How the Affective Reactivity Index (ARI) works for teachers as informants

Highlights

• Researchers and clinicians need sound instruments to assess irritability in several contexts.

• The Affective Reactivity Index (ARI) was originally designed for youth and parents.

• The use of the ARI answered by the teacher is supported by good psychometric properties.

• Teachers may inform of the difficulties children have in expressing irritability in school.

• The teacher's view can be useful when planning treatment by helping to identify treatment targets.

Abstract

Background

The Affective Reactivity Index (ARI) is a brief instrument originally designed as a self- and parent report. However, the view of teachers, who can observe social situations that may give rise to irritability, is relevant. The goal is to provide the measurement qualities of the ARI score as reported by teachers.

Method

Children formed part of a longitudinal study on behavior problems in Barcelona (Spain) and they were assessed when they were 7 (N = 471) and 11 years old (N = 454) with questionnaires about psychopathology, anger and aggressive behavior, and a diagnostic interview answered by the parents, youths and teachers. Confirmatory factor analysis, measurement invariance, reliability and validity were studied for the ARI answered by teachers.

Results

The 6-item, 1-factor model fitted well. Almost full metric invariance and partial scalar invariance was obtained across sex and over age. The ARI scores largely converged with other teacher-reported measures of anger and irritability, and with other measures of psychopathology, aggressive behavior, and callous-unemotional traits at a medium level. The associations with parent's measures were medium to low, and very low for child self-reported measures. The ARI scores significantly differentiated children with and without psychopathology and functional impairment, both cross-sectionally and longitudinally.

Limitations

Only one child self-report measure of irritability included. Limited internal consistency of some scale scores. Findings are mostly generalizable to Spanish children.

Conclusions

ARI could be a suitable instrument for measuring irritability as reported by teachers. The teacher's view can be useful when planning treatment by helping to identify treatment targets.

Introduction

DSM-5 (American Psychiatric Association, 2013a) has shifted from a purely categorical conceptualization of mental disorders to incorporate the notion that psychopathology occurs along dimensional continua. To this effect, the DSM-5 task force and some working groups have developed dimensional measures that evaluate the presence and severity of symptoms that cut across diagnostic boundaries and may inform clinical decision making (Clarke and Kuhl, 2014). One of these cross-cutting symptoms is irritability. Irritability is a shared symptom in different disorders, such as anxiety disorders, depressive disorders, oppositional defiant disorder (ODD), bipolar disorder, post-traumatic stress disorder, and disruptive mood dysregulation disorder, among others (Toohey and DiGiuseppe, 2017). Irritability can be defined as an elevated proneness to anger relative to peers at the same development level (Brotman et al., 2017; Stringaris et al., 2018; Vidal-Ribas et al., 2016) and is characterized by easy annoyance, a low or decreased threshold for frustration, touchiness, and anger/temper outbursts. Irritability is a frequent reason for mental health referral and predicts negative outcomes from childhood to adulthood (Copeland et al., 2014; Ezpeleta et al., 2016). Meta-analyses report that irritability is associated with future depression, anxiety problems, and ODD (Vidal-Ribas et al., 2016). Irritability also predicts ADHD (Shaw et al., 2014), comorbidity (internalizing and externalizing), difficulties with peers (Ezpeleta et al., 2016), functional impairment (Wiggins et al., 2018), suicide (Benarous et al., 2018), and, in adulthood, adverse health, educational and social outcomes (Copeland et al., 2014). There is a consequent need to measure irritability (see Toohey and DiGiuseppe, 2017, for a review).

The Affective Reactivity Index (ARI; Stringaris et al., 2012) is the instrument proposed by the American Psychiatric Association (2013b) to assess the cross-cutting symptom of irritability. The ARI contains 6 irritability symptom items plus one impairment item that must be scored if present during the previous six months. According to Stringaris et al. (2012), the items cover (a) the threshold for an angry reaction, (b) the frequency, and (c) the duration of the feelings/behaviors. The questionnaire was designed to be answered by youth (6–17 years old) and their parents.

Supplementary material Table 1 synthesizes the current available research on the psychometric properties of the ARI. The questionnaire has been studied in several countries (UK, US, Brazil, Australia and China) with samples of children/youth aged 5–19 years. Most of the studies have used the child and parent version, but Mulraney et al. (2014b) also studied an adult version. Regarding reliability, internal consistency is in the high range (alpha between 0.84 and 0.90 for youth, between 0.80 and 0.92 for parents and 0.80 for the adult version score). Test-retest reliability is acceptable for youth (ICC: 0.66) and good for parents (0.82) (Pan and Yeh, 2018) and adults (0.80) (Mulraney et al., 2014b). Parent-child agreement ranges between 0.42 (Pan and Yeh, 2018) and 0.73 (Stringaris et al., 2012). All the developed versions show that there is consistency in the answers to the items and the scores are more stable for parents. Regarding validity, confirmatory factor analyses has obtained a one-factor good fit for parents (Mulraney et al., 2014b; Stringaris et al., 2012) and the fit was acceptable for youth in one study (DeSousa et al., 2013) but not in two others (Mulraney et al., 2014b; Stringaris et al., 2012); the adult version has also shown poor fit. Parent and youth ARI scores differentiate control groups from groups with severe mood disorders, bipolar disorders, ODD (only parent) (Pan and Yeh, 2018; Stringaris et al., 2012), and DSM-IV diagnoses (Mulraney et al., 2014a). ARI scores converge to a large extent with other irritability measures (Pan and Yeh, 2018) and within a small to medium range with measures of psychopathology, impairment, and adaptation (Mulraney et al., 2014b; Pan and Yeh, 2018). Peer problems for the parent version and prosocial problems for the youth version of the Strengths and Difficulties Questionnaire (SDQ) are not associated (Mulraney et al., 2014b; Stringaris et al., 2012). Cross-sectionally, the youth and parent ARI scores have been associated with SDQ emotional problems, conduct problems, and hyperactivity (Mulraney et al., 2014a, 2014b; Stringaris et al., 2012). No study has made longitudinal predictions. The validity results indicate that the ARI is related to outcomes that have been shown to be related to irritability and that ARI scores can differentiate groups with different psychopathologies. In summary, these results indicate not only a broad use of the questionnaire internationally, but also an added interest in extending the initial use to adults. In all the versions the ARI presents good psychometric properties.

However, none of the previous studies have tested how the ARI scores work when answered by teachers. Discrepancies between informants are the norm in child psychopathology assessment (De los Reyes and Kazdin, 2005), highlighting the need to obtain information from various contexts (De Los Reyes, 2011). Regarding ODD, non-equivalence across parents and teachers has been found for some items of its dimensions (one of which was irritability), as it has been observed that parents tend to rate ODD behaviors as more frequent than teachers (Ezpeleta and Penelo, 2015). Teachers are familiar with children's normative development and may be the best reporter of peer and social relations and behaviors at school (Konold and Pianta, 2007). Teachers can observe the child in social situations that may give rise to irritability and its expression. For some disorders, the diagnosis system may require the presence of the symptoms in more than one setting and teachers may inform whether the symptoms are present in the school context (Evans et al., 2016). In this line, the information provided by the teacher may help the clinician to identify pervasive and severe symptoms that occur in several contexts and require special attention, which could potentially facilitate treatment planning and boost treatment efficacy (De Los Reyes et al., 2015). The goal of the study was to test the psychometric properties of the ARI answered by teachers. Since appropriate and proper comparison of a construct between groups and across times depends first on ensuring equivalence of meaning of the construct (e.g., Putnick and Bornstein, 2016), invariance of the measurement model across sex and over measurement occasions was analyzed. Based on the position of the teacher as observer in the school context, we expected high convergence between the ARI scores and other constructs related to irritability, such as aggressive behavior, psychopathology, functional impairment, and other measures of irritability. In line with the literature, we expected the same-informant values to be higher than the cross-informant ones. We also expected higher irritability scores to be associated with worse outcomes cross-sectionally and longitudinally, also differentiating between groups with and without problems. This is the first study to report on ARI predictive validity longitudinally.