Revista de Psicología Clínica con Niños y Adolescentes Copyright© 2023 RPCNA

Vol. 10 nº. 1- Enero 2023 - pp 9-19 www.revistapcna.com - ISSN 2340-8340

doi: 10.21134/rpcna.2023.10.1.1

Revista de Psicología Clínica con Niños y Adolescentes

What works in the psychological treatment of

child conduct problems? An umbrella review
of meta-analytic studies
Estrella Romero, María Álvarez-Voces, Beatriz Díaz-Vázquez, & Laura López-Romero
University of Santiago de Compostela

Abstract
Conduct problems (CP) constitute a major field for child clinical psychology, in terms of not only prevalence, but also developmental, social, and
clinical relevance. This study carried out an umbrella review of meta-analyses on the efficacy of treatment/indicated prevention of child CP (mean
age range up to 12 years). Following a registered protocol, several databases (i.e., Web of Science, PsycINFO, PROSPERO, and The Cochrane
Library) were searched for meta-analyses published from January 2002 to March 2022. Nine meta-analytic publications met the required criteria, and
the main characteristics and findings of the studies were systematically described. Weighted effect sizes (ESs) were calculated through RStudio
program. Analyses of heterogeneity, publication bias, quality (AMSTAR-2), and credibility were also conducted. Results indicate that parent training
has been the most studied intervention, with a weighted ES of d = -0.49 (95% CI -.67 to -.32). The reviewed studies identified several moderators
for the efficacy of parent training, including individual (severity of CP), family (financial disadvantage), and intervention (delivery format) characte-
ristics. Results from child-centered play therapy were also analyzed, d = -.34 (95% CI -.40 to -.28), but the reduced number of meta-analyses and
the weakness detected by quality assessment suggest the need of cautiously considering the pooled effects. In general, evidence seems to be
suggestive of the efficacy of treatment of CP, particularly for parent training. Nevertheless, efficacy seems to be moderate, heterogeneity indexes
are high, and quality assessments of meta-analyses are often suboptimal. This study suggests several avenues to strengthen knowledge in this field.
Keywords: Conduct problems; Intervention; Children; Meta-analysis; Umbrella review.

Resumen
¿Qué funciona en el tratamiento psicológico de los problemas de conducta en niños? Revisión umbrella de estudios meta-analíticos. Los prob-
lemas de conducta (PC) son un foco prioritario de atención en la psicología clínica infantil. Este estudio presenta una revisión umbrella de me-
ta-análisis sobre la eficacia del tratamiento/prevención indicada de los PC infantiles (rango de edad media hasta 12 años). Siguiendo un protocolo
registrado, se buscaron meta-análisis (enero 2002 a marzo 2022) en Web of Science, PsycINFO, PROSPERO y The Cochrane Library. Nueve
meta-análisis cumplieron con los criterios requeridos y sus características y hallazgos fueron descritos sistemáticamente; además, con RStudio
se calcularon los tamaños del efecto (TEs) ponderados. Se realizaron análisis de heterogeneidad, sesgo de publicación, calidad (AMSTAR-2) y
credibilidad. El entrenamiento parental fue la intervención más estudiada, con un TE medio ponderado de d = -.49 (IC del 95%: -.67 a -.32), y se
identificaron diversos moderadores de eficacia, incluyendo características individuales (gravedad de los PC), familiares (desventaja económica) y
de intervención (formato de administración). También se analizaron los resultados de la terapia de juego centrada en el niño, d = -.34 (IC del 95%:
-.40 a -.28), pero el reducido número de meta-análisis y la debilidad detectada por la evaluación de la calidad sugieren la necesidad de considerar
con cautela los efectos agrupados. En general, la evidencia es sugestiva de eficacia del tratamiento de los PC, específicamente del entrenamiento
parental. No obstante, la eficacia parece ser moderada, los índices de heterogeneidad elevados y las evaluaciones de calidad de los meta-análisis
no suelen ser óptimas. Este estudio sugiere varias vías para reforzar el conocimiento en este campo.
Palabras clave: Problemas de conducta; Intervención; Niños; Meta-análisis; Umbrella review.

Conduct problems (CP) are among the most relevant distur- including oppositionality, vindictiveness, aggression, temper tan-
bances addressed by clinical child and adolescent psychology. Under trums, noncompliance and rule breaking. These patterns may reach
the concept of CP a variety of behavioural patterns are subsumed, clinical significance, conforming disorders enlisted by the diagnos-

Corresponding author:
Estrella Romero.
Facultad de Psicología, University of Santiago de Compostela.
Campus Sur, s/n, 15781 Santiago de Compostela.
E.mail: estrella.romero@usc.es

What works in the psychological treatment of child conduct problems? An umbrella review of meta-analytic studies
10 Psychological treatment of child conduct problems

tic classifications, such as oppositional-defiant disorder (ODD) and perspective (Aromataris et al., 2015; López-López et al., 2022), the
conduct disorder (CD). Both at clinical and subclinical levels, CP present study portrays the meta-analytic results on this topic, with the
represent a significant public health concern, as they are one of the aim of identifying the types of programs that have been most studied
most prevalent types of problems in children and adolescents (e.g., during the last two decades, the efficacy associated to the different
Merikangas & Kessler, 2022), and, additionally, CP seem to have therapies, the strength of such evidence, and the limitations to be
experimented a sharp increase in the context of the pandemic times addressed by future research.
(Lebrun-Harris et al., 2022; Romero et al., 2020). During the last
years, large-scale surveys have shown that CP are the primary cause Method
for referring children to mental health services (Encuesta Nacional
de Salud de España [ENSE], 2017; Ghandour et al., 2019). The impli- Registration and guidelines
cations of CP for behavioural, emotional, and social development
are also well described in the literature, including an increased risk The protocol for this study was registered on PROSPERO (regis-
for academic failure, depression, drug abuse, social exclusion, and tration number CRD42022322438). This study was conducted accord-
criminality, thus entailing substantial health, educational and social ing to the recommendations of the Preferred Reporting Items for
costs (Rissanen et al., 2021). Systematic Reviews and Meta–Analyses (PRISMA 2020) (Page et al.,
The existence of diverse developmental profiles among children 2021), as well as recommendations for umbrella reviews (Fusar-Poli &
with CP has been one major tenet in developmental psychopathol- Radua, 2018) and the Assessing the Methodological Quality of System-
ogy for the last decades (e.g., Cicchetti, 2016). Even when the specific atic Reviews (AMSTAR–2) (Shea et al., 2017). The PRISMA checklist is
number and shape of the different trajectories is still a debated issue provided in Table A1 (see Appendix, https://osf.io/6wvb5/).
(Girard et al., 2018; Gutman et al., 2019), a big body of research has
supported the need to differentiate at least between childhood-onset Search strategy
and adolescence-onset CP (Moffitt, 1993; Patterson, 1982). While
adolescent-onset CP tend to be linked to the developmental tasks A comprehensive search was performed by two researchers
of the teenaging years and, therefore, they present a higher chance (M.A.V., B.D.V.) in the following data bases: Web of Science, PROS-
of remission after adolescence, childhood-onset CP tend to be asso- PERO, PsycINFO and The Cochrane Library. The descriptors used for
ciated to a complex net of interwoven determinants, operating in a the search are presented in Table 1. Language restrictions were applied
cascade mode (Masten & Cicchetti, 2010); this process may lead to (English and Spanish). Other restrictions were publication date
chronic CP patterns, which are increasingly difficult to treat as the (01/01/2002 to 15/03/2022), age range of participants (mean age up to
child grows up. Based on this evidence, the intervention on child- 12 years old) and article type: meta-analysis and network meta-analy-
hood-onset CP becomes a major challenge for mental health pro- sis. The final literature search was conducted on 18 March 2022. The
viders; intervening on children’s CP is not only important in terms complete search strategy is described in PROSPERO and in the
of treatment needs, but it is also considered a major opportunity for Appendix (Table A2).
prevention of future clinical and psychosocial dysfunctions (Harley
et al., 2008). Table 1. Search Strategy for the Meta-Analyses
For many years, researchers have tried to design, implement, and
evaluate treatment strategies for children’s CP. Psychological treat- Search items Descriptors
ments have aimed to change children’s CP through the intervention 1. Condition or (conduct problem*) OR (conduct disorder*) OR
in different settings and domains: e.g., parenting/family relations domain being (CD) OR (disruptive behaviour*) OR (externaliz*)
(Brinkmeyer & Eyberg, 2003), teacher/classroom interactions (Leflot studied OR (oppositional defiant disorder*) OR (ODD)
et al., 2010), and children’s cognitions and skills (Lochman & Wells,
2. Intervention (treatment*) OR (invervent*) OR (therap*) OR
2002). To apprehend and summarize evidence on this topic, several
(psychotherap*)
reviews have been conducted so far, with results that have conveyed
both optimism (Furlong et al., 2012) and pessimism (Bakker et al., 3. Age (child*)
2017) on the efficacy of available treatments. Combination #1 AND #2 AND #3
Among the published reviews, meta-analyses occupy a central
space, as they provide quantitative estimations of treatment effects,
and they allow delineation of possible individual, social or treat- Studies retrieved from the databases were encoded in a bibli-
ment-related moderators. Meta-analytic reviews have been per- ographic manager software (Refworks). These studies were system-
formed in the field of children’s CP (e.g., Comer et al., 2013; Ray et al., atically reviewed by two researchers independently of each other
2015), yet there is a paucity of overarching studies synthetizing what (M.A.V., B.D.V.). Moreover, the reference list of all studies was
we have learned from such quantitative reviews. This study attempts checked for additional relevant records. Interrater reliability was
to contribute to this line, by systematically collecting and analyzing Kappa = .89, which can be rated as strong (McHugh, 2012). Diver-
results from meta-analyses on the efficacy of psychological treatments gences in study selection were resolved by two independent research-
for children with CP. In the view of the developmental implications ers (E.R.T., L.L.R.).
of CP (with or without the ODD/CD diagnostic labels), our review
includes meta-analyses dealing with strict-sense treatment and with Inclusion and exclusion criteria
indicated prevention (Gordon, 1987). Therefore, we considered studies
on interventions for children that were referred for clinical assistance, The eligibility criteria for this umbrella review were based on
but also interventions for children whose CP were identified through the Population, Intervention, Comparator, Outcome, Study design
screening procedures in non-clinical populations. With an umbrella (PICOS) statement (Higgins et al., 2019), as shown in Table 2.
 Estrella Romero, María lÁ varez-Voces, Beatriz Díaz-Vázquez, & Laura López-Romero 11

Table 2. Eligibility Criteria for Inclusion of Meta-Analyses a total of seven authors who were contacted, one of them provided the
necessary missing data.
Inclusion criteria Exclusion criteria
P Participate Children whose Children with Correction of primary study overlap
mean age is up to 12 neurodevelopmental
years old with CP problems (e.g., autism The overlap of individual studies among the included meta-anal-
defined on the basis of spectrum disorder) or yses can lead to a significant distortion of the results (Munder et al.,
standardized tools. CP intellectual disabilities 2013). Therefore, following the instructions of Fusar-Poli & Radua
were referred to both (2018), when two or more meta-analyses overlapped more than 50%
clinical (e.g., conduct
of their individual studies, we included the one with the largest data-
disorder, oppositional
base and the most recent one.
defiant disorder)
and subclinical
After elimination, we observed that the highest overlap was 15
(e.g., disruptive, studies (Leijten et al., 2018; Lundahl et al., 2006), followed by 11 (Lei-
externalizing) levels jten et al., 2008; Menting et al., 2013). Out of the total of 278 studies
we observed 30 overlaps (10.79%). Additional information on over-
I Intervention Psychological Pharmacological
lapping articles is included in Table A5 in Appendix.
interventions interventions,
specifically designed to prevention programs
treat CP (universal and
Data analysis
selective prevention),
interventions not Quantitative data analysis was carried out with the RStudio pro-
specifically designed gram for Windows. As an umbrella pooling approach, bare bones
to reduce child meta-analysis of standardized differences was performed using the
CP, interventions psychometric method of Hunter & Schmidt (2004). As input data, the
specifically designed ESs reported for each study were considered. The index used to obtain
to treat internalizing the overall ES was Cohen’s d; for the studies that did not report this
disorders or ADHD index, the ES was transformed into it. Then, the ES weighted by the
C Comparison Comparison group: no With individuals sample size (pooled Cohen’s d), the standard error (SE) and the 95%
treatment, treatment as without CP confidence intervals (CI) were calculated. Heterogeneity was calcu-
usual, another type of lated with the I2 and Q statistics. In relation to the I2 statistic, hetero-
treatment and waiting- geneity was considered low if it was 25%, moderate if it was 50% and
list control group high if it was 75% (Higgins et al., 2003).
O Outcome Effects on CP,
moderating treatment - Publication bias
variables
Publication bias was analysed using the File-safe N with the
S Study Metanalysis and Qualitative systematic
Orwin’s approximation to detect the number of studies with an
network meta-analysis reviews
ES of zero that would reduce the observed average ES to a specific
value (e.g., half or quarter). Funnel plots and asymmetry indexes
Data extraction with Egger’s Test could not be performed, as the number of resulting
meta-analyses was deemed too low to provide reliable results, accord-
Two researchers (B.D.V., M.A.V.) extracted data independently ing to the guidelines by Guyatt et al. (2011).
using a previously defined protocol. Divergences were reviewed
by a third (L.L.R.) and fourth (E.R.T.) researcher. Qualitative data Quality assessment and credibility of evidence
extracted included: (a) first author and year of publication, (b) coun-
tries of the articles included in the meta-analysis, (c) participant data The methodological quality of each meta-analysis was assessed by
(sample size, gender distribution, mean age range), (d) study objec- two researchers (B.D.V. and M.A.V.) who worked independently with the
tives, (e) description of the control group, (f) instruments used to Assessment of Multiple Systematic Reviews tool (AMSTAR-2); result-
measure CP, (g) informants for the outcome measures, (h) type of ing minor discrepancies were discussed to reach a complete agreement.
intervention studied, (i) intervention target, (j) intervention charac- AMSTAR-2 classifies reviews into: (a) critically low-quality review, (b) low
teristics and (k) main outcomes (see Table A3 in Appendix). Quanti- quality review, (c) moderate quality review and (d) high quality review.
tative data included: (a) total number of individual studies included To assess credibility, studies were classified into 4 different classes
in each meta-analysis (k), (b) reported ES in relation to the decrease (Belbasis et al., 2015; Bellou et al., 2017; Fusar-Poli & Radua, 2018):
in externalizing problems after the application of the intervention –– Class I – Convincing: number of cases >1000, p < 10-6, I2 < 50%,
(e.g., Cohen’s d, Hedges’ g), (c) 95% confidence interval of the ES, and prediction interval excluding the null, no small-study effects and
(d) heterogeneity (I2 and Q) (see Table 3). In addition, Table A4 (see no excess significance bias.
Appendix), shows the search strategy (database, search period) and –– Class II – Highly suggestive: number of cases >1000, p < 10-6, larg-
the number of WOS citations for each meta-analysis. est study with a statistically significant effect and class I criteria
When it was not possible to extract some relevant data (e.g., gen- not met.
der distributions, individual studies entered to calculate each ES, –– Class III – Suggestive: number of cases>1000, p < 10−3 and class
country of origin of the studies), we tried to contact the authors. From I–II criteria not met.
12 Psychological treatment of child conduct problems

Table 3. Results of all Meta-Analysis Included in the Review (K = 9)

Reference Participants Studies ESa ES 95 % CI Heterogeneity

(N) (k) (Cohen’s d) (p)
Burkey et al. 4,441 11 SMD = -0.56 -.56 -0.79 to –0.32 I2 = 73.6%
(2018)
Comer et al. 3,042 36 Hedges’ g = .82 -.82 -1.01 to -.63 Q = 299.3
(2013) (p<.001)
Leijten et al. 4,892 IP = 50 d = -0.55 -.55 -0.70 to –0.39 I2= 66.66%
(2018) 3,327 T = 45 d = - 0.69 -.69 -0.84 to – 0.54 -
Lundahl et al. 1,838 29 (five years d = .44 -.44 - Q = 49.24
(2006) 648 or less) d = .31 -.31 - p < .05
10 (5-10 years) Q = 19.65
p < .05
Menting et al. 4,745 50 d = .27 -.27 .21 to .34 (p < .001) -
(2013)
Parker et al. 504 14 Hedges’ g = -.34 -.34 -.52 to -.17 Q=5.08
(2021) (p<.001) p = .97
I2 = 0%
Ray et al. 407 7 d =.34 -.34 .10 to .58 Q = 2.16
(2015) (p <.05) (p >.05)
Van IJzendoorn 1,03 13 r = .07 -.14 -0.42 to -0.09 I2 = 63.83%
et al. (2022) (p <.01)
Veenman et al. 1,081 At risk: 6 d = -.26 -.26 -0.35 to -0.04 Q = 18.58
(2018) 828 Clinical: 7 d = -.19 -.19 (p =.01) I2 = 73.00%
(p < .001)
Q = 6.78
I2 = 11.47%
(p = .34)

Note. (-) = data not reported in the article. ES (Cohen’s d): ES calculated from the ESs reported by the articles.
a
The signs used by each author are respected.

Figure 1. PRISMA Flow Diagram Illustrating Literature Research and Selection Process


Identification of studies via databases and registers Identification of studies via other methods




Identification

Records identified from (n= 333): Records removed before

 PsycInfo (n = 136) screening: Records identified from:
Cochrane (n = 174) Records marked as ineligible Citation searching (n = 1)
 Prospero (n = 11) by automation tools (n = 14)
 WOS (n = 12)







 Records screened (n = 319) Records excluded
 (n = 284)


Screening



 Reports assessed for eligibility
(n = 35)
 Reports excluded (n= 27):
 Average age of the range
over 12 (n = 14)
 Repeated articles (n = 8)
 Exclusion criteria (n = 5)
 etc.



Included

Studies included in review

 (n = 9)



Note. Extracted from “The PRISMA 2020 statement: an updated guideline for reporting systematic reviews” (Page et al., 2021).


 Estrella Romero, María lÁ varez-Voces, Beatriz Díaz-Vázquez, & Laura López-Romero 13

–– Class IV – Weak: p < .05 and class I–III criteria not met. follow-ups (Comer et al., 2013). However, some studies reported that
–– Non-significant: p > .05. the benefits of the intervention were maintained up to six-month (Bur-
key et al., 2018) and one-year follow-ups (Leijten et al., 2018).
Results Studies also found that the largest reductions in CP were
found in those groups with more severe problems at the start of
Literature search results / Study selection treatment (Lundahl et al, 2006; Menting et al., 2013).
In terms of theoretical orientation, behavioural and non-be-
The literature search identified 333 relevant articles and 319 havioural interventions were compared (Comer et al., 2013);
unique citations across all sources. Of these, 35 full papers were results showed that behavioural interventions offered better results
assessed for eligibility and nine reviews were selected for synthesis in reducing externalising behaviours than non-behavioural treat-
(see Figure 1 for the PRISMA flow chart). The list of articles excluded ments (e.g., family systems approaches, nondirective counselling).
following full-text review, together with the reasons for exclusion, is The efficacy of specific therapeutic techniques was evaluated
also available in Appendix (Table A6). by the examined studies. Specifically, three techniques were asso-
ciated with stronger program effects; positive reinforcement as a
Study characteristics general technique, praise as a specific operationalization of posi-
tive reinforcement, and the use of natural or logical consequences
A total of nine meta-analyses evaluating the efficacy of the treatment as a discipline technique (Leijten et al., 2018).
of CP were included. A list of the individual studies included in each Finally, considering the level of intervention, Leijten et al.
meta-analysis can be found in Table A7 in the Appendix. It should be (2018) found that indicated prevention and treatment were the
noted that three meta-analyses did not provide such information. Descrip- most effective interventions in reducing behavioural problems
tion of the study characteristics are presented in Table A3 in Appendix. compared to the other types of prevention. Some other studies
The oldest publication was by Lundahl et al. (2006) and the most recent by also noted that treatment was more powerful than prevention
Van Ijzendoorn et al. (2022). The most cited article in WOS was Lundahl approaches in reducing externalizing problems (Burkey et al.,
et al. (2006) with a total of 608 citations so far (see Table A4 in Appendix). 2018; Menting et al., 2013).
Three meta-analyses did not provide data on the country of origin of the
individual studies. Of those meta-analyses that did provide data, the most Moderator variables
frequent country of origin was USA (57.14%), Canada (42.86%) and the The examined studies highlighted some relevant variables as
Netherlands (42.86%). potential moderators of intervention efficacy.
The total sample consisted of 26,763 boys and girls with detected With respect to age of children, the studies found that pro-
clinical or subclinical CP. Percentages by gender could not be reported grammes were effective at all ages (Burkey et al., 2018) or that there is
because five of the nine meta-analyses included did not report such data. no significant difference considering age (Lundhal et al., 2016). How-
The type of intervention most studied by the meta-analyses was parent ever, some authors reported larger effects in samples of older youth
training (6/9); a relevant number of studies were focused, particularly, and with higher percentages of males (Comer et al., 2013).
on Incredible Years (3/9). Other interventions addressed were play ther- Some family characteristics were also analysed. As for ethnic
apy (3/9), classroom programs (1/9) and counselling (1/9). The target of characteristics, treatment effects were consistent across samples of
the interventions was parents (6/9), children (6/9), and teachers (2/9). varying ethnic compositions (Comer et al., 2013; Parker et al., 2012).
A few studies were specifically focused on randomized controlled trials Some studies considered the marital status of parents: while Lundahl
(RCTs; 3/9), while some others included both randomized and non-ran- et al. (2006) found that being single did not moderate the effective-
domized designs (6/9). ness of parent training, Menting et al. (2013) corroborated these find-
Intervention characteristics varied widely across studies due to the dif- ings, but pointed to a certain tendency towards downward modera-
ferent nature of the therapies studied. Interventions included both individ- tion of being single. In terms of socio-economic status, some authors
ual and group formats, with an average duration ranging from seven days found that economically disadvantaged families had fewer benefits
to four years. As for the control groups, the most frequent types were wait- than those who were not economically disadvantaged (Lundahl et
ing list [WL] (5/9) and no treatment [NT] (4/9). As outcome measures, al., 2006). Additionally, the delivery mode (individual vs. group) was
25 different instruments for measuring CP were used, the most frequent found to affect the strength of the effects in specific social groups, i.e.,
being: Child Behaviour Checklist [CBC] (7/9), Eyberg Child Behaviour parents who participated only in individually delivered parent train-
Inventory [ECBI] (4/9), Strengths and Difficulties Questionnaire [SDQ] ing changed significantly more than those who participated only in
(3/9) and Behavioural Assessment System for Children [BASC] (3/9). group-delivered parent training among economically disadvantaged
families (Lundhal et al., 2006). With regards to the delivery mode,
Main findings from the individual meta-analyses no significant difference was found between self – directed parent
training and face – to – face parent training (Lundhal et al., 2006).
Efficacy of the interventions With respect to the type of control group used as comparison, it
Most of the treatments assessed showed to be effective in reduc- was reported that effects of parent training varied depending on the
ing CP; only one study analyzing videofeedback-based interventions type of control group, with Treatment as Usual (TAU) comparisons
showed no substantial treatment effects for externalizing problems yielding even stronger effects than comparisons with no active treat-
(Van ljzendoorn et al., 2022). Parent training and play therapy were the ment (Comer et al., 2013). On the other hand, the number of sessions
most investigated interventions across all the reviews, with significant attended by parents was positively elated to the intervention effects
short- and long-term effects on the reduction of behavioural problems. (Menting et al., 2013). Additionally, with regards to parent training,
Timing of post-treatment assessment significantly moderated the effects measured through teacher reports were smaller than the ones
effects, with immediate effects being larger than effects at subsequent reported by parents (Menting et al., 2013).
14 Psychological treatment of child conduct problems

Figure 2. Forest Plot of the Results of the Meta-Analyses of All Interventions

Note. Black boxes represent the ESs of the studies, the lines through them correspond to the 95% CI.

Figure 3. Forest Plot of The Results of The Meta-Analyses of Parent Training Interventions

Note. Black boxes represent the ESs of the studies, the lines through them correspond to the 95% CI.

Quantitative results: Umbrella pooling of ESs plus sign indicates that they increased. Results expressed in Cohen’s d
ranged from -0.82 (95% CI = -1.01; -.63) (Comer et al., 2013) to -.14
Table 3 displays the results of each meta-analysis. Different types (95% CI = -.04; -.19) (Van Ijzendoorn et al., 2022).
of indicators for ES were found; Cohen’s d was used in eight of the 12 The I2 statistic ranged from 0% (Parker et al., 2021) to 73.6% (Bur-
associations, Hedges’ g in two, the correlation coefficient r in one, and key et al., 2018), and was classified as moderate heterogeneity (Hig-
SMD in one. Therefore, all measures were transformed to Cohen’s d. gins & Thomson, 2002) in four studies (Burkey et al., 2018; Leijten et
Some studies used the plus sign (+) in ES (e.g., Comer et al., 2013) to al., 2018; Van Ijzendoorn et al., 2022; Veenman et al., 2018).
indicate that the treatment was effective; however, other authors used
the minus sign (-) (e.g., Veenman et al., 2018) to indicate a decrease Pooling results from all types of psychological interventions
in behavioural problems. For our study, we decided to homogenise Considering all nine meta-analyses, 278 studies (median k= 13.5)
the results so that a minus sign indicates that CP decreased, and a and 26,783 individuals (median n = 1,460), the average ES weighted
 Estrella Romero, María lÁ varez-Voces, Beatriz Díaz-Vázquez, & Laura López-Romero 15

Figure 4. Forest Plot of the Results of The Meta-Analyses of Play Therapy Interventions

Note. Black boxes represent the ESs of the studies, the lines through them correspond to the 95% CI.

by the sample size was d = -0.49 (95% CI -.67 to -.32). The correspond- (2022), the risk of bias raised some concern in more than half of the
ing Z value of –5.47 was statistically significant (p < .0001). Evidence studies (56%).
of significant heterogeneity was supported by the significance test for The AMSTAR-2 rating of overall confidence in the reviews was
the observed value of Q and the I2 statistical (Q11 = 5688.38, p < .001, considered critically low in six studies, moderate in two studies and
I2 = 99.76%). It was also found that 87.5% of the meta-analyses had a high in one study. At first, the interobserver agreement for all items
significant ES. Figure 2 provides a forest plot of the ESs, the weights, was 77.7%. Differences in criteria were discussed and finally the
and the 95% intervals for all included studies. agreement obtained was 100%. The items that were most frequently
unmet or that omitted information were registering the protocol
Pooling results from parent training interventions beforehand (item 2), providing a list of excluded studies and justifying
Six meta-analyses grouped 167 individual studies (median k = 37) the exclusions (item 7), items related to the risk of bias (items 9, 12,
and 16,480 individuals (median n = 2,582) to evaluate the specific effi- 13), and reporting on the sources of funding of the studies included in
cacy of parent training interventions. The mean ES weighted by sam- the review (item 10). The AMSTAR-2 assessment is presented in Table
ple size was d = -.45 (95% CI -.61 to -.29). The corresponding Z-value A8 in the Appendix.
of -5.60 was statistically significant (p < .001). Evidence of significant Regarding credibility, 10 associations obtained suggestive evi-
heterogeneity was found by Q-test and I2 statistic (Q = 910.78, p < dence (meeting the following criteria: number of cases > 1,000, p
.001, I2 = 99.24%). The highest ES was d = -.69 (Leijten et al., 2018) < 10−3 and class I–II criteria not met), one obtained weak evidence
and 91.67% of the associations had a significant ES, i.e., all but one (meeting the criteria p < .05 and class I-III not met) (Ray et al., 2015)
of the meta-analyses (Van Ijzendoom et al., 2022). Figure 3 shows a and one was not significant because it did not meet the criterion of
forest plot of these associations. statistical significance (Van Ijzendoorn et al. 2022).

Pooling results from play therapy interventions Publication bias across studies
The two meta-analyses who evaluated the specific efficacy of play
therapy included 21 studies (median k = 10.5) and 911 individuals Orwin’s method estimated that 12 (for all treatments), six (for
(median = 455.5). The mean ES weighted by sample size was d =-.34 parent training) and two (for play therapy) missing studies with a
(95% CI -.40 to -.28); the corresponding Z value of -11.59 was statis- null effect would be necessary to reduce the pooled ES by half in each
tically significant (p<.001). No evidence of significant heterogeneity association. Thus, the number of meta-analyses required with zero ES
was found by Q test and statistical I2 (Q = 0, p=1, I2 = 0%). The ES in is unlikely to be met for the “all treatments” analysis; robustness of
the two meta-analyses was d = -.34, significant in both cases. Figure 4 results for specific types of interventions (and, especially, play ther-
shows a forest plot of these meta-analyses. apy) appear to be weaker in terms of publication bias.

Risk of bias in included studies and quality and credibility Discussion

assessment
The development of effective and feasible treatments for children’s
Assessment of potential risk of bias was reported in three of the CP is a major aspiration for clinical child psychology. Many experi-
nine meta-analyses using the Cochrane Collaboration Tool (Higgins ences and trials have been reported in the scientific literature; in par-
et al., 2011). The meta-analysis by Burkey et al. (2018) indicated that allel, a variety of meta-analytic studies have been published, provid-
32% of their individual studies were at high risk of bias. Their main ing material for higher-level umbrella reviews (Fusar-Poli & Radua,
problem was blinding of outcome informants. The meta-analysis by 2018). This study aimed to organize and summarize the evidence pro-
Leijten et al. (2018) indicated that, for most of the individual studies, vided by meta-analyses performed during the last 20 years, so that we
the risk of bias was low, with appropriate blinding of outcome inform- can achieve a more comprehensive view of what has been studied and
ants, treatment of incomplete data, analysis of dropouts, and selective with which results.
outcome reporting. As for the meta-analysis by Van Ijzendoorn et al. Our review shows that family programs, and, specifically, parent
16 Psychological treatment of child conduct problems

training programs are, by far, the most studied interventions in this In fact, the degree of participation in the programs was also
field. Parent training programs are grounded on the evidence proving identified as another efficacy moderator (Menting et al., 2013); this
that parenting practices are a major component in the pathways to CP finding suggests a dose-effect relationship, and reinforces the need
(Patterson, 1982). They are typically based on Operant Conditioning to prioritize motivation and adherence, because even the best quality
Learning and Social Learning Theory, and they aim to induce changes programs work just if parents are appropriately involved (Nix et al.,
in the parent-child interactions, by breaking coercive relationships 2009). In this sense, along with good delimitation of effective contents
and by promoting consistent and effective discipline. Parenting pro- and training techniques, research in this field should go deeper into
grams have gained recognition over the years, and, in fact, they are the factors and processes that influence implication and permanence
often considered as first-line treatment for CP at childhood (NICE, in the interventions (Hackworth et al., 2018).
2013). Our review estimates a pooled weighted ES of d = -.45, which In terms of effect maintenance, the reviewed meta-analyses prove
is significant, but still not high according to the usual thresholds that the effects of parent training may be significant up to one-year
(Cohen, 1988). follow-up (Leitjen et al., 2018; Lundahl et al., 2006); however, results
In this line, identifying the factors that boost or reduce the also show that effects are attenuated over time (Lundahl et al., 2006).
effect of parent training emerges as a meaningful question, and the It has been suggested that a continued care model may be appropriate
meta-analyses reviewed by this umbrella provide some relevant for families with CP, so that the effective parenting can be supported
insights. For example, some results distil which ingredients may be and adapted to developmental and family changes (Lundahl et al.,
more active within parent training programs (Leitjen et al., 2018): 2006). It has also been suggested that effect maintenance could be
positive reinforcement (particularly, social praise) and natural/logi- bolstered by including non-behavioural components (e.g., commu-
cal consequences seem to be the associated to stronger effects in the nication styles, improving emotional family climate) that may con-
intervention on CP. Such evidence may help researchers and practi- tribute to the internalization of the values and principles of positive
tioners in selecting potentially effective programs, and also in design- parenting (Lundahl et al., 2006). Yet, the usefulness of including such
ing new parenting-based interventions. relational components is still controversial (Leitjen et al., 2018), and
The reviewed studies could also identify some moderators of the more long-term evaluations of parenting programs need to be con-
effects of parent training, considering either participants’ or interven- ducted (see, for example, Romero et al., 2017). The need of long-term
tion’s characteristics. In this regard, results show that parent training studies in this field is also implied by Van Ijzendoorn et al. (2022),
may not be equally useful for all the intervention levels (Leitjen et al., who suggest that attachment-oriented components, potentially effec-
2018; Menting et al., 2013): effects are stronger for treatment than for tive to sustain the parenting changes, may take longer to show their
indicated prevention (Burkey et al., 2018; Menting et al., 2013), and effects on child behaviour.
they are stronger for indicated than for universal or selective preven- Despite the gaps that still must be covered (also addressed
tion levels (Leijten et al., 2018). This pattern of results may respond below), in general, the meta-analyses based on more than 160 indi-
to the different motivations for enrolling in treatment and prevention vidual studies and 16,000 participants support parent training as an
conditions; parenting programs are often perceived as quite demand- effective intervention for CP problems. Interestingly, Comer et al.
ing by participants (Nock & Ferriter, 2005), and it may be plausible (2013) found that the effects are stronger when parent training is
that treatment (i.e., help-seeking) parents are more willing to comply compared to TAU than to non-active treatment, thus revealing that
with attendance and tasks required by the programs, in comparison the usual service given to CP children is inadequate, and that the
to parents who have been recruited in the community after screening evidence-supported treatments should be better disseminated. As
processes. Alternatively, the difference in efficacy between treatment other authors have claimed (Lebrun-Harris et al., 2022), attention to
and prevention conditions may be related to CP severity. Previous child mental health should be strengthened, especially considering
research has already shown that intervention effects may be stronger the increasing rates of behavioural and emotional problems in young
for more severe CP cases (e.g., Leitjen et al., 2013), as there is a wider people. Such reinforcement may involve not only an improvement
range of potential improvement; this pattern was also found in the in quantity of resources, but also in quality of training and interven-
meta-analyses reviewed by the present study (e.g., Lundahl et al., tions, with more effective insertion of research results into the clin-
2006; Menting et al., 2013). ical practice.
Differential effects related to delivery modalities were also exam- Apart from parent training, not many eligible meta-analyses were
ined (Lundahl et al., 2006). Group formats were found to be as effica- focused on other specific types of treatments. An exception comes
cious as individually administered programs (e.g., Comer et al., 2013), from child-centered play therapy, with two studies entirely devoted
a finding that favors the group delivery when cost-effectiveness is a to this approach (Parker et al., 2021; Ray et al., 2015). Child-centered
priority. Nonetheless, this finding seems to be moderated by socioec- play therapy, with roots in humanistic theory, is a modality of per-
onomic conditions: disadvantaged families tend to obtain less benefit son-centered counselling based on the use of toys and plays to meet
from parent training, and this holds particularly true when the pro- the developmental needs of children (Landreth, 2012); with a non-di-
grams are administered in groups (Lundahl et al., 2006). The stronger rective approach, child-centered play therapy emphasizes the com-
efficacy of individual vs. group-oriented interventions for disadvan- munication with children, with a supportive, safe, and caring envi-
taged families may be due to multi-problematic circumstances (Taus- ronment, which promotes the development of healthy attitudes and
enfreund et al., 2016), including financial strain, lower literacy, psy- skills. Our study, pooling results from meta-analyses on child-cen-
chosocial stress, and lack of social support, which may raise personal tered play therapy, shows a significant ES (d = -.34); nevertheless, a
and situational barriers to the participation in group interventions. cautious consideration of this ES is sensible, as it stems from a low
This may be an important take-home message, as disadvantaged fam- number of meta-analyses and individual studies, with relatively low
ilies may need a more flexible, individualized approach that nourishes sample sizes, and with weak AMSTAR-2 and credibility assessments.
motivation and engagement, considering the net of family problems Research on child-centered play therapy should be encouraged, as
coexisting with CP. preliminary results may be promising; however, more investigation
 Estrella Romero, María lÁ varez-Voces, Beatriz Díaz-Vázquez, & Laura López-Romero 17

is needed not only to make this evidence more robust, but also to treatments will provide valuable feedback for theory building on eti-
describe longer-term effects and to identify efficacy moderators. ology and developmental pathways of CP. So far, results point at the
efficacy of specific treatment approaches, but results also indicate that
Limitations and future directions there is much room for improvement. As in other health disciplines
(e.g., Allen et al., 2009), rather than wondering what works, research
Results of this umbrella review need to be seen in the light of may need to be pushed into more refined questions: what works, for
some limitations. First, although high-quality studies (e.g., based on whom, and under which circumstances.
RCTs) are increasingly conducted in this field, research quality is quite
uneven, often including small sample sizes, inadequate reporting of Funding
important data, and blurry definition of treatments. Likewise, the
AMSTAR-2 quality of the reviewed meta-analyses was often compro- This study was supported by grant PID2019-107897RB-I00, with
mised. Of note, most meta-analyses were in fact conducted before the funds from MCIN/AEI/10.13039/501100011033
dissemination of the AMSTAR-2 chart in the scientific community
(Shea et al., 2017); thus, it is expected that future reviews will meet
the stringent requirements of current protocols. Overall, a proac-
tive consideration of potential reviews should be already adopted by
primary studies, as the lack of information in the individual studies References
drives a domino effect on the quality of future secondary (meta-anal-
yses) and tertiary (umbrella) reviews. Second, this umbrella review Allen, D., Gillen, E., & Rixson, L. (2009). Systematic review of the effectiveness
could just pool results from two treatment types (i.e., parent train- of integrated care pathways: what works, for whom, in which circum-
ing and child-centered play therapy). Other types, like child-centered stances? International Journal of Evidence‐Based Healthcare, 7(2), 61–74.
cognitive-behavioural therapy (CBT) could not be analyzed; although https://doi.org/10.1111/j.1744-1609.2009.00127.x
meta-analyses have been conducted on child-centered CBT (e.g., Batt- Aromataris, E., Fernandez, Ritin., Godfrey, C. M., Holly, C., Khalil, H., &
agliese et al., 2015), they typically cluster together results from chil- Tungpunkom, P. (2015). Summarizing systematic reviews: methodolog-
dren and adolescents, thus precluding the specific analyses required ical development, conduct and reporting of an umbrella review approach.
by this review. Expectedly, more research will be accumulated dur- International Journal of Evidence-Based Healthcare, 13(3), 132–140.
ing the following years, separately addressing treatment effects on https://doi.org/10.1097/XEB.0000000000000055
individuals in different developmental stages; in this line, the effect Bakker, M. J., Greven, C. U., Buitelaar, J. K., & Glennon, J. C. (2017). Prac-
of age on treatment effects deserves more consideration; results on titioner Review: Psychological treatments for children and adolescents
this topic are far from being clarified, as it is suggested by the pres- with conduct disorder problems–a systematic review and meta‐analysis.
ent review. Third, this study provided pooled ESs for all treatments, Journal of Child Psychology and Psychiatry, 58(1), 4–18. https://doi.
and for two specific types of treatments, but significant heterogeneity org/10.1111/jcpp.12590
was found across studies; to the extent that more meta-analyses can Battagliese, G., Caccetta, M., Luppino, O. I., Baglioni, C., Cardi, V., Mancini, F.,
be further conducted, the sources of such variability should be a tar- & Buonanno, C. (2015). Cognitive-behavioural therapy for externalizing
get of future umbrella reviews. Relatedly, the study of moderators of disorders: A meta-analysis of treatment effectiveness. Behaviour Research
treatment effects should be a priority for research (see also McMahon and Therapy, 75, 60–71. https://doi.org/10.1016/j.brat.2015.10.008
et al., 2021). Although the reviewed studies show some interesting Belbasis, L., Bellou, V., Evangelou, E., Ioannidis, J. P., & Tzoulaki, I. (2015).
results, much more is needed to be known, for example, in terms of Environmental risk factors and multiple sclerosis: an umbrella review
comorbidities, outcome informants, intervention length, combined of systematic reviews and meta-analyses. The Lancet Neurology, 14(3),
treatments, or modality of delivery; specifically, the effects of digital- 263–273. https://doi.org/10.1016/S1474-4422(14)70267-4
ized interventions as compared to traditional formats deserve a closer Bellou, V., Belbasis, L., Tzoulaki, I., Evangelou, E., & Ioannidis, J. P. (2016).
view, given the current bourgeon of e-Health interventions, and the Environmental risk factors and Parkinson’s disease: an umbrella review
need of rendering programs accessible to the communities (Florean of meta-analyses. Parkinsonism & Related Disorders, 23, 1–9. https://doi.
et al., 2020). Finally, the diversity of children with CP should be con- org/10.1016/j.parkreldis.2015.12.008
sidered. For the last years, the phenotypic, etiological, and develop- Brinkmeyer, M. Y., & Eyberg, S. M. (2003). Parent-child interaction therapy for
mental heterogeneity of CP has been emphasized (Waller et al., 2013). oppositional children. In A. E. Kazdin & J. R. Weisz (Eds.), Evidence-based
Accordingly, the need of tailored interventions has been highlighted psychotherapies for children and adolescents (pp. 204–223). The Guilford Press.
(Glenn et al., 2019; Romero et al., 2019), for a better tuning of pro- Broidy, L. M., Nagin, D. S., Tremblay, R. E., Bates, J. E., Brame, B., Dodge, K.
grams to the specific needs of children and families; the effects of such A., Fergusson, D., Horwood, J. L., Loeber, R., Laird, R., Lynam, D. R., Mof-
interventions should come at the forefront for further research. fitt, T. E., Pettit, G. S. & Vitaro, F. (2003). Developmental trajectories of
childhood disruptive behaviours and adolescent delinquency: A six-site,
Concluding remarks cross-national study. Developmental Psychology, 39(2), 222–245. https://
doi.org/10.1037/0012-1649.39.2.222
Our umbrella review indicates that there is suggestive evidence *Burkey, M. D., Hosein, M., Morton, I., Purgato, M., Adi, A., Kurzrok, M.,
of efficacy for treatment of childhood CP. Evidence is more robust Kohrt, B. A., & Tol, W. A. (2018). Psychosocial interventions for disruptive
for parent training, with a number of moderators suggested in terms behaviour problems in children in low‐ and middle‐income countries: A
of individual, family or implementation factors. Further research on systematic review and meta‐analysis. Journal of Child Psychology and Psychi-
effective and sustainable treatments emerges as a major practical need atry, 59(9), 982–993. http://dx.doi.org/10.1111/jcpp.12894
nowadays, in a socio-historical context of recognized vulnerability for Cicchetti, D. (Ed.) (2016). Developmental psychopathology (Vol. 3, Maladapta-
children’s mental health; at the same time, the inquiry into effective tion and Psychopathology, 3rd ed.). Wiley.
18 Psychological treatment of child conduct problems

Cicchetti, D., Rogosch, F. A., Lynch, M., & Holt, K. D. (1993). Resilience in mal- Higgins, J. P., Altman, D. G., Gøtzsche, P. C., Jüni, P., Moher, D., Oxman, A. D.,
treated children: Processes leading to adaptive outcome. Development and Savović, J., Schulz, K. F., Weeks, L., & Sterne, J. A. (2011). The Cochrane
Psychopathology, 5(4), 629–647. https://doi.org/10.1017/S0954579400006209 Collaboration’s tool for assessing risk of bias in randomised trials. Bmj,
Cohen, J. (1988). Statistical power analysis for the behavioural sciences, 2nd ed. 343. https://doi.org/10.1136/bmj.d5928
Erlbaum. Higgins, J. P., Thomas, J., Chandler, J., Cumpston, M., Li, T., Page, M. J., &
*Comer, J. S., Chow, C., Chan, P. T., Cooper-Vince, C., & Wilson, L. A. S. Welch, V. A. (eds). (2019) Cochrane handbook for systematic reviews of
(2013). Psychosocial treatment efficacy for disruptive behaviour problems interventions. 2nd Edition. John Wiley & Sons.
in very young children: A meta-analytic examination. Journal of the Amer- Higgins, J. P., Thompson, S. G., Deeks, J. J., & Altman, D. G. (2003). Measuring
ican Academy of Child & Adolescent Psychiatry, 52(1), 26–36. http://dx.doi. inconsistency in meta-analyses. Bmj, 327(7414), 557–560. https://doi.
org/10.1016/j.jaac.2012.10.001 org/10.1136/bmj.327.7414.557
Florean, I. S., Dobrean, A., Păsărelu, C. R., Georgescu, R. D., & Milea, I. (2020). Hunter, J. E., & Schmidt, F. L. (2004). Methods of meta-analysis: Correcting
The Efficacy of internet-based parenting programs for children and ado- error and bias in research findings. Sage.
lescents with behavior problems: A meta-analysis of randomized clinical Lahey, B. B., Van Hulle, C. A., Waldman, I. D., Rodgers, J. L., D’Onofrio, B. M.,
trials. Clinical Child and Family Psychology Review 23, 510–528. https:// Pedlow, S., Rathouz, P. & Keenan, K. (2006). Testing descriptive hypothe-
doi.org/10.1007/s10567-020-00326-0 ses regarding sex differences in the development of conduct problems and
Furlong, M., & McGilloway, S. (2012). The Incredible Years Parenting program delinquency. Journal of Abnormal Child Psychology, 34, 737–755. https://
in Ireland: A qualitative analysis of the experience of disadvantaged par- doi.org/10.1007/s10802-006-9064-5
ents. Clinical Child Psychology and Psychiatry, 17(4), 616–630. https://doi, Landreth, G. L. (2012). Play Therapy. The Art of the Relationship. Routledge.
org/10.1177/1359104511426406 Lebrun-Harris, L. A., Sappenfield, O. R., & Warren, M. D. (2022). Missed
Fusar-Poli, P., & Radua, J. (2018). Ten simple rules for conducting umbrella and delayed preventive health care visits among US children due to the
reviews. Evidence-Based Mental Health, 21(3), 95–100. https://doi. COVID-19 pandemic. Public Health Reports, 137(2), 336–343. https://doi.
org/10.1136/ebmental-2018-300014 org/10.1177/00333549211061322
Ghandour, R. M., Sherman, L. J., Vladutiu, C. J., Ali, M. M., Lynch, S. E., Bitsko, Leflot, G., van Lier, P. A. C., Onghena, P., & Colpin, H. (2010). The role of
R. H., & Blumberg, S. J. (2019). Prevalence and treatment of depression, teacher behavior management in the development of disruptive behaviors:
anxiety, and conduct problems in US children. The Journal of Pediatrics, An intervention study with the good behavior game. Journal of Abnormal
206, 256–267. https://doi.org/10.1016/j.jpeds.2018.09.021 Child Psychology 38, 869–882. https://doi.org/10.1007/s10802-010-9411-4
Girard, L. C., Tremblay, R. E., Nagin, D., & Côte, S. M. (2019). Development of *Leijten, P., Gardner, F., Melendez-Torres, G. J., Van Aar, J., Hutchings, J.,
aggression subtypes from childhood to adolescence: A group-based mul- Schultz, S., Knerr, W., & Overbeek, G. (2018). What to teach parents
ti-trajectory modelling perspective. Journal of Abnormal Child Psychology, to reduce disruptive child behaviour: two meta-analyses of parenting
47(5), 825–838. https://doi.org/10.1007/s10802-018-0488-5 program components. Journal of the American Academy of Child and
Glenn, A. L., Lochman, J. E., Dihsion, T., Powell, N. P., Boxmeyer, C., Kass- Adolescent Psychiatry, 58(2). https://doi.org/10.1016/j.jaac.2018.07.900
ing, F., Qu, L., & Romero, D. (2019). Toward tailored interventions: Leijten, P., Raaijmakers, M. A., de Castro, B. O., & Matthys, W. (2013).
Sympathetic and parasympathetic functioning predicts responses to an Does socioeconomic status matter? A meta-analysis on parent training
intervention for conduct problems delivered in two formats. Prevention effectiveness for disruptive child behaviour. Journal of Clinical Child &
Science, 20, 30–40. https://doi.org/10.1007/s11121-017-0859-0 Adolescent Psychology, 42(3), 384–392. https://doi.org/10.1080/15374416
Gordon, R. (1987). An operational classification of disease prevention. In J. .2013.769169
A. Steinberg & M. M. Silverman (Eds.), Preventing mental disorders (pp. Lochman, J. E., & Wells, K. C. (2002). Contextual social–cognitive mediators
20–26). Department of Health and Human Services. and child outcome: A test of the theoretical model in the Coping Power
Gutman, L. M., Joshi, H., & Schoon, I. (2019). Developmental trajectories program. Development and Psychopathology, 14(4), 945–967.
of conduct problems and cumulative risk from early childhood to ado- López-López, J. A., Rubio-Aparicio, M., & Sánchez-Meca, J. (2022). Overviews
lescence. Journal of Youth and Adolescence, 48(2), 181–198. https://doi. of reviews: Concept and Development. Psicothema, 34(2), 175–181.
org/10.1007/s10964-018-0971-x https://doi.org/10.7334/psicothema2021.586
Guyatt, G. H., Oxman, A. D., Schünemann, H. J., Tugwell, P., & Knottnerus, *Lundahl, B., Risser, H. J., & Lovejoy, M. C. (2006). A meta-analysis of parent
A. (2011). GRADE guidelines: A new series of articles in the Journal of training: Moderators and follow-up effects. Clinical Psychology Review,
Clinical Epidemiology. Journal of Clinical Epidemiology, 64(4), 380–382. 26(1), 86–104. http://dx.doi.org/10.1016/j.cpr.2005.07.004
https://doi.org/10.1016/j.jclinepi.2010.09.011 McHugh, M. L. (2012). Interrater reliability: the kappa statistic. Biochemia
Hackworth, N. J., Matthews, J., Westrupp, E. M., Nguyen, C., Phan, T., Medica, 22(3), 276–282.
Scicluna, A., Cann, W., Bethelsen, D., Bennetts, S. K., & Nicholson, J. Masten, A., & Cicchetti, D. (2010). Developmental cascades. Develop-
M. (2018). What influences parental engagement in early intervention? ment and Psychopathology, 22(3), 491–495. https://doi.org/10.1017/
Parent, program and community predictors of enrolment, retention and S0954579410000222
involvement. Prevention Science 19, 880–893. https://doi.org/10.1007/ McMahon, R. J., Goulter, N., & Frick, P. J. (2021). Moderators of psychosocial
s11121-018-0897-2 intervention response for children and adolescents with conduct prob-
Harley, M., Murtagh, A., & Cannon, M. (2008). Conduct disorder–psychiatry’s lems. Journal of Clinical Child & Adolescent Psychology, 50(4), 525–533.
greatest opportunity for prevention: A commentary on ‘From conduct https://doi.org/10.1080/15374416.2021.1894566
disorder to severe mental illness: associations with aggressive behaviour, *Menting, A. T. A., de Castro, B. O., & Matthys, W. (2013). Effectiveness of the
crime and victimization’ by Hodgins et al. (2007). Psychological Medicine, incredible years parent training to modify disruptive and prosocial child
38(7), 929–931. https://doi.org/10.1017/S0033291708002699 behaviour: A meta-analytic review. Clinical Psychology Review, 33(8),
Higgins, J. P., & Thompson, S. G. (2002). Quantifying heterogeneity in a 90–913. http://dx.doi.org/10.1016/j.cpr.2013.07.006
meta‐analysis. Statistics in Medicine, 21(11), 1539–1558. https://doi.
org/10.1002/sim.1186
 Estrella Romero, María lÁ varez-Voces, Beatriz Díaz-Vázquez, & Laura López-Romero 19

Merikangas, K. R., Nakamura, E. F., & Kessler, R. C. (2022). Epidemiology Romero, E., López-Romero, L., Domínguez-Álvarez, B., Villar, P., & Gómez-
of mental disorders in children and adolescents. Dialogues in Clinical Fraguela, J. A. (2020). Testing the effects of COVID-19 confinement in
Neuroscience, 11(1), 7–20. https://doi.org/10.31887/DCNS.2009.11.1/ Spanish children: The role of parents’ distress, emotional problems and
krmerikangas specific parenting. International Journal of Environmental Research and
Ministerio de Sanidad, Consumo y Bienestar Social del Gobierno de España Public Health, 17(19), 6975. https://doi.org/10.3390/ijerph17196975
(2017). Encuesta Nacional de Salud [ENSE] https://www.sanidad.gob. Romero, E., Rodríguez, C., Villar, P., & Gómez-Fraguela, X. A. (2017). Inter-
es/estadEstudios/estadisticas/encuestaNacional/encuestaNac2017/ENSE vention on early-onset conduct problems as indicated prevention for
2017_notatecnica.pdf substance use: A seven-year follow up. Adicciones, 29(3), 150–162.
Moffit, T. E. (2003). Life-course-persistent and adolescence-limited antisocial Schaeffer, C. M., Petras, H., Ialongo, N., Masyn, K. E., Hubbard, S., Poduska, J.,
behaviour: A 10-year research review and a research agenda. En B. & Kellam, S. (2006). A comparison of girls’ and boys’ aggressive-disruptive
B. Lahey, T. E. Moffit y A. Caspi (Eds.), Causes of conduct disorder and behaviour trajectories across elementary school: Prediction to young adult
juvenile delinquency (pp. 49–75). Gilford Press. antisocial outcomes. Journal of Consulting and Clinical Psychology, 74(3),
Moffitt, T. E. (1993). Adolescence-limited and life-course-persistent antisocial 500–510. https://doi.org/10.1037/0022-006X.74.3.500
behaviour: A developmental taxonomy. Psyhological Review, 100, 674–701. Schardt, C., Adams, M. B., Owens, T., Keitz, S., & Fontelo, P. (2007). Utilization
https://doi.org/10.1037/0033-295X.100.4.674 of the PICO framework to improve searching PubMed for clinical ques-
Munder, T., Brütsch, O., Leonhart, R., Gerger, H., & Barth, J. (2013). Researcher tions. BMC Medical Informatics and Decision Making, 7, 16. https://doi.
allegiance in psychotherapy outcome research: An overview of reviews. org/10.1186/1472-6947-7-16
Clinical Psychology Review, 33(4), 501–511. https://doi.org/10.1016/j. Shea, B. J., Reeves, B. C., Wells, G., Thuku, M., Hamel, C., Moran, J., Moher, D.,
cpr.2013.02.002 Tugwell, P., Welch, V., Kristjansson, E., & Henry, D. A. (2017). AMSTAR
National Collaborating Centre for Mental Health (2013). Antisocial behaviour 2: A critical appraisal tool for systematic reviews that include randomised
and conduct disorders in children and young people: The NICE guideline on or non-randomised studies of healthcare interventions. Bmj, 358. https://
recognition, intervention and management. London: The British Psycholog- doi.org/10.1136/bmj.j4008
ical Society and The Royal College of Psychiatrists. Tausendfreund, T., Knot-Dickscheit, J., Schulze, G. C., Knorth, E. J., & Gri-
Nix, R. L., Bierman, K. L., McMahon, R. J., & The Conduct Problems Preven- etens, H. (2016). Families in multi-problem situations: Backgrounds,
tion Research Group. (2009). How attendance and quality of participation characteristics, and care services. Child & Youth Services, 37(1), 4–22.
affect treatment response to parent management training. Journal of Con- https://doi.org/10.1080/0145935X.2015.1052133
sulting and Clinical Psychology, 77(3), 429–438. https://doi.org/10.1037/ *van IJzendoorn, M. H., Schuengel, C., Wang, Q., & Bakermans-Kranenburg,
a0015028 M. J. (2022). Improving parenting, child attachment, and externalizing
Nock, M. K., & Ferriter, C. (2005). Parent management of attendance and behaviours: Meta-analysis of the first 25 randomized controlled trials on
adherence in child and adolescent therapy: A conceptual and empirical the effects of Video-feedback Intervention to promote Positive Parenting
review. Clinical Child and Family Psychology Review, 8(2), 149–166. and Sensitive Discipline. Development and Psychopathology, 1–16. https://
https://doi.org/10.1007/s10567-005-4753-0 doi.org/10.1017/S0954579421001462
Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mul- *Veenman, B., Luman, M., & Oosterlaan, J. (2018). Efficacy of behavioural
row, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., classroom programs in primary school. A meta-analysis focusing on rand-
Glanville, J., Grimshaw, J. M., Hróbjartsson, A., Lalu, M. M., Li, T., Loder, omized controlled trials. PLoS ONE, 13(10), 23. http://dx.doi.org/10.1371/
E. W., Mayo-Wilson, E., McDonald, S., . . . Moher, D. (2021). The PRISMA journal.pone.0201779
2020 statement: An updated guideline for reporting systematic reviews. Waller, R., Gardner, F., & Hyde, L. W. (2013). What are the associations
Bmj, 372, 71. https://doi.org/10.1136/bmj.n71 between parenting, callous–unemotional traits, and antisocial behaviour
*Parker, M. M., Hunnicutt Hollenbaugh, K. M., & Kelly, C. T. (2021). Exploring in youth? A systematic review of evidence. Clinical Psychology Review, 33,
the impact of child-centered play therapy for children exhibiting behav- 593–608. http://dx.doi.org/10.1016/j.cpr.2013.03.001
ioural problems: A meta-analysis. International Journal of Play Therapy,
30(4), 259–271. http://dx.doi.org/10.1037/pla0000128 *Articles included in the umbrella review
Patterson, G. R. (1982). Coercitive family process: A social learning approach to
family intervention (vol. 3). Castalia Publishing Company. Supplementary materials
*Ray, D. C., Armstrong, S. A., Balkin, R. S., & Jayne, K. M. (2015). Child-cen-
tered play therapy in the schools: review and meta-analysis. Psychology in Additional supporting information (Appendix) can be found online at
the Schools, 52(2), 107–123. https://doi.org/10.1002/pits.21798 https://osf.io/6wvb5/
Rissanen, E., Kuvaja‐Köllner, V., Elonheimo, H., Sillanmäki, L., Sourander, Table A1. PRISMA Checklist
A., & Kankaanpää, E. (2022). The long‐term cost of childhood conduct Table A2. Search Strategy Used
problems: Finnish Nationwide 1981 Birth Cohort Study. Journal of Child Table A3. Main Results of ach Study Included in the Umbrella Review
Psychology and Psychiatry, 63(6), 683–692. https://doi.org/10.1111/ Table A4. Databases Searched, Search Period and Citations in WOS for Each
jcpp.13506 Meta-Analysis
Romero, E., Gómez-Fraguela, X. A., Villar, P., & Rodríguez, C. (2019). Prevención Table A5. Overlapping of Individual Studies Included in the Meta-Analyses
indicada de los problemas de conducta: entrenamiento de habilidades Table A6. Excluded Items After Abstract and Full Text Review and Reasons (N=27)
socioemocionales en el contexto escolar. Revista de Psicología con Niños y Table A7. Individual Studies Included in the Selected Meta-Analyses
Adolescentes, 6(3), 39-47. https://doi.org/10.21134/rpcna.2019.06.2.1 Table A8. AMSTAR-2 Quality Assessment