Early Detection of Behavioral and Emotional Problems in School-Aged Children and Adolescents: The Parent Questionnaires

1.1. The Importance of Early Detection.

It is increasingly clear that mental health is largely based on appropriate life experience and conduct and on any corrective actions that must take place early in the development of the individual. The awareness of this fact makes itincreasingly necessary for a civil society to guarantee conditions of mental hygiene (“primary prevention”), which aims to avoid or correct all the factors and elements that can negatively affect mental health, and of a “secondary prevention”, which aims to identify early and adequately treat those who present symptoms of a disorder.

The identification phase of secondary prevention is strongly based on diagnostic screening, for which the most practical place to implement is the school. Diagnostic screenings in school are strongly recommended [1-4].

For an early identification in school age, the most suitable tools appear to be those that involve parents in asking for information on the emotions and behavior of their children. Teachers may be helpful in reporting externalizing behaviors, but they can hardly detect internalizing problems unless they are very relevant. Children at least 8 years old and adolescents can answer appropriate self-administered questionnaires, which are reliable, within certain limits, for the internalizing disorders, but not for the externalizing ones. Obviously, the use of the three sources (“multi-informant assessment”) gives the possibility of having more reliable data, even if further work is needed to evaluate the differences found.

1.2. The Questionnaires For Parents

Parent questionnaires are probably the most used tools. They can be used both in the clinic as a first or further data collection to be used in subsequent evaluations, and for epidemiological studies. They can also be used by the pediatrician, who can have them filled out by the parents to assess the possible existence of problems such as to recommend the intervention of a pedopsychiatrist.

The questionnaires can be “targeted”, that is specific for a certain problem (e.g. anxiety disorders or a type of anxiety disorder such as social anxiety) or “broad-band”, that is aimed at exploring at least the most frequent and predominant clinical conditions. In this mini-review, we focus on broad-band instruments, indeed aiming to underline the importance of an extensive exploration of psychopathology, in order to give the clinician as much information as possible on the problems presented by the child-adolescent.

In their formulation, the questionnaires can present the items grouped according to the clinical problems to be explored or distributed randomly. This second case makes a direct evaluation of the results impractical, thus requiring the use of a correction grid.

Like all diagnostic tools, these questionnaires must be “validated”. This term refers to different procedures, of which the most important, indeed fundamental, is the ability to differentiate the subject with pathology from one without. It is also important that the instrument is able to identify the type of pathology (e.g. depressive) and not simply indicate in a generic way that the subject may have a pathology. In the validation procedure, the comparison with another well-validated instrument is frequently used. However, this comparative validation between instruments (“concurrent validity”) cannot be a guarantee of the ability to identify a clinical state of pathology. It is necessary that the results provided by the instrument are compared with the definitive clinical data, that is the clinical diagnose(s) obtained through appropriate multiple evaluations. The closer the results are to the clinical data, evaluated as “sensitivity”, “specificity”, “accuracy” and ROC evaluation, the greater the clinical validity of the instrument. The other types of validation are elements that do not add much to the clinical criterion validity.

1.3. Purpose of the Review

This review aims to analyze those questionnaires for parents of school-aged subjects, who explore behavioral and emotional problems on a wide range; only those for which there are adequate validation data, are taken into consideration. The analysis concerns the composition of the questionnaires, in relation to the problem areas explored, the number of items and their clinical relevance, and assesses their different characteristics and the results of comparative studies among them.

3.1. Characteristics of Some Parent Questionnaires

The characteristic elements of the broadband, structured and validated parent questionnaires are described below, following the order of first publication. A summary is shown in Table 1.

3.1.1. Pediatric Symptom Checklist (PSC)

PSC [4-6] is a 35-item questionnaire for subjects 4 years old and above. PSC gives a total problem score, with a cut-off score of 28 as a criterion for a disorder; however, further studies identified cutoff scores ranging from 12 to 24.

In a screening on school-age children, Jellinek et al. [7] found PSC having a specificity of 0.68 and a sensitivity of 0.95. Several children, whose pediatricians' ratings had indicated adequate functioning, were identified by the PSC as having substantial psychosocial dysfunction and requiring further evaluation.

To overcome the limit of a global assessment for the presence of pathology, subsequently, the authors indicated 3 groups of items (see the website of the Massachusetts General Hospital): 1.attention problem (5 items), 2.internalizing (5 items), 3.externalizing (7 items) subscales, being the other 18 items unattributed. The relevant 17 items have been proposed as Pediatric Symptom Checklist-17 (PSC-17). The validation study of PSC-17 [8] using ROC procedures, established a cut-off of 15 (on a maximum score of 34).

Mean sensitivity of PSC resulted .75 (95% CI .08) and mean specificity .88 (95% CI .04) on the basis of 28 studies evaluated by Lavigne et al. [9, 10].

PSQ is available online, free of charge.

3.1.2. Child Behavior Check-List (CBCL)

This parent questionnaire [10] includes 113 (+7) questions related to behavioral problems of children and adolescents 6-18 years. It is by far the most widely used (over 2300 citations in PubMed).

Initially, on the basis of factor analysis, the CBCL was subdivided into 8 syndromic scales: 1. Anxious/Depressed, 2. Withdrawn/Depressed, 3. Somatic Complaints (all 3 summed as Internalizing disorders), 4. Social Problems, 5. Thought Problems, 6. Attention Problems, 7. Rule-Breaking Behavior, 8. Aggressive behavior (7 and 8 summed as Externalizing disorders), plus one scale as “other problems”. A “dysregulation profile” (CBCL-DP) has been proposed [11] for the case of simultaneous extreme values on the syndrome scales 1. Anxious/Depressed, 6. Attention Problems, and 8. Aggressive Behavior. The CBCCL-DP has been suggested to be associated with disruptive behavior disorders, suicidal behavior, and reduced need for sleep. However, according to Deutz et al. [12], it only “reflects a broad syndrome of dysregulation that exists in addition to specific syndromes of emotional symptoms, conduct problems, and hyperactivity-inattention”.

An obsessive-compulsive scale (CBCL-OCD) with 8 items has also been proposed [13]; however, among the 8 items only 3 result specific of OCD according to the DSM (item 9: Can’t get his/her mind off certain thoughts - obsessions; item 31: Feels he/she might think or do something bad; item 66: Repeats certain acts over and over - compulsions), the other referring to anxiety or various though problems.

The original syndromic scales of CBCL, derived from the first-factor analysis, do not correspond to the current classification of psychopathological groups; moreover, several items do not appropriately refer to the indicated syndromic area. This led to a new grouping of items under 6 scales [14] consistent with the DSM-IV-TR definitions [15] and valid also for the DSM-5 [16].

The six DSM-oriented scales of CBCL comprise 55 of the total 113 items, selected for their correspondence to the symptoms indicated by the DSM-IV (still valid for the DSM-5) as belonging to the following diagnostic groups: affective (= depressive; 13 items), anxiety (6), somatic (7), attention-deficit hyperactivity (7), oppositional-defiant (5) and conduct (17) problems.

Note that the “somatic problem” scale is not an index of a “Somatic symptom disorder”. This is not identified by the number of somatic symptoms (which gives the score at the scale), but by the excessive focus on the symptoms, which causes emotional distress and anxiety. Therefore, the clinical hallmark of the “Somatic symptom disorder” is a focalized anxiety, and not the presence of a greater or lesser number of somatic symptoms. Therefore, the somatic scale does not correspond to a clinical disorder, and consequently, the actual DSM-oriented scales are 5, including 48 items.

Validity and reliability of the DSM-oriented scales have been documented [14]. Normative data for children and adolescents 6 to 18years old are available and the kit that is sold includes the ability to get T scores by entering raw data into a computer program. Two age groups are distinct: 6-11 and 12- 18 years old.

Mean sensitivity of CBCL resulted in .63 (95% CI .08) and mean specificity .84 (95% CI .06) on the basis of 22 studies evaluated by Lavigne et al. [9].

There is a version for teacher (TRF), in which about 20 CBCL items have been modified.

The use of CBCL is covered by copyright.

3.1.3. Strengths and Difficulties Questionnaire (SDQ)

The SDQ parent version [17], used also for teachers, is a brief screening questionnaire that can be completed by the parents of children-adolescents aged 4 and older. The scale asks for 25 attributes, a few positive and others negative. The 25 items are divided between 5 scales, the first 4 exploring 1) emotional symptoms (5 items), 2) conduct problems (5 items), 3) hyperactivity/inattention (5 items), 4) peer relationship problems (5 items); added together, the 20 items generate a total “difficulty” score. Further, 5 items constitute the “prosocial behaviour” score (“strength” score).

The SDQ identified over 70% of individuals with conduct, hyperactivity, depressive and some anxiety disorders, according to Goodman et al. [18]. SDQ scores above the 90th percentile predicted a substantially raised probability of independently diagnosed psychiatric disorders [19]. In relation to the specific problem of ADHD, the predictive validity of the SDQ was satisfactory [20].

A dysregulation profile of the SDQ (SDQ-DP), like the CBCL-DP, according to Deutz et al. [12] has the limitation already reported under the heading CBCL.

The mean sensitivity of SDQ resulted in .65 (95% CI .08) and mean specificity .76 (95% CI .07) on the basis of 32 studies evaluated by Lavigne et al. [9].

German, Finnish and Dutch versions of SDQ are validated. The parent-SDQ (and teacher-rated SDQ) resulted valid and reliable for different ethnic groups within the Dutch society, however, with differences in reliability and validity of the subscales, which makes its interpretation difficult for certain ethnic groups [21].

The use of the SDQ is free, available online.

3.1.4. Child and Adolescent Symptom Inventory 4 (CASI-4) Parent Checklist

The CASI-4R brings together the CSI-4, usable for children 5 and 12 years old and contains 97 items, and ASI-4, for teenagers 12-18 years containing 120 items [22-24]. The CASI-4R Parent Checklist for subjects 5 to 18 years contains 142 items.

Since 2013, updated according to the new nosographic changes made from the DSM-IV to the DSM-5, there is the CASI-5 version, that includes all of the items from the CASI-4R (http://www.checkmateplus.com/product/casi5.htm).

The CASI-4&5 are 4-points scales. Items are grouped in relation to the psychopathological areas they explore. The symptoms of the following disorders are assessed: ADHD, oppositional defiant disorder, conduct disorder, generalized anxiety disorder, social phobia, separation anxiety disorder, major depressive episode, manic episode, dysthymic disorder, schizophrenia, autistic/Asperger’s disorder, anorexia, and bulimia. One or two key symptoms of each of the following disorders are also included: obsessive-compulsive disorder, specific phobia, panic attack, motor tics, vocal tics, and substance use.

Reliability, convergent and discriminant validity, and clinical utility (intended as sensitivity and specificity versus DICA-P interview diagnoses) were confirmed by Sprafkin et al. [25].

There is also a Teacher Checklist (CASI-4R-TC) containing 105 items.

The use of CASI-4&5 is covered by copyright.

3.1.5. Behavior Assessment System for Children-2&3 (BASC-2&3)

This “system” is a comprehensive set of rating scales that in addition to Parent Rating Scale (PRS) includes a Teacher Rating Scales (TRS, with 100 to 139 items), a Self-Report of Personality (SRP, completed by the child or adolescent) and a Student Observation System [26].

The BASC-PRS has three versions with varying numbers of items for different age groups (preschool, 134 items; child, 160 items; adolescent, 150 items); it uses a 4-point scale.

It includes the following scales: Activities of Daily Living, Adaptability, Aggression, Anxiety, Attention Problems, Atypicality, Conduct Problems, Depression, Functional Communication, Hyperactivity, Leadership, Learning Problems, Social Skills, Somatization, Study Skills, and Withdrawal.

According to Kamphaus & Frick [27], the BASC-PRS exhibits good correlations with analogous scales from other parent rating scales. In Korean children, Song et al. [28] found BASC-PRS valid for measuring developmental psychopathology.

The BASC series is covered by copyright.

3.1.6. Child and Adolescent Behavior Inventory (CABI)

The CABI [29, 30] questionnaire consists of 75 questions to parents/caregivers. These explore a wide range of problem areas: somatic, anxiety, phobias, obsessive-compulsive, insecurity, depression, irritability, oppositional-defiant, conduct, impulsivity, hyperactivity, attention deficit, reality evaluation, social relationships, sphincter control, bulimia, anorexia, sex interest, smoking, alcohol and substance abuse, school performance and being bullied.

In the CABI, some problems that can belong to two or more disorders are grouped separately: “sleep problems”, located among somatic symptoms, according to DSM-5 can be part of both depression and generalized anxiety; “irritability”, held as a separate subscale, can be part of depression, generalized anxiety and oppositional defiant disorder. In this way, the pedopsychiatrist more correctly assesses their clinical significance.

Psychometrics properties, including internal consistency, factor analysis, normative data together with a comparative and clinical criterion evaluation on a small number of cases are reported by Cianchetti et al. [29, 30]. Predictive validity for the clinical diagnosis on 462 subjects has been recently published [31]. The normative data are different in relation to gender and age, and 3 age groups are distinct: 6-10, 11-13 and 14-18 years old.

The use of the CABI is free, available on Cianchetti et al. [29] or by direct request to the author.

3.2. Studies Comparing The Parent Questionnaires

A comparison between the above-described questionnaires was carried out almost exclusively between CBCL and the others, probably because CBCL is largely the most widespread and therefore has been taken as the main reference tool. To make it easier for the reader to evaluate the differences between the questionnaires, the items of CBCL, CABI and SDQ are compared with the DSM-5 diagnostic criteria in Tables (2-5).

3.2.1. PSC

The PSC as a total score was effective in identifying subjects with psychosocial problems taking as reference the CBCL total assessment scores [32, 33]. Similar results were obtained with a simultaneous comparison with total scores of CBCL and of SDQ by Vogels et al. [34]. The reduced form PSC-17 was compared with the total scores of CBCL, showing excellent classification accuracy [35]. It should be noted that these are data relating to the total score, therefore indicating the presence of problems, without specification on their type.

3.2.2. CBCL

As indicated above, see each of the other instruments.

3.2.3. SDQ

Scores from the SDQ and CBCL have been reported as highly correlated and equally able to discriminate psychiatric from non-psychiatric cases, with the SDQ significantly better than the CBCL at detecting inattention and hyperactivity, and at least as good at detecting internalising and externalising problems [36].

Comparable diagnostic accuracy of SDQ and CBCL in detecting common emotional and behavioral disorders was also found by Kovacs & Sharp [37].

In a study by Sheldrick et al. [38], where total scale values were considered, the SDQ showed a mean positive predictive value around 77%, against CBCL around 61%, and mean sensitivity around 19%, against CBCL around 38%. In the study by Kuhn et al. [39], SDQ and DAWBA adequately predicted the presence of an ICD-10 disorder.

The Dutch version of the SDQ, similar to the English and German versions, has equal validity as the Dutch ASEBA for screening children, according to Janssens & Deboutte [40].

3.2.4. CASI-4, CSI/ASI-4

The ratings of the parent version of the Child Symptom Inventory (CSI-4) converged and diverged in a theoretically consistent pattern with respective scales of the CBCL and the Diagnostic Interview for Children and Adolescents-Revised-Parent Version (DICA-P), and boys with specific DICA-P diagnoses received significantly higher corresponding CSI-4 parent symptom ratings than boys not so diagnosed [25].

3.2.5. BASC

The validity of the BASC-PRS has been found comparable to that of the CBCL/4-18 for assessing childhood ADHD and disruptive behavior [41], while according to Ostrander et al. [42] for distinguishing ADHD students from non-ADHD students the BASC model was more parsimonious and accurate than the CBCL. While the correlations between similarly named scales on the BASC-2 and CBCL were nonsignificant, according to Jacola et al. [43], BASC-2 and CBCL were not statistically different from each other in sensitivity to change of youth treatment outcome [44]. More recently, Gabrielli et al. [45] stated that the BASC-2 PRS, when compared to the CBCL, consistently performed well as a measure of behavioral outcome in the assessment of youth in foster care.

3.2.6. CABI

The comparison of the CABI with CBCL had the purpose of verifying their degree of predictivity towards the final clinical diagnosis, i.e., the ability to discriminate the pathological from the non-pathological, which is the basic objective of these tools. The CBCL, as mentioned above, it is able to provide data only in relation to 5 areas of clinical diagnosis (the 6th, “somatic symptoms” is not a clinically autonomous area, a said above in the section “CBCL”). For this reason, even if the CABI explores a much broader range of pathologies, the comparison has taken place in relation to these 5 areas, i.e., 1. affective problems (a term used by CBCL to indicate “depression”), 2. anxiety, 3. ADHD, 4. ODD and 5. CD. A study on 462 subjects [29] found that the accuracy values (probability of correct classification) were high for both instruments, and significantly better for CABI anxiety and ADHD scales, and for CBCL ODD and CD scales; no significant difference was found for depression scales. The areas under the curve of the receiver operating characteristic analysis confirmed anxiety and ADHD scales of the CABI having a better predictive ability than those of the CBCL, with not statistical differences between the other scales.

4.1. Considerations for the Use of Tools

The parent questionnaires above described obtained the general validation criteria for the proposed use. The choice of the questionnaire to be used must, therefore, be based on the different characteristics.

1. The first point should be the number of emotional and behavioral problems for which the questionnaire can give the clinician information on their presence or not in the examined subject. The clinician must decide whether it is sufficient for him to have information on the main externalizing and internalizing conditions, in which case he can use a short tool like SDQ and PSC. Instead, if he requires a predictive orientation extended to many more clinical problems, he must use one of the other tools. He must keep in mind that the CBCL gives information only on 5 psychopathological areas (depression, anxiety, ADHD, ODD, CD), even if these are the most frequent, while BASC, CABI and CASI give information on almost all the psychopathological areas and this it is important because screening should allow wide-ranging explorations.

2. Another point is the commitment required by the parent to complete the questionnaire, which depends on the number of items. It is likely that in a screening in schools, which parents do not always access very willingly, requesting answers to a high number of items is a disincentive for a careful and correct compilation. BASC and CASI, and CBCL to a lesser extent, are therefore disadvantaged in this condition.

If, on the other hand, the compilation is proposed to a parent who has asked for a consultation for the child, he is certainly interested in responding scrupulously and will commit himself to a high number of items, even if it will be easier for him to deal with an intermediate number.

It should also be considered how many items are really useful for assessing the presence of the clinical disorder. For example, the CBCL has 113 items (+ 7 somatic), but only 48 explore the pathologies that can be identified by the questionnaire; therefore, the parent is called to respond on a high number of items that the clinician is not able to use in relation to the pathology.

3. In cases of use of the questionnaires as a preliminary, in association with the clinical examination, the grouping of the items according to the psychopathological areas, as happens in the CABI and the CASI, is an advantage for the examiner. Before facing the direct interview with the parent, with a glance at the answers he can become aware of the problems in relation to which he will have to deepen the interview.

Furthermore, the parent, who comes to consult for a specific problem, faced with a wide series of questions on different situations, will be able to realize problems that he had not spontaneously detected; thus, signaling them, he will help the clinician not to miss any comorbidities, which are frequent in developmental psychopathology.

4. A final criterion for the choice, not insignificant in case of very wide use as for the screening, is the cost that involves the use of those tools that are covered by copyright, like BASC, CASI and CBCL (Table 1).

CONCLUSION

As observed, the tools available are not few, but each has its strengths and weaknesses. The next commitment should be to perfect the existing ones in order to reach the choice of a more limited number of questionnaires that allow a wide exploration.

To facilitate the use in preparation for the clinical visit, it would be advisable to group the items according to the problems explored.

Last but not least, in the interest of the whole community, it is desirable that the tools be freely available to all psychiatrists

Summary Points

• Early detection of behavioral and emotional problems in children and adolescents is relevant.
• During the school years, it can avail of the collaboration of parents through questionnaires, a series of these being described here.
• Different parent questionnaires are examined, whose characteristics are outlined in relation to the accuracy of the information they provide for identifying undiagnosed pathologies.
• Parent questionnaires are also a useful source of preliminary information for the clinical interview.
• For the choice of the instrument, the clinician will evaluate, in addition to the possible costs, the ability to explore and identify a large number of problematic conditions in the most accurate way.