Future Directions: The Phenomenology of Irritable Mood and Outbursts: Hang Together or Hang Separately¹

ABSTRACT

Recognition of the importance of irritable mood and outbursts has been increasing over the past several decades. This “Future Directions” aims to develop a set of recommendations for future research emphasizing that irritable mood and outbursts “hang together,” but have important distinctions and thus also need to “hang separately.” Outbursts that are the outcome of irritable mood may be quite different from outbursts that are the trigger or driving force that make youth and his/her environment miserable. What, then, is the relation between irritable mood and outbursts? As the field currently stands, we not only cannot answer this question, but we may also lack the tools to effectively do so. Here, we will propose recommendations for understanding the phenomenology of irritable mood and outbursts so that more directed and clinically useful assessment tools can be designed. We discuss the transdiagnostic and treatment implications that relate to improvements in measurement. We describe the need to do more than repurpose our current assessment tools, specifically interviews and rating scales, which were designed for different purposes. The future directions of the study and treatment of irritable mood and outbursts will require, among others, using universally accepted nomenclature, supporting the development of tools to measure the characteristics of each irritable mood and outbursts, understanding the effects of question order, informant, development and longitudinal course, and studying the ways in which outbursts and irritable mood respond to treatment.

During a psychiatric evaluation for severe outbursts, 10-year-old Jack was asked why he thought he lost his temper so often. His response was to ask about the boiling point of water. When told it was 212 degrees, he said “well, I feel like I’m always at 211 degrees. It doesn’t take much for me to boil over.”

Introduction

The general definition of irritability is a quickness or proneness to anger. Anger itself is an internal, negative affective state occurring in response to a perceived threat or a response to frustration. We might never know someone is angry unless it shows in some way as when the person loses his or her temper and has an outburst. If the person does not recover quickly from the outburst, we deem it pathological especially if the outburst is vicious verbally or if it becomes physical in some way at which point, we speak of reactive aggression. Although aggression is defined as occurring with the intent to cause harm, reactive aggression is said to occur in the context of anger without a specific goal in mind (Leibenluft & Stoddard, 2013; Toohey and DiGiuseppe, 2017).

Jack’s definition was insightful. His simmering irritability was not recognized. It was his boiling over that prompted referral for treatment. His classmate Jill, on the other hand, often started the day happy, but if she perceived something to be unfair, or it did not go quite the way she anticipated, she would explode, getting into trouble with her teacher and alienating peers. She recognized her “little mistake” (her term) too late and remained annoyed and irritable for the rest of the day so that Mom knew there was a problem as soon as she got off the school bus.

Irritability and its manifestations are responsible for many mental health referrals and emergency department visits (Copeland et al., 2013; Evans et al., 2022; Farquharson et al., 2023; Stringaris et al., 2018). In fact, it is the Jacks, Jills and others that the authors encounter daily that are at the heart of this “Future Directions.” That is, unlike some of the excellent recent articles on ways forward to study irritability in general (e.g. Beauchaine & Tackett, 2020; Brotman et al., 2017; Evans et al., 2022; Leibenluft et al., in press), or those focused specifically on treatment directions (e.g. Breaux et al., 2023; Brotman & Kircanski, 2022), this is a pragmatic discussion of how to move forward to best identify, understand, and ultimately treat the young people who embody irritability. We begin with a discussion of the language needed to communicate clearly about irritability and provide a history of what has complicated information acquisition about irritability. The bulk of the discussion centers on how to better understand phenomenology including course of illness, measures, effects of informants and comorbidity and ends with a discussion of treatment needs focused especially on the behavioral manifestations of irritability including triggers to outbursts and damage control.

Should Irritable Mood and Outbursts “Hang Together or Hang Separately”?

Clearly, everyone is irritable sometimes. Like many psychiatric (and other) symptoms, irritability is normally distributed (Beauchaine & Tackett, 2020). What makes irritability pathological is the threshold at which a reaction takes place, the frequency with which the manifestations occur, how bad the reaction gets, and how long it lasts. Those parameters determine impairment, help-seeking, and outcome. Finally, irritability is a symptom of many conditions, and predicts impairment over and above those conditions, though it is unclear if its phenomenology varies with those conditions (Klein et al., 2021; Toohey and DiGiuseppe, 2017).

That irritability has both mood and behavior aspects has become increasingly clear (Burke et al., 2014; Carlson, Singh, et al., 2023; Moore et al., 2019; Silver et al., 2022; Silver et al., in press; Vidal-Ribas et al., 2016). The roots of irritability may start early, beginning sometimes as early as infancy/preschool. Even these early manifestations significantly predict later internalizing and externalizing symptoms and diagnoses (Finlay-Jones et al., 2024; Momany and Troutman, 2021; Sorcher et al. 2022).

Where internalizing and externalizing components of irritability have been studied separately, data reveal that irritable mood or what some have called “tonic irritability” predicts subsequent internalizing disorders like depression and anxiety (Cardinale et al., 2021; Silver et al., 2021; Silver et al., 2023), whereas the behavioral component, which has been called “phasic irritability,” temper loss, or outbursts, predicts later externalizing psychopathology in children (Silver et al., 2023) and adolescents (Hawes et al., 2020; Silver et al., 2021).

Clearly, then, irritability’s subjective internalizing (irritable mood) and observable externalizing (outbursts) aspects were deemed important enough to be given scientific names: tonic and phasic irritability at a 2014 NIMH workshop (Copeland et al., 2015). The tonic component was defined as “a persistently angry, grumpy, or grouchy mood;” phasic irritability was conceptualized as “behavioral outbursts of intense anger.” Although these are important concepts to identify and distinguish from one another (Carlson & Klein, 2018; Carlson, Singh, et al., 2023; Copeland et al., 2015; Klein et al., 2021), these names potentially obscure the symptoms and behaviors they represent, confusing rather than clarifying the phenomena. Adding further to the obfuscation, the dimensions of “tonic” and “phasic” were initially meant to codify the temporal course of irritability when the issue had been distinguishing episodic course as seen, for instance in mood disorders as “phasic,” and chronic course (“tonic”) representing the children with long-standing, non-episodic mood and outburst behaviors (Leibenluft, 2011).

Recommendation #1: Let’s keep our vocabulary clear. We suggest that face valid terms be used like irritable or angry mood instead of “tonic irritability” as it reflects mood and outbursts instead of “phasic irritability.”

The terms “tonic and phasic irritability” sound scientific but it is then necessary to translate back to what is really meant. The terms also create yet another literature for similar concepts. “Tonic irritability” is simply irritable/angry mood. “Phasic” confuses episodes with outbursts. Let us use the term “outbursts” instead. The definition of outbursts would be the same as described in disruptive mood dysregulation disorder (DMDD): developmentally and situationally inappropriate outbursts manifested verbally by verbal rages and/or behaviorally by physical aggression toward people and/or property (American Psychiatric Association, Diagnostic and Statistical Manual 5^th edition, DSM 5, p. 156).

Consider R Codes

Until we have a better means for doing so, diagnostic criteria provide a working hypothesis by which to establish with some reliability the frequency of a specific disorder in the general child, adolescent, and adult population. However, we also need a way to identify and code irritable mood and outbursts without invoking disorders that children do not have (Carlson, Farquharson, et al., 2023; Carlson, Singh, et al., 2023; Fristad, 2021). Non-psychiatric physicians and professional coders responsible for billing, make liberal use of ICD 10–11 “R codes” for important symptoms and laboratory values that describe a patient’s reason for seeking help. R codes may be used in addition to a diagnosis (which in psychiatry are the “F” codes we use) or stand alone. There are “R codes” (R40-R46) which identify signs and symptoms involving cognition, perception, emotional state and behavior. R 45.4 is the billable code for “irritability and anger.” Since tantrums/outbursts lack their own code, the American Psychiatric Association’s (APA) DSM-5 Revision Subcommittee Co-Chair, Michael First, MD has suggested using R 45.89 for coding outbursts, “mental symptoms not otherwise specified” until the National Committee on Health Statistics (NCHS) is persuaded that outbursts have their own code (Michael First, MD, personal communication, 7/22/2022). The NCHS coding committee felt that R codes for hostility (R 45.5), violent behavior (R 46.6.) and homicidal ideation (R 45.85) were good enough though we contend they are not the same as emotionally impairing, angry outbursts or tantrums. Coding has implications for research conducted on the electronic medical record, where it is often difficult to identify the behaviors leading to referral and hospitalization. As such, outbursts are overlooked (Carlson, Farquharson, et al., 2023).

Recommendation 2: Add R codes to the child’s existing diagnoses to modify them and highlight the irritable mood or impairing emotional outbursts which are the reasons for seeking help. Convince the National Committee on Health Statistics’ classification committee to accord outbursts their own code. Until then we suggest the best-fitting R code or use R 45.89.

In the presence of comorbid conditions (e.g., for children with attention deficit hyperactivity disorder (ADHD), oppositional defiant disorder (ODD), etc.), the R code would provide more accurate information regarding why the child was referred for evaluation and treatment Carlson, Farquharson, et al. (2023).

Unintended Consequences of Inconsistent or Ill-Defined Terminology

Irritability (angry mood and outbursts) is a symptom like fever or pain that plays different roles in different disorders. It is an explicit criterion in a major depressive and manic episode, generalized anxiety disorder, and post-traumatic stress disorder. It is a defining characteristic (i.e. loss of temper) in oppositional defiant disorder (ODD), as “irritability” and outbursts in disruptive mood dysregulation disorder (DMDD), and as “outbursts” in intermittent explosive disorder (IED). Irritability is an important symptom without even being identified officially in the criteria in attention deficit hyperactivity disorder (ADHD) and autism spectrum disorder (ASD). Irritability thus occurs in internalizing, externalizing, neurodevelopmental, psychotic and stress disorders, as well as disorders or conditions caused by medical or neurological conditions (Toohey and DiGiuseppe, 2017).

Diagnosis is the mechanism we use to help us formulate and communicate about how to categorize significant defining symptoms and behaviors and determine what might be explanatory. Since 1980, the process has been to identify a clinically important condition, then develop criteria, then develop interviews (and rating scales) with acceptable reliability to establish the condition’s presence, absence and severity. Whether DMDD proves to be a valid condition or not, it has provided a way to examine at least some aspects of irritability in clinical settings. DMDD, however, emerged as the unintended consequences of four unrelated Diagnostic and Statistical Manual (DSM) decisions:

1. Emotion dysregulation (including irritability), once an important symptom of ADHD (Laufer & Denhoff, 1957) was excluded from the core ADHD criteria in DSMIII because of feared lack of reliability and specificity (personal communication, D.P. Cantwell, MD 1/15/1996). As a result, questions about volatile emotions were excluded from systematic interviews for ADHD so that when we encountered hyperactivity, impulsivity and severe irritability/outbursts (which were chronic) “prepubertal mania” appeared to be the only place to diagnose them. Emotion regulation and irritability have finally been re-recognized as important symptoms of ADHD (Barkley, 2010; Faraone et al., 2019; Karalunas et al., 2019; Shaw et al., 2014).

2. Not including symptoms of aggression that are mood/anger related (i.e. reactive aggression) or assuming oppositional defiant disorder (ODD) would suffice as a diagnosis has made this important dimension diagnostically homeless (Carlson, Singh, et al., 2023; Stepanova et al., 2023). ODD addresses only loss of temper and behaviors that are “noxious and negativistic,” not impulsively or reactively aggressive (Burke et al., 2021).

3. In important research interviews, ratings of thoughts and feelings were merged with behaviors. “Feeling very irritable or quite angry OR has frequent homicidal thoughts or thoughts of hurting others OR throws and breaks things around the house,” are all nested in a single rating of severe irritability in one diagnosis module, depression. The response to depression irritability was then used for other disorders (like mania) where irritability was a symptom.

4. Until DSM-5 (American Psychiatric Association [APA], 2013), manic episodes were poorly operationalized (i.e. no onset and offset were specified) so it was difficult to distinguish episodic from chronic irritability.

The result of these decisions was that mood disorders, specifically mania, became the disorder in which hyperactivity, impulsivity, distractibility, and chronic irritability could occur in one condition (e.g. Mick et al., 2005). Hence, the bipolar controversy unfolded between severe ADHD with emotion dysregulation and prepubertal bipolar disorder (Biederman & Jellinek, 1998; Carlson & Klein, 2014, 2018 for review;). Severe mood dysregulation (SMD; Leibenluft et al., 2003) was developed to better distinguish chronic and episodic severe irritability. Its DSM-5 successor, DMDD (Leibenluft, 2011) emerged to try to solve the problem of attributing outbursts to bipolar disorder.

Recommendation #3: Re-conceptualizing and re-wording criteria and interviews require reestablishing reliability and field-testing criteria wording by a large group of professionals (e.g. Leibenluft et al., in press), raters who do the interviews, and patients who are interviewed to ensure everyone understands the questions the same way and that responses mean the same thing to interviewers and respondents.

Changing criteria occurs for a variety of reasons like trying to decrease response burden or, in the case of the emotionality in ADHD, deciding it lacks reliability and specificity or because those in charge have strong feelings or wish to include one particular line of research. This has unintended consequences that defeat the purpose for which criteria and their operationalization were intended. The PROMIS methodology (Ader, 2007), discussed later provides some guidance for how criteria changes should proceed.

Which Comes First, Irritable Mood or Outbursts

Lack of clarity over the relationship between irritable mood and outburst behavior is underscored by the fact that we have yet to decide where DMDD belongs. Does it belong in the DSM-5 mood disorders section, where it has been placed? Or is it a behavior disorder as classified by ICD-11 as a form of irritable oppositional defiance (Evans et al., 2017)? The answer could be “it depends.” If proneness to anger identifies an internalized mood state, one might not even realize a person is feeling angry until there is an obvious manifestation of it. It is also possible to be angry without loss of temper. The high correlation between tonic and phasic irritability in both community and clinical samples [e.g. between r = 0.60 and r = 0.82 (Silver et al., 2021, 2023; Degroot et al., in press)] suggests that as children become more easily annoyed and angry, they are more likely to lose their temper, have an outburst and have more serious outbursts. The concepts clearly “hang together.” The fact that loss of temper and outbursts are not synonymous is evident in the finding that 20% of a community sample of 6-year-olds was identified as having a significant loss of temper with only 55.4% having actual tantrums though 78% were also considered irritable by their parents. In contrast, in a clinic sample of 6-year-olds, half often lost their tempers and 92% had tantrums which led to referral but only 24% were considered irritable (Carlson et al., 2016). This suggests that feeling in an angry/irritable mood, expressing anger, loss of temper and tantrums/outbursts need to be defined, distinguished and organized so we can predict who of those with angry moods follow through to temper loss and outbursts and who does not. They do “hang separately.”

Also unclear is that irritable mood between outbursts (the DMDD definition of chronic irritable mood) may occur either because the mood is truly chronically negative (presumably like Jack) or because the child is fine until stressed but re-regulates so slowly that the next triggers boil them over quickly (like Jill).

Furthermore, irritable mood and outbursts have not been consistently elicited. The Great Smoky Mountains study analyzed extant data by first selecting youths with outbursts/”phasic irritability” not irritable mood. They reported that across childhood, “tonic” and “phasic” irritability often co-occurred (63%). However, while outbursts could occur alone (in 24% of youth reports), bad mood/”tonic irritability” rarely did (4% with only irritability; Copeland et al., 2015). In other words, outbursts usually occurred in irritable children and adolescents but could occur with a normal mood 24% of the time. However, simply being frequently angry or grouchy without outbursts was rare.

On the other hand, Laporte et al. (2021) used the Development and Well Being Assessment (DAWBA; Goodman et al., 2000) which asks about irritable mood first. In their community sample of 10–12 year olds, 45% had both irritable mood and outbursts, 21.3% had outbursts only but 33.7% had irritable mood only. Perhaps by starting with irritable mood, they found higher rates than those reported by Copeland et al. (2015). If the Copeland study had started with irritable mood, would the findings have agreed? We do not know.

Finally, a serious limitation of our knowledge base accumulated over the past 20 years is that we have relied on data collected for different purposes. Irritability as a symptom was taken initially from interview depression modules. ODD scales or interview modules have provided symptoms of temper loss (not necessarily outbursts) and anger. Irritable mood and temper loss have been assembled post hoc to construct irritability and DMDD. We thus cannot honestly answer the question of whether and when chronic irritability/anger leads to outbursts or whether outbursts are independent of chronic irritable mood. Different studies either address the question differently or ignore it.

Recommendation #4: We have two suggestions. In existing studies of mood and outbursts, compare results with mood elicited first (like LaPorte et al., 2021) versus where outbursts are ascertained first (like Copeland et al., 2015). Second, in future studies, ascertain irritability both independent of a diagnosis and within relevant disorders clarifying if the outbursts occurred while the mood was grouchy or the outbursts occurred first and then the mood was grouchy (because the child got into trouble for instance).

There are precedents for understanding how behaviors unfold. In psychosis, there are diagnostic differences between those whose psychosis appears at the peak or depth of their manic or depressive episode versus psychosis that starts first and is followed by depression. The same is true for a suicide attempt that occurs during depression and one that occurs impulsively. Ecological Momentary Assessment has been used to unpack the relationship between frustration and mood changes (Naim et al., 2021; Tseng et al., 2023). The strategy could also be used to examine the relationship between mood and outbursts.

Premature Skip Outs and Subsyndromal Symptoms

The DAWBA (Goodman et al., 2000) is the best comprehensive interview to ascertain irritability symptoms. It begins with a question about irritable mood and then asks in detail about outbursts when the child is angry (rather than outbursts and then mood). However, it progresses only if irritability or outbursts occur more than once or twice a week. We thus lose the opportunity to learn about less frequent but equally destructive outbursts. Severe symptoms resulting in the police being called once a month are certainly worth ascertaining and we cannot know about them with a priori decisions about what is important. That further raises the question of how to handle subsyndromal (including less frequent) irritable behaviors.

Recommendation #5: We need to be careful about “skip outs” in interviews until we know what we are missing when we skip. We also need a way to highlight clinically symptomatic and impaired youth who miss a criterion and suggest an “other specified” designation for DMDD.

Interviews paired with inventories that first screen for outbursts and the circumstances surrounding them, like the Emotional Outburst Inventory (EMO-I; Carlson et al., 2022) would lighten interview burden. The EMO-I collects information from parents about outburst behaviors, their frequency, duration, episodicity and manifestations. The interviewer would have the information in advance and decide whether, for example, monthly outbursts, or episodic irritability that need physical restraint or daily outbursts of swearing are worth following up. Item Response Theory approaches with data collected with the entire range of outburst frequencies and severities might replace current intuition-based approaches for skip-outs.

The term “other specified” has been attached to various disorders in DSM to identify those who do not quite meet full criteria and serves the useful purpose of providing diagnostic coverage and honesty. Currently, we do not have an “other specified” code for DMDD. The aforementioned R codes would also provide a way of noting significant symptoms that do not meet criteria for a given disorder.

Defining the Characteristics of Irritability and Outbursts: Onset, Offset, Course of Illness

Whether chronic irritability exists as a discrete disorder, as in DMDD, or a dimension with clinically significant ramifications, we need to understand the initial appearance of symptoms that culminate in the full-fledged manifestations (onset), and how they resolve (offset). Infants can be fussy, reactive and irritable, called temperamental negative affectivity which can be lifelong (e.g. Beauchaine & Tackett, 2020; Finlay-Jones et al., 2024; Momany and Troutman, 2021). Irritability may be a developmental delay in emotion regulation (e.g. the child who “never outgrew the terrible two’s”) that first starts in preschool (Dougherty et al., 2013; Dougherty et al., 2015; Silver et al., 2022) and has enduring effects (Caspi & Silva, 1995). We have noted that early childhood tantrums are likely developmental. The question is whether later onset irritability is developmental in nature (i.e. tantrums), or psychopathology-related (e.g. anxiety, mood, psychosis drug- or stress-related). That is, do younger children’s outbursts differ structurally or in terms of outcome than older age onset irritability, or outbursts (Potegal et al., 2009).

While there are studies of predictors of outcome (e.g. Dougherty et al., 2013, 2015; Sorcher et al., 2022; Silver et al., 2023), and even informant variance (Kessel et al., 2021) associated with different ages, we are unaware of specific studies relating clinical features and outcome to age at onset. Unfortunately, absent information about age of onset, we cannot use existing data to answer this question. There is a precedent, however, in that there are certainly different implications of age of onset for conditions like bipolar disorder (Carlson & Pataki, 2016), schizophrenia (Driver et al., 2020), and conduct disorder (e.g. Moffitt et al., 1996) that might be true for irritability/DMDD as well.

The concept of offset or cessation of irritability is relevant to episodic versus tonic irritability. An episode has a fairly discreet start and end; chronic conditions are less well defined. At what point irritability (or DMDD) is considered remitted has not been established but that defines the distal end of an episode.

Several longitudinal studies of irritability have identified subgroups within a given cohort that have several unique trajectories, two of which are stable i.e. a low irritability subgroup of about 60-to 80%, and a high irritability persistent subgroup of about 5–10% of the sample (e.g. Ezpeleta et al., 2016; Reef et al., 2010; Riglin et al., 2019; Yu et al. 2023). The high irritability persistent group may be more like children in clinical samples. However, there are also groups whose irritability symptom trajectory is unstable, either worsening or improving. Differences in those subgroups are also informative. Irritability as a developmental dimension generally seems to decline over time (Copeland et al., 2015); correlations at different ages vary from r = 0.29 to 0.88 (Vidal-Ribas et al, 2016) with cortical maturation. This is why “developmentally inappropriate” is a useful diagnostic qualifier. Also, ADHD polygenic risk scores are associated with increasing adolescent irritability and severe, stable symptom patterns while depression risk scores are associated only with increasing adolescent irritability (Riglin et al., 2019).

Given different longitudinal trajectories, it is not surprising that DMDD itself has been considered unstable (Vidal-Ribas et al., 2016) i.e. children with the disorder at one follow-up do not have it at the next (Axelson et al., 2012; Brotman et al., 2006; Dougherty et al., 2016; Mayes et al., 2015). Furthermore, a closer look at longitudinal studies reveals that the irritability problems persisted even if the diagnosis itself did not. Deveney et al. (2015) found only 39.5% of children with DMDD still met criteria at 4-year follow-up but 37.2% had sub-threshold symptoms. A similar sample after 3 years showed only 21.1% still met DMDD criteria but 57.9% still had tantrums (Mulraney et al., 2021). In both cases, patients were still irritable and impaired even when the diagnostic criteria for DMDD were no longer met (the youth was subsyndromal), i.e. missed a criterion and thus did not have the diagnosis. This relates to how one defines “offset.” A decline of outburst frequency from several times a week to several times a month may reflect improvement but not necessarily remission or diagnosis change. Conversely, it is possible for mood to improve but outbursts to remain. We hope that does not mean a child goes from having DMDD to having Intermittent Explosive Disorder (IED).

Recommendation # 6: Until proven irrelevant, we need to ascertain age of onset of significant irritable mood and outbursts and define offset. This is best obtained by interview and separate from a specific diagnosis.

Unfortunately, the KSADS-PL (Kaufman et al., 1997), the most frequently used interview for childhood mental health disorders, ascertains the worst episode of a mood disorder, not first episode and asks about age of onset only to determine if onset was before age 10, a DMDD criterion. The PAPA/CAPA (Angold et al., 1999; Angold & Costello, 2000) asks about irritable mood in the depression module and temper tantrums in the ODD section and the items are assembled for DMDD after the fact. Age of onset is not addressed. The DAWBA (Goodman et al., 2000) asks about mood and outbursts together but again, does not ask about age at onset.

Onset and offset are important to the concept of chronic versus episodic irritability and are at the heart of the creation of DMDD and distinguishing it from mania. It is also important to know if irritability onset is acute, tracking with mood disorders, or more gradual, traveling with neurodevelopmental disorders like ADHD and autism spectrum disorders. There is a precedent for defining remission and recurrence in mood disorders. The 1988 MacArthur Foundation-funded Depression Task Force defined remission from a depressive episode whereby the person must be virtually asymptomatic [(i.e. fewer than 2 symptoms and to only a mild degree) for at least 2 months (Frank et al., 1991)]. Similar criteria could be established for DMDD.

Understanding Outcomes

In DSM-5, DMDD is considered a mood disorder because longitudinal studies have found depression and anxiety in adult follow-up samples. However, when factors describing irritable mood and outbursts are separated and traced longitudinally, they have different outcomes (e.g. Silver et al., 2023; Sorcher et al., 2022) and predict both internalizing and externalizing outcomes. We find that irritability predicts depression and anxiety because those conditions are ascertained in both child and adult interviews (Vidal-Ribas et al. 2016). Irritability predicts ODD when the child/adolescent interview asks about it. However, where follow-up studies into adulthood have used adult psychiatric interviews like the Structured Clinical Interview for DSM IV (SCID, Spitzer et al., 1992) or the Young Adult Psychiatric Assessment (YAPA, Angold et al., 1999), it was not possible to ascertain if conditions like ADHD, ODD or conduct disorder were still present or predicted because those conditions were not part of those follow-up interviews.

Interview methodology also changes with respondent age from an integrated assessment of parent and child informants in youth to one that relies only on the young adult. Barkley et al. (2002) reported years ago that young adults may not be reliable informants about their ADHD and ODD symptoms. We truly do not yet know what happens to mood AND outburst symptoms. Externalizing disorders may have been unrecognized, underdiagnosed or in partial remission, even when queried.

Where information on mood and outbursts was ascertained in adults [The World Health Organization’s Composite International Diagnostic Interview (CIDI, Kessler et al., 2004)], Liu et al. (2021) found rates of tonic (65.3%) and phasic (63.8%) irritability in unipolar MDD which were related to greater severity, chronicity and younger age of onset. Fava et al. (2010) found that depressed adults with lifetime irritability (consisting of anger attacks), compared to those with non-irritable depression, had higher rates of comorbid of anxiety disorders and impulse control disorders (ADHD, ODD, and IED). Absent data specifically linking childhood irritability and irritability in adults with depression, where a comprehensive adult interview was used, data are suggestive of a similar parallel relationship between irritability and internalizing and externalizing disorders.

Recommendation #7: To truly understand the evolution of childhood psychopathology, adult psychiatric interviews should routinely incorporate questions about childhood psychopathology. Follow-up studies of irritability should incorporate methodology that includes other informants.

Baseline predictors of adult outcomes of childhood irritability conditions are clearly important, but we also need to know if DMDD and its externalizing comorbidities persist, attenuate, or disappear in adulthood. Prospectively tracking signs of improvement, persistence or progression are critical for outcome prediction as are identifying patterns that are waxing and waning, comorbid, and each of their responses to intervention.

Defining the Characteristics of Irritable Mood and Outbursts – the Role of Rating Scales

Rating scales for emotion dysregulation and related concepts abound (Althoff & Ametti, 2021; Freitag et al., 2023; Mazefsky, Conner, et al., 2021). Those used specifically to identify irritability include Achenbach and Rescorla (2001) Child Behavior Checklist (CBCL), Teacher Report Form (TRF) and the Youth Self Report (YSR). Irritability-focused items include “sullen, stubborn, irritable,” “hot temper/tantrums” and may or may not include “argues a lot” and “sudden changes in mood” (Aebi et al., 2013; Evans et al., 2020). However, these items conflate different concepts as well as having only one item each for irritable mood and outbursts.

The Affective Reactivity Index (ARI; Stringaris et al. 2012) is a 6-item questionnaire specifically designed to measure irritable mood and although it asks if the child loses his/her temper, it does not fully address outbursts. It will be useful to determine if it can function as a treatment outcome measure.

Several informative rating scales use the PROMIS guidelines (Ader, 2007; Cella, 2007) which combine what Achenbach (2020) calls the “top down” approach (taking advantage of items that are already known) and the “bottom-up” strategy (making de novo observations of the phenomenon being rated with input from knowledgeable stakeholders rather than just rearranging other rating scale items). Wakschlag et al. (2012) used this combined approach to develop the Multidimensional Assessment of Preschool Disruptive Behavior (MAP-DB), a comprehensive scale of temper loss that describes not only the behavioral components of a tantrum but also identifies potential triggers, anger regulation (trouble calming down), context (where and why the tantrum occurs) and irritability (becomes frustrated easily, has a short fuse). The MAP-BD, now called the Multidimensional Assessment Profile Temper Loss Scales, or MAPS-TL, has been extended to identify parsimonious screening items for infants/toddlers (Wiggins et al., 2021; Wiggins, Ureña Rosario, Zhang et al., 2023; Wiggins, Ureña Rosario, MacNeill et al., 2023), preschool (Wiggins et al., 2018), early elementary school children (Hirsch et al., 2023), preadolescents (Alam et al., 2023), and adolescents (Kirk et al., 2023).

Mazefsky et al. (2018) also followed PROMIS guidelines to develop the Emotion Dysregulation Inventory (EDI), initially to describe emotion dysregulation in children on the autism spectrum but also normed it in typically developing children (Conner et al., 2021). Both the EDI and a recently published scale called the PROMIS EC Measure of Emotional Distress (Sherlock et al. 2022) which measures anger/irritability, anxiety and depression in early childhood have the potential to measure treatment outcomes since they ask about past week behaviors. To our knowledge, they have not yet been used for that purpose.

Bottom-up attempts to obtain direct observations of outbursts include the Disruptive Behavior-Diagnostic Observation Scale (DB-DOS) for observing preschool children as a supplement to parent report (Wakschlag et al., 2008). In older children, Ceresoli-Borroni et al. (2019) developed an impulsive aggression (outburst) outcome measure for drug studies by creating a diary or list of behaviors that the parent or child/teen could complete near the time of the impulsive aggression or outburst. The final list of behaviors was used by families over a 2-week period to record what the youth did soon after an aggressive episode occurred. There was no attempt to define differences between irritable mood and outbursts.

Carlson et al. (2010, 2022) also used a bottom-up approach to develop an observation scale to rate outburst components and duration on an inpatient unit. Beginning with items from observations of preschool tantrums (Potegal and Davidson, 2003) and obtaining input from experienced inpatient nursing staff, they created the Children’s Agitation Scale (Carlson et al., 2010; Potegal et al., 2009) for nurses to rate an outburst every 15 min recording what they saw so that both behaviors and their duration were tracked. Outburst items were ultimately incorporated into a screening inventory now called the “Emotional Outburst Inventory (EMO-I; Carlson et al., 2022) previously described.

Two recently developed interviews that assess irritable mood, tantrums and impairment are the Early Childhood Irritability-Related Impairment Interview (E-CRI) (Wakschlag et al., 2020) for preschool children and the Clinician-Affective Reactivity Index (CL-ARI, Haller et al., 2020) for older children. As these are new instruments, clinical uptake has not been determined.

Recommendation #8: We need practical measures that specifically address irritable mood and outbursts and their severity in clinical samples and that elicit the time course of both mood and outbursts as well as their frequency and severity. We suggest that inpatient units and special education settings are not only in need of such practical irritability measures, they are also good settings in which to develop “real world,” “bottom up” measures or try the recently developed ones.

The MAPS-TL (Wiggins, Ureña Rosario, Zhang et al., 2023) scales and Mazefsky’s et al. (2018; Mazefsky, Yu, et al., 2021). Emotion Dysregulation Inventory has generated useful screens. They could be used to generate scores that could be used as an index of severity and, presumably, as an index of treatment response or factor analyzed to determine utility as measures of chronic and phasic irritability for baseline and weekly tracking.

Defining the Characteristics of Irritable Mood and Outbursts – Informant Effects

Since irritability has both internalizing and externalizing components, accurate information is needed from the person who can discuss their internal state and from those around them who can identify resulting behavior. However, modest inter-informant correlations (Achenbach et al., 1987) beset irritability measures, too. On the ARI (Stringaris et al., 2012) correlations between parent and child/adolescent range from r = 0.23 (Kessel et al., 2021) to 0.58 (Stringaris et al., 2012). Using a multi-trait, multi-method factor analysis from scales given to multiple informants, Zik et al. (2022) reported a correlation of r = 0.23; while 19% of the total variance on the parent report ARI was explained by irritability, 55% was explained by the informant choice (in this case, parent).

De Los Reyes et al. emphasize that informant discrepancies provide unique and valuable information (De Los Reyes et al., 2023). Degroot et al. (in press) found that parent/teacher concordance was higher (r = 0.52) for phasic than tonic (r = 0.34) irritability. Using the outburst rating from the Child Mania Rating Scale (Pavuluri et al., 2006), concordant parent/teacher endorsements occurred most often for children diagnosed with externalizing disorders (ADHD, ODD, conduct disorder, or any combination of these). Discordant ratings were more common with anxiety and depressive disorders (Carlson & Blader, 2011).

Recommendation #9: To understand the condition we are defining, we need to revisit methodology about how and when to integrate informants. We need to ensure all informants understand the questions asked the same way being especially mindful that children and adults with behavior problems have a higher rate of language comprehension issues (Carlson et al., 2020).

Bird et al. (1992) reported that for child psychiatric epidemiology there was no statistically optimal way to combine parent and child data and use any positive response. That rule comes at a cost of understanding the phenomenology of a condition, at the very least. Bird, of course, was addressing the needs of epidemiology not phenomenology. “Bottom up” visual analog scales may help language-limited children to identify mood and anger. Systematic debriefings of parent/child units to understand sources of agreement and disagreement on interviews and rating scales they complete might bring further understanding to discrepancies in ratings where they occur.

Defining Characteristics of Irritable Mood and Outbursts – Effects of Comorbidity

Irritability has been studied in ADHD and autism, anxiety, obsessive-compulsive disorder (OCD), depression, and Tourette disorder. The psychopathological significance of irritability is not consistent across disorders and the relationship may depend on what is measured. This is another reason irritability should be assessed both independent of diagnosis and within relevant disorders. Greater irritability may be associated with greater disorder severity and impairment (e.g. Elvin et al., 2023; Shimshoni et al., 2020 for anxiety; Guzick et al., 2021 for OCD; Sherwood et al., 2021 for depression) or confounded with comorbid disruptive behavior disorders (Stringaris et al., 2013).

One-third to one-half of children with ADHD-combined type experience low frustration tolerance and short tempers (Shaw et al., 2014; Faraone et al., 2019; Eyre et al., 2017). The ADHD/irritability or dysregulation combination is frequent enough that some have felt that it should be considered a subtype (Karalunas et al., 2014, 2019), independent of comorbid ODD. When multiple measures of mood between outbursts and anger/outbursts/aggression were examined in a recruited sample of children with ADHD, irritability symptoms and DMDD, a high correlation between mood and outburst measures was found (r = 0.58 to 0.79) but outburst severity tracked with ADHD severity not mood with larger associations for temper outbursts and aggression than for mood between outbursts (Cardinale et al., 2021).

Irritability/dysregulation and outbursts often play a large role in children with ASD (e.g. Conner et al., 2021). It is a sufficiently serious clinic problem that irritability in autism is the single irritability condition for which the Food and Drug Administration has approved the use of medications (Fung et al., 2016).

Where outbursts rather than irritable mood were the focus of study, samples of children with anxiety (Johnco et al., 2015) and OCD (Storch et al., 2012) were found to have frequent “rage” episodes. In Tourette disorder, various samples evidenced “rages” occurring between 20% and 67% (Conte et al., 2020) of children. Data were inconclusive about the relationship between rage attacks and tic severity but there did appear to be an association between rages, ADHD, OCD, irritable mood and externalizing behaviors (Conte et al., 2020).

Finally, DMDD’s diagnostic criteria preempts the diagnosis of ODD. If DMDD is truly different, then the symptoms should occur without the negativistic, headstrong component of ODD which has a different set of predictors (Burke et al., 2021). If it does not, both diagnoses should be given.

Recommendation #10: Not only should irritability be ascertained independent of diagnosis but also future studies should compare descriptions and formulations of irritability (mood and outbursts) within as well as across disorders to better understand how irritability interacts with the various conditions. Characteristics that might change include whether irritable mood or outbursts are part of the initial symptom, or complication and sign of worsening.

Until we have a better understanding of DMDD, the only truly exclusionary diagnoses should be bipolar disorder and intermittent explosive disorder.

Treatment – What Lights the Fuse (Triggers)

Given the degree of impairment that accompanies irritability, effective treatment is a critical goal of future research. Two metaphors describe the relationship of mood and irritability. The first is Jack’s description of simmering (“irritable mood”) and boiling over (“outbursts”). The current treatment approach is to lower the boiling point by improving the problem or disorder driving the irritable mood so that boiling over happens less easily. The use of stimulant treatment for ADHD appears to decrease impulsive aggression (otherwise known as outbursts) (Blader et al., 2021) and possibly decrease outburst duration (Spring et al., 2024), improving ADHD symptom severity in a variety of ways.

The second metaphor is of a bomb and resulting explosion. The use of second-generation antipsychotics for irritability in autism is an example of reducing the explosion (outbursts) independent of diagnosis. The primary condition, autism, is not the focus of treatment (e.g. McCracken et al., 2002).

To expand the bomb metaphor, what lights the fuse to the bomb is the trigger. We have said little about triggers or antecedents to outbursts, but they are an important point of intervention and may well change according to the primary disorder. Leibenluft (2017) has reviewed a number of mechanisms that underpin irritability. While not strictly speaking “triggers,” they include frustrative non-reward or becoming easily frustrated, hostile interpretation bias, atypical response to emotional stimuli and cognitive inflexibility/resistance to change which may occur in neurodevelopmental disorders. Moreover, investigators interested in these models have developed psychotherapeutic interventions to add to time-tested parent management therapy (Dekkers et al., 2022; Waxmonsky et al., 2021) and anger management (Sukhodolsky et al., 2016). Newer treatments include Exposure-Based Cognitive-Behavioral Therapy for Youth with Severe Irritability and Temper Outbursts (Naim et al., 2023), Regulation Focused Psychotherapy (Prout et al., 2021), and Dialectical Behavior Therapy for Children (Perepletchikova et al. 2017).

Recommendation #11: Systematic study of triggers for irritable mood and outbursts in different psychopathological conditions would facilitate the development of more targeted preventive interventions both in the long term for children at risk to learn emotion regulation [(the goal for screening measures like the MAPS-TL, (Wiggins, Ureña Rosario, Zhang et al., 2023)] and in keeping people in the child’s environment from lighting fuses.

Information on outburst antecedents is often obtained in schools as part of the functional behavioral assessment generated for children with temper problems. Collaborations with special education classrooms would be useful to help determine if certain triggers were more common in specific conditions. Also, parents, teachers and treatment staff often report outbursts occur “for no reason.” There is always a reason. “No reason” means the antecedent has been overlooked or unrecognized. Ecological momentary assessment (EMA) to obtain on the spot, so to speak, information from the informant about the sequence of events prior to the outburst would be helpful in uncovering triggers.

Treatment – the Length of the Fuse

De-escalation, cognitive-behavioral strategies, parent training and anger management address increasing frustration tolerance or lengthening the fuse. Among available psychotherapies, the MATCH approach (Evans et al., 2020; Evans & Santucci, 2021) has a structure for examining irritability focused treatments for different conditions and includes evidence-based applications of parent management training and cognitive-behavioral treatments. Previously mentioned newer psychotherapies show promise. For the most seriously afflicted, treatment requires multiple interventions (McClellan et al., 2023) and is not quick.

Breaux et al. (2023) recently reported a meta-analysis of nonpharmacologic, pharmacologic and combined pharmacologic and nonpharmacologic studies for treating irritability. The pooled effect size for randomized controlled psychopharmacologic treatments was an impressive Hedges g of 1.89 (95% CI 1.42–2.36) for autism spectrum disorder but a more modest 0.64 (95%CI 0.51–0.78) for ADHD/DMDD/severe mood dysregulation/disruptive behavior disorders. Antipsychotics alone or combined with other treatments (usually stimulants) produced the largest effects but the studies were complicated by considerable heterogeneity and medication side effects. Whether psychotropic medication reduced irritability by moderating mood or decreasing the frequency, severity or duration of outbursts needs to be specifically empirically evaluated as primary endpoints in clinical trials.

While the Breaux et al. study (2023) outlines important goals for future research, most of the studies conducted on irritability were for autism. That is because there is a specific indication for treatment with a rating scale that the United States Food and Drug Administration (USFDA) and investigators have accepted, namely the irritability subscale of the Aberrant Behavior Checklist (ABC; Aman et al., 1985).

Recommendation #12: We need to develop clinically relevant outcomes that would be considered valid (e.g. by USFDA) for testing in new clinical trials. This will require developing an effective, treatment-responsive, child-relevant measure for both irritable mood and outbursts which addresses frequency, severity and duration of both.

Aman et al.’s (1985) ABC-Irritability symptom scale, used to obtain the “irritability in autism” indication, is unfortunately not pertinent to conditions besides autism. For FDA consideration, the question is whether a single transdiagnostic scale would be acceptable or if disorder-specific scales are needed, i.e. separate rating scales for irritability in ADHD, DMDD, Tourette Disorder, etc. We may also need different measures for inpatient studies where day-to-day changes are monitored versus outpatient studies where weekly or longer metrics are used. A “bottom up” methodology, with a developmentally relevant measure of outbursts is needed that addresses both severity and duration of outbursts.

Treatment – the Size of the Explosion and Damage Control

It is the size of the explosion and how long it lasts that causes the most damage and that we frankly are at a loss to handle. Much has been written about eliminating the use of seclusion and restraint (Carlson et al., 2020; Masters et al., 2002; Slaatto et al., 2021). However, the goal should be to eliminate or reduce outbursts not just reduce seclusion and restraint. What suffices in an emergency room, where the goal is to discharge the child (or adult) as quickly as possible, is different from having to manage someone day after day as in school, hospital or residential programs. Unfortunately, we lack good solutions to the explosion other than damage control. De-escalation serves to prevent or diminish the immediate outburst. It does not address future outbursts. De-escalation is turning down the heat on the stove to prevent the water from boiling over. Once the boil-over occurs, it is necessary to get the pan off the heat quickly! Preventing future boil-overs is not the immediate goal. A review of de-escalation and restraint studies in residential facilities concluded that interventions that helped reduce seclusions and restraints differed from those that lead to a reduction in conflicts and aggression (Slaatto et al., 2021). Without sufficiently effective and well-tolerated treatments, the most seriously dysregulated children are often vulnerable to capricious forms of polypharmacy, hospitalization and residential care due to the magnitude of impairment related to their irritability and outbursts (Barclay et al., 2021; Carlson, Singh et al., 2023; McClellan et al., 2023; Sorter et al., 2022).

Recommendation #13: We need studies to advance our understanding of outcomes associated with interventions like PRN medication, seclusion/isolation, evacuating classrooms or hospital wards until the outburst burns itself out. Ultimately, we need data to successfully select specific and optimal first-line strategies for the immediate, short- and long-term management of youth with irritability leading to outbursts. We also need new treatments to address difficult-to-treat behaviors that lead to their reduction and hopefully extinction.

A consensus conference of researchers, experienced clinicians, policy makers and hospital administrators would be helpful to move this process along. The question would be who would organize and pay for it. In years past, multidisciplinary consensus conferences were held to decide on who should be included in medication studies of mania in children (Carlson et al., 2003) and to plot a course to study impulsive aggression (Jensen et al., 2007). Perhaps, the field is ready to tackle some of the same issues with what we are learning about irritability in children.

Who are We Treating?

Sample demographics are needed to understand who is being represented in a study. However, we also need to know the sample severity including both the scalable outcome measures and overall normed measures of recognized, comorbid symptom dimensions. For instance, combined tantrums and irritability were reported in 8.7% of a community sample of parents of 6-year-olds (Dougherty et al., 2016) and almost half of a clinic sample of 6-year-olds from the same community. CBCL aggression T score was 64 in the former (significantly higher than the rest of the community sample) but the T score was 73 in the clinic sample (Carlson et al., 2016). The clinic sample severity was 1 standard deviation worse than the community sample. Even though the terms “tantrum” and “irritability” were the same, the severity clearly was not. Similarly, in the MATCH psychotherapy study (Evans & Santucci, 2021), the overall sample severity T scores for internalizing (69.5 ± 8.04), externalizing (69.85 ± 5.38) and total problem (T 71.23 +4.31) [CBCL (Achenbach & Rescorla, 2001)] dimensions could be compared with the optimized stimulant plus risperidone or valproate study (Blader et al., 2021) where the sample’s admission aggression T score was 76.95 ± 8.99 and total problem T score was 71.82 ± 5.63. We can surmise with data that severity of externalizing behaviors were relatively similar in these studies.

Cardinale’s study (Cardinale et al., 2021) of irritable ADHD children reported hyperactivity/impulsivity T scores (Conners, 2008) of 79.24 + 9.04 and an ARI (Stringaris et al., 2012) score of 8.99 ± 1.70) out of 12. Children were clearly hyperactive and irritable though without a specific outburst measure or at least an aggression score we do not know if they had outbursts or how severe they were. We might also glean that even with a respectable treatment effect size of 1.0, the children in that sample will still be significantly symptomatic at the end of treatment. That fact is important to one’s expectations for treatment response.

Recommendation #14: In addition to an outcome measure of irritability, treatment studies should report well-replicated, recognized, and normed measures of the conditions associated with irritability, like ADHD and aggression so we can compare sample severity, and judge subsequent intervention impact (Carlson & Klein, 2019). An effect size for any finding needs the perspective of the sample’s initial severity, not just magnitude of change.

Summary and Conclusions

In this “Future Directions,” we have emphasized the importance of understanding the phenomenology of irritability. We begin by asking for the use of face-valid terms for irritable mood and outbursts as well as keeping the two concepts separate rather than blending them so it is clear what is actually being studied. Although Ben Franklin warned the 13 colonies at their birth with the Declaration of Independence to “hang together” or “hang separately,” for irritable mood and outbursts, we need both hanging together and separately as well as an understanding of the relationship between them.

Course of illness has always been integral to validating psychopathology (Robins & Guze, 1970) – not just the outcome but the pathways to getting there. We and others advocate for the development of severity measures that include items for both irritable mood and outbursts, reworking measures of mood and aggression as well as a “bottom up” component with new ideas and an eye to understanding the implications of the inevitable informant variance that will occur. We desperately need treatment-responsive measures and an FDA indication for treating irritability across childhood disorders.

Finally, we need a way of accurately identifying the young people we see with disabling outbursts. If they meet our DSM criteria for DMDD, or IED the problem is solved. If they miss by a criterion, there should be an “other specified” category. Most importantly, though, if they have another disorder, like ADHD, or ODD, or PTSD, or a mood disorder and we are not yet sure that their irritable mood and/or outbursts are a disorder, we should use R codes rather than invoking a disorder they do not have.

The biggest issue, of course, will be who will do this work, and how it will be paid. In years past, multidisciplinary consensus conferences were held to decide on who should be included in medication studies of mania in children (Carlson et al., 2003) and to plot a course to study impulsive aggression (Jensen et al., 2007). Perhaps, the field is ready to tackle some of the same issues with what we are learning about irritability in children.

Disclosure Statement

Dr. Carlson receives grant support from NIH and honoraria for lectures and courses given for the American Academy of Child and Adolescent Psychiatry.

Dr. Althoff receives or has received grant support from NIH and the Klingenstein Third Generation Foundation, honoraria for an editorial position at JAACAP and for presentations at the MGH Psychiatry Academy, and has an ownership interest in WISER Systems, LLC.

Dr. Singh has received or will receive research support from the National Institutes of Health, Patient Centered Outcomes Research Institute, AbbVie, Intracellular Therapeutics, and Bausch Health. She is on a data safety monitoring board for a study funded by the National Institute of Mental Health. She is on the national scientific advisory board for Skyland Trail. She receives honoraria from the American Academy of Child and Adolescent Psychiatry, and royalties from American Psychiatric Association Publishing.