Introduction

Anxiety disorders are the most prevalent psychiatric disorders among youths (Kessler et al., 2005) and have a major impact on their academic, social, and family functioning (Rapee et al., 2009). The emotional suffering and negative impact on daily life are partly explained by the maladaptive cognitive and behavioral strategies often used by these children. In particular, anxious children tend to overestimate the likelihood of threatening situations and underestimate their ability to cope with those situations (Beck, 2005). Consequently, these children avoid feared or anxiety-provoking situations or stimuli; in other words, behavioral avoidance is their most frequent behavioral coping strategy (Barrett et al., 1996; Chorpita et al., 1996; Vasey & Ollendick, 2000).

Behavioral Avoidance

Behavioral avoidance is a trademark of many emotional disorders, especially anxiety disorders (Barlow, 2002), and it is considered an important maintenance factor for many difficulties/psychopathologies (e.g., Chu et al., 2014). Avoidance can be active (e.g., escape behaviors or overt behaviors to prevent harm) or passive (e.g., staying in one’s comfort zone or staying away from situations or stimuli that might trigger anxiety) (Branchi & Ricceri, 2013). By avoiding or escaping from situations perceived as threatening, the child experiences short-term relief. However, avoidance prevents children from (1) learning how to deal with anxious feelings, (2) learning how to handle anxiety-evoking situations, and (3) gathering information that could help them disconfirm their beliefs about the potential threat of the situation and/or about their self-efficacy in dealing with it (Kendall et al., 2000; Vasey & Dadds, 2001). Therefore, by avoiding anxiety-evoking situations, anxiety is maintained and can even worsen. In addition, children often avoid situations that are important to them, which can significantly interfere with their daily lives (Vasey & Ollendick, 2000).

To break this vicious cycle, cognitive‒behavioral therapy (CBT) for anxiety disorders in children focuses on reducing dysfunctional cognitions (such as the overestimation of danger, which is targeted by cognitive restructuring) and on reducing behavioral avoidance (targeted by exposure techniques, in which a child exposes himself or herself to anxiety-evoking stimuli, preventing behavioral avoidance). Targeting behavioral avoidance is expected not only to decrease fear/anxiety (due to the direct effects of exposure - e.g. habituation - Groves & Thompson, 1970) but also to reduce the negative interference of fear/anxiety in the child’s life (because the child can face previously avoided situations/stimuli) and dysfunctional beliefs through the integration of corrective information and fear memory (pathological fear structure), which disconfirms the pathological, unrealistic associations, further decreasing fear (Foa & Kozak, 1986; Richard et al., 2007).

Assessment of Behavioral Avoidance

Given the importance of behavioral avoidance in maintaining children’s emotional disorders, it is critical to accurately assess it in research and clinical settings. However, the measures specifically developed to assess children’s behavioral avoidance are unexpectedly limited. Behavioral avoidance tests are the most common method of measuring avoidance. These tests involve asking the child to approach a feared object or situation and measuring the physical distance from that object or situation, which is used as an estimate of behavioral avoidance (Field et al., 2008; Klein et al., 2011). In addition, self- and other-report measures assessing children’s emotional disorders tend to focus on somatic and cognitive symptoms, with little or no reference to behavioral symptoms (particularly avoidance). Some measures include a few items related to avoidance but do not include separate avoidance scales [e.g., the Multidimensional Anxiety Scale for Children (MASC) (March et al., 1997); the Screen for Child Anxiety Related Emotional Disorders (SCARED) (Birmaher et al., 1997)]. In addition, the few measures that include behavioral avoidance subscales tend to be disorder-specific [e.g., the Social Anxiety Scale for Children–Revised (SASC-R) (La Greca & Stone, 1993); the Mobility Inventory for Agoraphobia (Chambless et al., 1985)] and do not allow a transdiagnostic assessment of avoidance across emotional disorders.

In response to this gap, Whiteside et al. (2013) developed a multi-informant questionnaire to assess childhood behavioral avoidance: the paired Child Avoidance Measure–Self Report (CAMS) and the Child Avoidance Measure–Parent Report (CAMP). The CAMS and CAMP each consist of 8 items that assess children’s behavioral avoidance of anxiety triggers from the perspective of the child and from the perspective of his or her parents, respectively. In the original validation study, exploratory and confirmatory factor analyses showed that both scales presented a one-factor structure in nonclinical and clinical samples. In addition, both scales exhibited good internal consistency (CAMS Cronbach’s α = 0.89 for the nonclinical sample and 0.88 for the clinical sample; CAMP Cronbach’s α = 0.90 for the nonclinical sample and 0.91 for the clinical sample) and adequate test-retest reliability (r = .56 for the CAMS and r = .75 for the CAMP in the nonclinical sample). Additionally, the construct validity of the CAMS and CAMP was supported by significant and positive correlations with preexisting measures of avoidance, anxiety symptoms, social anxiety, and disability (Whiteside et al., 2013).

Moreover, the CAMS and CAMP were also shown to discriminate between clinical and nonclinical participants. Evaluation of whether the measurement model of the CAMS and CAMP was consistent across ages (i.e., above and below 12.5 years) suggested that the imposition of a constraint that factor loadings be equal across age groups did not result in a significant reduction in model fit (Whiteside et al., 2013).

The Present Study

To our knowledge, no other study has evaluated the factor structure and psychometric properties of the CAMS and CAMP in cultures beyond that of the US. The present study aimed to fill this gap by examining the factorial structure and psychometric properties of the CAMS and CAMP in a Portuguese population, recruiting both clinical and nonclinical samples. The investigation of the psychometric properties of these scales in other cultures is particularly relevant given (1) the importance of behavioral avoidance for understanding and treatment of anxiety disorders in children and (2) the lack of alternative measures allowing transdiagnostic assessment of this construct, which limits research in this area and the comprehensive assessment of the effectiveness of CBT interventions.

Therefore, the first goal of this study was to confirm the factorial structure of the CAMS and CAMP in a clinical sample (children with anxiety-related disorders and their parents), as well as in a nonclinical sample of children and their parents. In both the clinical and nonclinical samples, we expect to confirm the one-factor structure found by Whiteside et al. (2013). Additionally, we aimed to assess the reliability of both scales and evaluate their construct validity by examining their correlations with the Revised Child Anxiety and Depression Scale (RCADS); we predicted these correlations would be positive and significant. We also aimed to assess the ability of the CAMS and CAMP to discriminate between clinical and nonclinical participants. Based on the original validation study (Whiteside et al., 2013), we expected CAMS and CAMP scores to be significantly higher in the clinical sample of anxious children than in the nonclinical sample of children of the same age. Finally, we aimed to assess the sensitivity of the CAMS and CAMP to changes after children participated in the Unified Protocol for the Transdiagnostic Treatment of Emotional Disorders in Children (UP-C; Ehrenreich-May et al., 2017a, b), Ehrenreich-May et al., 2017a group CBT intervention with a strong focus on exposure therapy. We expected a decrease in CAMS and CAMP scores after the UP-C intervention.

Materials and Methods

Participants

Clinical Samples

The clinical sample consisted of 172 parent‒child dyads. Children were 6–13 years old (M = 9.15, SD = 1.89; 58.7% female) and had a primary diagnosis of an anxiety disorder (e.g., social phobia, separation anxiety disorder, specific phobia) or anxiety-related disorder (e.g., illness anxiety disorder, obsessive-compulsive disorder, posttraumatic stress disorder). Their parents (89% mothers) were 28–59 years old (M = 40.83, SD = 5.03), and the majority were married (69.2%), employed (82.6%), had a university degree (57%), and had a family income between 800€ and 2000€ per month. Most families lived in urban areas (61.6%) from five districts in central Portugal.

Nonclinical Sample

The nonclinical sample of children comprised 288 children who were 8–13 years of age (M = 9.60, SD = 1.30; 52.3% female) and in third (31.6%), fourth (19.8%), fifth (22.9%), or sixth (25.7%) grade. The nonclinical sample of parents comprised 210 parents (93.3% mothers) who were 27–52 years of age (M = 40.59, SD = 4.68). The majority were married (60.5%), employed (83.8%), had a university degree (74.8%), and lived in urban areas (77.6%) in several districts of Portugal.

Procedure

This study was approved by the Ethics Committee of the Faculty of Psychology and Education Sciences of the University of Coimbra (FPCEUC), the Ethics Committee of the Centro Hospitalar Tondela -Viseu of a public hospital in central Portugal, and by the Board of Directors of the six collaborating schools. Parents provided written informed consent, and children provided their verbal assent to participate in the study.

Clinical Sample

Data were collected from the clinical sample between August 2020 and April 2022. Children and parents participated in a feasibility study (Caiado et al., 2022) and in a randomized controlled trial evaluating the efficacy of the UP-C (Ehrenreich-May et al., 2017a, b transdiagnostic group intervention for children with emotional disorders and their parents. To be included in the studies, children had to be between 6 and 13 years old and have a primary diagnosis of an anxiety, anxiety-related, or depressive disorder. Additionally, both children and their parents had to be fluent in Portuguese and free of any cognitive disability or learning difficulty that could prevent them from understanding and answering the questionnaires.

Children and parents were referred to participate in the study by (1) mental health professionals of the collaborating hospital; (2) school psychologists from the collaborating schools; or (3) parents’ self-registration on the project website, indicating their willingness to participate in the study. All parents and children were then contacted by the research team, who scheduled an interview to provide detailed information about the project, assess children’s eligibility to participate in the study, and collect parents’ informed consent. Children who met the eligibility criteria completed the research protocol in the presence of a researcher. Although children with a primary diagnosis of a depressive disorder were included in the feasibility and efficacy studies, in the current study, only children with a primary diagnosis of an anxiety or anxiety-related disorder were included. The exclusion of children with a primary diagnosis of a depressive disorder was due to their low number (i.e., only 13 children), which would not allow the generalization of the results to children with a primary diagnosis of a depressive disorder. Therefore, these 13 children were excluded from the current study. The data analyzed in this study pertain to the baseline assessment of children and parents (i.e., treatment assessment). To examine the scales’ sensitivity to change, only data from participants who benefited from the UP-C intervention and had complete pretreatment and posttreatment evaluation data were included. Of the 104 participants who received the UP-C (31 from the feasibility study and 73 from the randomized controlled trial), only 98 children and 96 parents completed the CAMS and CAMP at both assessment points (pretreatment and posttreatment); therefore, only these were included for this particular analysis.

Nonclinical Sample

Data were collected from the nonclinical sample of children between December 2019 and March 2020 in six public schools in central Portugal. Parents of 588 children (aged 8–13 years) received, through their children’s teachers, the informed consent form and a letter explaining the objectives and ethical issues of the study. Of these children, a total of 288 returned their parents’ informed consent and met the inclusion criteria of the study (i.e., aged 8–13 years, fluent in Portuguese, and without cognitive disabilities or learning difficulties that could prevent them from understanding and answering the questionnaires). Children completed the research protocol, which included the CAMS, in their classroom in the presence of a researcher.

An independent sample of parents participated in a study focused on mindful parenting during the COVID-19 pandemic. Parents were invited to participate in the study through social media websites, and those interested completed the survey online. Informed consent was obtained from all participants by clicking on the option “I declare that I agree to participate in the study titled ‘Mindful parenting in times of a pandemic’. I understand that my participation is voluntary and that I have the right to withdraw at any time. I understand that my responses are anonymous and that the results of the study will be treated collectively and used for research purposes only. I understand that if I have any doubts or questions, I can contact the research team through the contacts provided.” Enrollment in the study occurred between September and November 2020. In total, 210 parents consented to participate in the study and completed the CAMP.

Measures

Child Avoidance Measure–Self Report

The CAMS (Whiteside et al., 2013) is an 8-item self-response measure designed for children that assesses their behavioral avoidance when faced with stimuli that elicit anxiety, fear, or worry. The questionnaire begins with the following instruction: “We want to know more about your fears and worries. Some examples of things that make kids feel scared and worried are meeting new people, taking tests, being away from parents, and going near dogs, snakes, spiders, bugs, or germs. Now, we want to know what you do when you are afraid. Please circle the number that shows how often you do these things when you are afraid. When I feel scared or worried about something…”. Children are asked to rate how often they engage in each avoidant behavior (e.g., “I try not to go near it”; “I try not to think about it”; “I try to avoid it”) on a 4-point Likert scale that ranges from 0 (almost never) to 3 (almost always). The total score is the sum of all items, with higher scores indicating higher levels of behavioral avoidance.

Child Avoidance Measure–Parent Report

The CAMP (Whiteside et al., 2013) is the parent-report version of the CAMS. It also has eight items that assess children’s behavioral avoidance when faced with stimuli that elicit anxiety, fear, or worry. Similar to the CAMS, the questionnaire begins with the following instructions: “The next statements refer to how children sometimes react to things that make them scared or worried. Please circle the number that indicates how often your child reacts this way when he or she is around the main things that make him or her feel scared or worried. Think about how your child reacts in general to things that scare him or her, not just unusual or infrequent things. When my child is faced with something that makes him or her feel scared or worried (meeting new people, being away from parents, and going near dogs or germs)…”. Parents are asked to rate their own perception of how often their child engages in avoidance behaviors (e.g., “She or he asks whether she or he can do it later”; “She or he asks me to take care of it for him or her”; “She or he tries to avoid it”) on a 4-point Likert scale that ranges from 0 (almost never) to 3 (almost always). The total score is the sum of all items, with higher scores indicating higher levels of child behavioral avoidance according to their parent’s perspective.

The Revised Child Anxiety and Depression Scale

The RCADS (Chorpita et al., 2000) is a 47-item scale that measures children’s self-reported anxiety and depression symptoms. It comprises one subscale assessing low mood or depressive symptoms (e.g., “I feel sad or empty”) and five subscales corresponding to various anxiety disorders, including separation anxiety disorder (e.g., “I fear being away from my parents”), generalized anxiety disorder (e.g., “I worry that something bad will happen to me”), panic disorder (e.g., “I suddenly become dizzy or faint when there is no reason for this”), social phobia (e.g., “I worry what other people think of me”), and obsessive-compulsive disorder (e.g., “I have to do some things over and over again, like washing my hands, cleaning or putting things in a certain order”). It is also possible to compute a total anxiety score (the sum of the five anxiety subscales) and a total internalizing score (the sum of all six subscales). In the present study, only the total anxiety score and the five anxiety subscales were used. Items are rated on a 4-point Likert scale ranging from 0 (never) to 3 (always). Higher scores indicate more severe symptomatology. In the current study, Cronbach’s alpha coefficients were 0.92 for the anxiety scale, 0.73 for the separation anxiety disorder subscale, 0.76 for the obsessive-compulsive disorder subscale, 0.84 for the generalized anxiety disorder subscale, 0.84 for the panic disorder subscale, and 0.87 for the social phobia subscale.

Statistical Analyses

The statistical analyses were conducted in SPSS version 26.0 (IBM SPSS, Chicago, IL) and AMOS (IBM® SPSS® AMOS™ version 24.0; IBM Corporation, Meadville, PA, USA). Descriptive statistics (including of demographic information) were computed using SPSS. Missing values were handled through SPSS’s Missing Values Analysis (MVA). Little’s MCAR test was conducted to test whether the data were missing completely at random (MCAR). Nonsignificant results in this test were obtained for all questionnaires with missing values [CAMS, χ2(14) = 16.22, p = .30; CAMP, χ2(33) = 26.88, p = .77; RCADS, χ2(683) = 686.63, p = .45], indicating that MCAR may be inferred (Tabachnick & Fidell, 2007). Therefore, missing values were imputed via the expectation maximization technique (Little & Rubin, 1987).

Confirmatory factor analysis (CFA) using maximum likelihood estimation was performed to evaluate the model fit of the one-factor structure of the CAMS and CAMP in both samples. The following fit indices were used to assess model fit: the comparative fit index (CFI), the root-mean square error of approximation (RMSEA) and the standardized root-mean-square residual (SRMR). According to Hu and Bentler (1999), a two-index strategy can be used as a criterion for adequate model fit: SRMR < 0.09 and RMSEA < 0.06 or CFI > 0.95. Factor loadings of 0.32 or above were considered meaningful (Tabachnick & Fidell, 2007).

SPSS was also used to calculate Cronbach’s alpha coefficients for the CAMS and CAMP in both samples, with values above 0.70 indicating good reliability (Nunnally, 1978). The value of Cronbach’s alpha coefficients if one item was deleted was analyzed. Item-total correlations were also computed and were considered adequate when above 0.30 (Ferketich, 1991).

Pearson correlation analyses between the CAMS and CAMP and of the CAMS and CAMP with the RCADS subscales were performed to explore the validity of the CAMS and CAMP. To compare the CAMS/CAMP mean scores between the clinical and nonclinical samples, an independent-sample t test was used. Paired-sample t tests were used to examine the sensitivity to change of the CAMS/CAMP. Cohen’s d (Cohen, 1998) was used to estimate the magnitude of group differences. Effect sizes of 0.20, 0.50, and 0.80 were taken to indicate small, medium, and large effects, respectively.

Results

CAMS and CAMP Factor Structure

Clinical Sample

The single-factor model of the CAMS showed poor model fit, χ2(20) = 68.37, p < .001; CFI = 0.91; RMSEA = 0.12; SRMR = 0.06, although all standardized factor loadings were significant (p < .001) and above 0.32. An examination of the modification indices suggested that the errors belonging to Items 4 (“I ask if I can do something else”) and 5 (“I try to avoid it”) and those belonging to 4 and 6 (“I refuse to do it”) might be correlated (Byrne, 2010). The respecified model had adequate fit, χ2(18) = 44.12, p = .001; CFI = 0.95; RMSEA = 0.09; SRMR = 0.05. All standardized factor loadings were significant (p < .001) and above 0.32 (see Fig. 1).

Fig. 1
figure 1

CFA model of the CAMS and standardized factor loadings for the clinical sample

* p < .05; *** p < .001

Full size image

The single-factor model of the CAMP showed poor model fit, χ2(20) = 76.68, p < .001; CFI = 0.93; RMSEA = 0.13; SRMR = 0.04, although all standardized factor loadings were significant (p < .001) and above 0.32. An examination of the modification indices suggested that the errors belonging to Items 1 (“She or he tries to move away from it”) and 2 (“She or he asks whether she or he can do it later”) and those belonging to Items 1 and 3 (“She or he tries not to go places where it is”) might be correlated (Byrne, 2010). The respecified model had adequate fit, χ2(18) = 48.94, p < .001; CFI = 0.96; RMSEA = 0.10; SRMR = 0.03. All standardized factor loadings were significant (p < .001) and above 0.32 (see Fig. 2).

Fig. 2
figure 2

CFA model of the CAMP and standardized factor loadings for the clinical sample

* p < .05; *** p < .001

Full size image

Nonclinical Sample

The single-factor model of the CAMS showed acceptable model fit, χ2(20) = 61.78, p < .001; CFI = 0.95; RMSEA = 0.09; SRMR = 0.04, and all standardized factor loadings were significant (p < .001) and above 0.32. Nevertheless, an examination of the modification indices suggested that the errors belonging to Items 7 (“I think it is best to stay away from it”) and 8 (“I try to stay away from it”) and those belonging to Items 6 (“I refuse to do it”) and 7 (Byrne, 2010) might be correlated. The respecified model had good model fit, χ2(18) = 28.64, p = .05; CFI = 0.99; RMSEA = 0.05; SRMR = 0.03. All standardized factor loadings were significant (p < .001) and above 0.32 (see Fig. 3).

Fig. 3
figure 3

CFA model of the CAMS and standardized factor loadings for the nonclinical sample

* p < .05; *** p < .001

Full size image

The single-factor model of the CAMP had good model fit, χ2(20) = 56.45, p < .001; CFI = 0.97; RMSEA = 0.09; SRMR = 0.04, and all standardized factor loadings were significant (p < .001) and above 0.32 (see Fig. 4).

Fig. 4
figure 4

CFA model of the CAMP and standardized factor loadings for nonclinical sample

* p < .05; *** p < .001

Full size image

Reliability

Clinical Sample

Both the CAMS and CAMP presented adequate reliability (Cronbach’s α coefficients: CAMS = 0.85, CAMP = 0.90). As presented in Table 1, the item-total correlations for the CAMS ranged from 0.34 (Item 2) to 0.78 (Item 8), and the item-total correlations for the CAMP ranged from 0.42 (Item 2) to 0.81 (Item 3 and Item 6). Item 2 appeared to reduce the internal consistency of both scales, as Cronbach’s alpha increased when this item was removed.

Table 1 Item-total correlations and Cronbach’s alpha values with one item removed
Full size table

Nonclinical Sample

The CAMS and CAMP also demonstrated adequate reliability in a nonclinical sample of children and parents (Cronbach’s α coefficients: CAMS = 0.84, CAMP = 0.92). The item-total correlations for the CAMS ranged from 0.45 (Item 2) to 0.69 (Item 7). The item-total correlations for the CAMP ranged from 0.52 (Item 2) to 0.83 (Item 6). As in the clinical sample, Item 2 appeared to reduce the internal consistency of the CAMS, as the Cronbach’s alpha of this scale increased when this item was removed (Table 1).

Correlations

Clinical Sample

As presented in Table 2, the CAMS and CAMP scores exhibited a significant positive correlation with moderate strength. The CAMS was strongly and significantly correlated with the overall score of children’s anxiety symptoms (RCADS total anxiety score) and weakly to moderately correlated with all anxiety disorders assessed by the RCADS subscales. The CAMP, on the other hand, was only significantly correlated with the separation anxiety disorder subscale, and this correlation was weak.

Table 2 Matrix of correlations among study variables in the clinical sample
Full size table

Differences in CAMS and CAMP Scores Between Clinical and Nonclinical Samples

As shown in Table 3, the mean CAMS and CAMP scores were significantly higher in the clinical sample of anxious children and parents than in the nonclinical sample, with medium to large effect sizes.

Table 3 Differences in CAMS and CAMP scores between clinical and nonclinical samples
Full size table

Sensitivity of the CAMS and CAMP to Changes After the UP-C Intervention

As presented in Table 4, in the clinical sample that received treatment for emotional disorders focused on exposure therapy (i.e., the UP-C intervention), the mean CAMS and CAMP scores decreased significantly from pretreatment to posttreatment, with a large effect size. Of note, the CAMS and CAMP scores at posttreatment also presented good internal consistency, with Cronbach’s alpha values of 0.89 and 0.91, respectively.

Table 4 Differences in CAMS and CAMP scores between pretreatment and posttreatment
Full size table

Discussion

Behavioral avoidance is a core component in the understanding and treatment of anxiety-related disorders. It is a critical maintenance factor of these disorders; while avoidance provides temporary relief, the fear is reinforced through such behavior, and dysfunctional beliefs are not refuted (e.g., Barlow, 2002; Kendall et al., 2000; Mowrer, 1947). Additionally, behavioral avoidance leads to significant impairment in children’s daily lives, as they avoid activities or situations that might be important for their socioemotional development (e.g., Vasey & Ollendick, 2000). In line with this finding, the exposure technique is one of the most important and effective components of CBT for emotional disorders. Therefore, the availability of reliable measures to assess behavioral avoidance is critical not only to conduct future studies to deepen our understanding of the transdiagnostic mechanisms underlying emotional disorders but also to assess the efficacy of the exposure components in CBT interventions (namely, compared with other components, such as cognitive restructuring and mindfulness).

The CAMS and the CAMP are two eight-item parallel questionnaires that assess the behavioral avoidance of children from the child and parent perspectives, respectively. The current study aimed to determine the factor structure and psychometric properties of the CAMS and CAMP in clinical and nonclinical samples of Portuguese children and parents, thereby contributing to the study of the CAMS/CAMP in a culture in which it has not previously been studied, specifically a European culture, and providing Portuguese clinicians and researchers with a reliable and valid measure to assess children’s behavioral avoidance.

As expected, the one-factor structure of the CAMS and CAMP proposed by Whiteside et al. (2013) was confirmed in both clinical and nonclinical samples. Additionally, as expected, in both samples, the CAMS and CAMP presented good internal consistency and item-total correlations above 0.30. Almost all the CAMS and CAMP items seem to contribute to the scales’ internal consistency, except for Item 2 on the CAMS (“I try not to think about it”), which slightly reduced the Cronbach’s alpha of the scale. In addition, Item 2 had the lowest item-total correlation. One possible explanation is that all the other items are more strongly related to behavioral avoidance (e.g., “I try not to go near it”), whereas Item 2 appears to be more strongly related to cognitive avoidance. Nevertheless, to retain the original factor structure of the CAMS, to enable comparison with other studies that employed the 8-item scale and considering that the correlation between Item 2 and the total score was above 0.30, we decided to retain this item in the scale (Ferketich, 1991).

The construct validity of the CAMS and CAMP was analyzed in the clinical sample by examining their associations with a measure of anxiety. As expected, higher levels of children’s behavioral avoidance in the CAMS were associated with higher levels of child’s anxiety symptoms (RCADS total anxiety score) and with higher scores of separation anxiety disorder, generalized anxiety disorder, panic disorder, social phobia, and obsessive-compulsive disorder (assessed by the RCADS subscales). These results are consistent with previous findings (Vasey & Ollendick, 2000; Whiteside et al., 2013), as well as with theoretical and clinical perspectives, according to which children with higher anxiety levels are expected to engage in more avoidance behaviors.

Contrary to the expectations and results of the original validation study (Whiteside et al., 2013), the CAMP was only significantly correlated with the separation anxiety disorder subscale, and this correlation was weak. This surprising result can be explained by the fact that parents might be more aware of avoidance behaviors in the context of the relationship with their children, as separation anxiety is the anxiety disorder that has the largest manifestations and impact on the parent‒child relationship. Most anxiety disorders are internalizing disorders, which can make it difficult for parents to recognize them and the associated avoidance behaviors. Additionally, research has shown that parents of anxious children are more likely to be overprotective (e.g., Clarke et al., 2013), which may make them less likely to notice their children’s avoidance, thus potentially normalizing it in some way and discouraging the child’s independence and courageous behavior. These findings have important clinical and research implications, as they suggest that children with anxiety disorders can recognize their avoidant behaviors more effectively than their parents. In addition, these findings highlight the importance of providing psychoeducation and intervention for parents during the treatment of anxious children. However, it is also important to note that the CAMS and RCADS are both completed by children; therefore, higher correlations are expected between these measures than between the CAMP (parent report) and RCADS (child report).

Additionally, higher levels of children’s behavioral avoidance from the child’s perspective were significantly associated with higher levels from the parent’s perspective. However, this association was moderate, implying that there was only moderate agreement between the child’s and parent’s perspectives regarding the intensity of the child’s behavioral avoidance, which is consistent with the findings regarding the correlations between the CAMP and RCADS.

Additionally, as expected, the levels of children’s behavioral avoidance, according to both child and parent reports, were significantly higher in the clinical sample of anxious children than in the nonclinical sample of children and parents, with medium to large effect sizes. These findings may indicate that the CAMS and CAMP can discriminate between clinical and nonclinical populations.

Finally, and as expected, the CAMS and CAMP scores decreased significantly, with a large effect size, after an exposure-focused CBT intervention (the UP-C), demonstrating that these scales are sensitive to therapeutic change. This result is consistent with numerous studies that concluded that exposure-focused CBT is effective for treating emotional disturbances and reducing behavioral avoidance (e.g., Richard et al., 2007; Schopf et al., 2020), as is the UP-C (UP-C; Ehrenreich-May et al., 2017a, b).

Clinical and Research Implications and Future Studies

The present study has important clinical and research implications. First, this study can provide Portuguese clinicians and researchers with a valid instrument for assessing behavioral avoidance in children, which is a core aspect of many emotional disorders. Since behavioral avoidance is an observable variable, its assessment through the CAMS and CAMP may allow a more objective evaluation of the efficacy of psychological or psychopharmacological interventions, particularly their ability to reduce children’s levels of behavioral avoidance. Furthermore, assessing behavioral avoidance may provide a better understanding of its role in the development and maintenance of youth emotional disorders, as well as illuminating the mechanisms of change underpinning CBT.

Second, since the CAMS/CAMP provides a transdiagnostic assessment of behavioral avoidance, it can be applied to a wide range of disorders or difficulties, making this measure particularly useful. In light of the increased interest in transdiagnostic therapies (e.g., Brown & Barlow, 2009; Ehrenreich et al., 2009), it is critical to have a psychometrically robust measure for assessing this construct. Moreover, in clinical settings, the CAMS/CAMP can be a useful instrument for determining children’s level of behavioral avoidance and, consequently, defining an appropriate therapeutic plan (e.g., gradual exposure). Furthermore, because of the discriminative capacity of this instrument, it may enable better identification of children at risk of developing emotional disorders. Finally, the CAMS/CAMP comprises only eight easy-to-understand items, which can facilitate its use in real-world clinical settings.

Because no other studies have been conducted to evaluate the factor structure and psychometric properties of the CAMS and CAMP, this study significantly contributes to the literature by supporting the psychometric robustness of both scales and by including clinical and nonclinical samples of children.

Despite these contributions, this study also has some limitations that should be noted and addressed in future studies. First, only children with anxiety or anxiety-related disorders were included in this study. Because behavioral avoidance is thought to be a transdiagnostic mechanism underlying emotional disorders, in future studies, it would be interesting to investigate the associations of the CAMS and CAMP scores with specific internalizing psychopathologies, including depression, which would require the recruitment of a larger sample of children with a variety of internalizing psychopathologies. Second, in the present study, the sample may not have been fully representative of the Portuguese population, as it was collected in only the central region of the country. Third, because the clinical sample of children was recruited in a therapeutic setting in which a change in behavioral avoidance was expected and because participants in the nonclinical sample answered the questionnaires only once, the test-retest reliability of the CAMS and CAMP could not be examined. Finally, the convergent validity of the scales was not assessed using other self- or other-report measures or subscales of behavioral avoidance.

Conclusions

In the present study, both the CAMS and CAMP exhibited a one-factor structure and adequate psychometric properties, enabling a multi-informant assessment of children’s behavioral avoidance in clinical and nonclinical samples of Portuguese children. This study’s contributions to research and clinical practice are particularly important because of the scarcity of other measures for transdiagnostic assessment of behavioral avoidance and the lack of other studies conducted in Europe to investigate the factor structure and psychometric properties of the CAMS and CAMP. Moreover, the CAMS and CAMP were able to discriminate between clinical and nonclinical children and were sensitive to therapeutic change. As a result, the CAMS and CAMP can be used to identify children at risk of developing emotional disorders, to examine the efficacy of psychological interventions in reducing behavioral avoidance and to study the role of behavioral avoidance in emotional disorders.