937. https://doi.org/10.1016/j.jaac.2011.06.006

49.

Smith

D. J.

McAra

(2004). Edinburgh study of youth transitions and crime: No. 2. Gender and youth offending. Center for Law and Society, University of Edinburgh.

50.

Svensson

Shannon

(2021). Immigrant background and crime among young people: An examination of the importance of delinquent friends based on national self-report data. Youth & Society, 53(8), 1335–1355. https://doi.org/10.1177/0044118x20942248

51.

Timmons-Mitchell

Bender

M. B.

Kishna

M. A.

Mitchell

C. C.

(2006). An independent effectiveness trial of multisystemic therapy with juvenile justice youth. Journal of Clinical Child and Adolescent Psychology, 35, 227–236. https://doi.org/10.1207/s15374424jccp3502_6

52.

van den Berg

J. W.

Smid

Schepers

Wever

van Beek

Janssen

Gijs

. (2018). Tre predictive properties of dynamic sex offender risk assessment instruments: A meta-analysis. Psychological Assessment, 30(2), 179–191. https://doi.org/10.1037/pas0000454

53.

Van der Stouwe

Asscher

J. J.

Stams

G. J. J. M.

Dekovic

Van der Laan

P. H

. (2014). The effectiveness of multisystemic therapy (MST): A meta-analysis. Clinical Psychology Review, 34, 468–481. https://doi.org/10.1016/j.cpr.2014.06.006

54.

Weiss

Han

Harris

Catron

Ngo

V. K.

Caron

Gallop

Guth

(2013). An independent randomized clinical trial of multisystemic therapy with non-court-referred adolescents with serious conduct problems. Journal of consulting and clinical psychology, 81(6), 1027–1039. https://doi.org/10.1037/a0033928