5.1. INTRODUCTION
This guideline is applicable to people above the age of 3 and of all levels of intellectual ability, who show symptoms of hyperactivity, impulsivity or inattention to a degree that severely impairs their mental or social development causing failure to make expected progress in the domains of intellectual development, personal relationships, physical or mental health or academic function. This includes people with ADHD whether or not they have other coexisting developmental or mental health disorders or whether the ADHD behaviours and symptoms result from genetic, physical environmental or social-environmental causes. This chapter sets out to look at the issue of diagnostic categorisation and assessment that should trigger the use of this guideline. Sections 5.3 to 5.14 address the validity of the diagnostic construct of DSM-IV-TR ADHD and ICD-10 hyperkinetic disorder as diagnostic categories that give rise to significant impairments. Sections 5.15 to 5.17 provide guidance for clinical practice.
For ADHD the question is whether a diagnostic category associated with clear evidence of impairment, that most people would consider requires some form of medical, social or educational intervention, can be reliably defined. To provide guidance for clinicians involved in the medical component of such intervention, the validity of the diagnostic concept of ADHD is addressed using the definition of a clinical disorder or illness as any condition that causes discomfort, dysfunction, distress or social problems to the person concerned. This part of the guideline addresses the question of validity of the diagnostic construct of ADHD and provides practice guidelines for the diagnostic process.
5.2. DEFINITIONS OF TERMS
The terminology applied to ADHD and related problems has been used in different ways at different times and by different groups of people. This section clarifies some of the major terms used in this chapter. A description of the diagnostic terms is provided in Chapter 2.
ADHD and hyperkinetic disorder
The terms ADHD (DSM-IV-TR) and hyperkinetic disorder (ICD-10) are used when talking about the specific diagnostic categories of ADHD as defined by DSM-IV-TR and hyperkinetic disorder as defined by ICD-10 respectively. The criteria for hyperkinetic disorder are more stringent that those for ADHD with hyperkinetic disorder forming a subgroup of the DSM-IV-TR ADHD combined type diagnosis (see Chapter 2). When discussing the disorder more broadly ‘ADHD’ is used as an umbrella term. Some of the earlier literature used the term ‘hyperactivity’ for the cluster of hyperactive, impulsive and inattentive symptoms. In this guideline the term ‘hyperactivity’ will be restricted to mean the combination of symptoms that define overactive behaviour and the term ‘ADHD symptoms’ used to refer to the combination of hyperactive, impulsive and inattentive symptoms.
Symptoms
The behavioural phenomena that describe ADHD will be referred to as symptoms of ADHD throughout this chapter. This choice of wording is intended to reflect the fact that the behavioural phenomena that characterise ADHD may not always be reported as observed behaviours, but may also be reported as subjective changes in mental state. For simplicity the term ADHD symptoms will be used whether the guideline is discussing impairing levels of behaviour or mental phenomena, or referring to the normal range of behaviour of these phenomena. For example, many people have low to moderate levels of ADHD symptoms, which do not reflect an impairing condition or mental health disorder.
Having said that, the GDG recognises that behaviours that describe ADHD are not strictly symptoms, as this term is usually used to refer to changes in physical or mental state associated with significant morbidity that is a change from a premorbid state: for example, symptoms experienced during an episode of depression or attack of anxiety. The behavioural and mental phenomena that characterise ADHD are in contrast trait-like, in the sense that they are non-episodic and may have been present from early childhood. Furthermore, in children the criteria are usually applied on the basis of parent and teacher reports of behaviour, rather than subjective reports of mental state phenomena. Older children and adults are usually able to provide detailed descriptions of their subjective experiences of inattention, hyperactivity and impulsivity.
Oppositional defiant disorder and conduct disorder
The use of these terms is restricted to mean the definitions of oppositional defiant disorder and conduct disorder as described in DSM-IV-TR. The GDG recognises, however, that the terms oppositional defiant disorder and conduct disorder are widely used outside of these narrow diagnostic definitions. Many studies cited in this review have used rating scale measures for aspects of oppositional defiant disorder and conduct disorder and people often use the term conduct disorder when they are talking about oppositional behaviour. We will therefore use the terms conduct problems or oppositional defiant problems when referring to these classes of behaviour where the DSM-IV-TR definitions have not been strictly applied.
5.3. THE VALIDITY OF ADHD AS A DIAGNOSTIC CATEGORY
The use of the diagnosis of ADHD has been the subject of considerable controversy and debate and the diagnosis itself has varied across time and place as diagnostic systems have evolved (Rhodes et al., 2006). Points of controversy identified by the GDG included both specific issues, such as the wide variation in prevalence rates reported for ADHD and the possible reasons for these differences, and the nature of the aetiological factors that increase the risk for ADHD, as well as more complex broader sociological and philosophical issues.
The GDG wished to evaluate evidence for the validity of the diagnostic category of ADHD and formulate a position statement on the use of the diagnosis. It is recognised that defining neurodevelopmental and mental health disorders is a difficult process because of the overlapping nature of syndromes, the complexity of the aetiological processes and the lack of a ‘gold standard’ such as a biological test. In this regard ADHD is similar to other common psychiatric disorders that rely on the identification of abnormal mental phenomena. Although biological tests for ADHD do not exist, the diagnosis can be reliably applied when data capture tools such as standardised clinical interviews used by trained individuals and operational diagnostic criteria are employed (for example, Taylor et al., 1986; Schwab-Stone et al., 1993; Schwab-Stone et al., 1994; Epstein et al., 2005).
In keeping with most common mental health disorders, the distinction between the clinical condition and normal variation in the general population is difficult to define on the basis of symptom counts alone. This is because there is continuity in the level of ADHD symptoms between those with an impairing mental health disorder and those who are unimpaired. The distinction between ADHD and normal variation in the general population requires the association of a characteristic cluster of symptoms and significant levels of impairment. This is comparable to normal variation for medical traits such as hypertension and type II diabetes, as well as psychological problems such as anxiety or depression. Controversial issues surround changing thresholds applied to the definition of illness as new knowledge and treatments are developed (Kessler et al., 2003) and the extent to which it is acknowledged that clinical thresholds are socially and culturally influenced and determine how an individual’s level of functioning within the ‘normal cultural environment’ is assessed (Sonuga-Barke, 1998). In considering these issues, a key question is to define the level of ADHD symptoms and associated impairments required to trigger the use of this guideline.
Undertaking a systematic review of diagnostic categories is not a straightforward exercise for behavioural and mental health disorders because in most cases definitive diagnostic tests for the presence or absence of disorder do not exist. The relative lack of a validated reference standard (indicated by SIGN diagnostic study quality assessment, see Appendix 16) means that the question of validity for the diagnosis of ADHD needs to draw on evidence from a wide range of sources. There is also potential for ascertainment bias, particularly in clinic-referred populations, and considerable variability resulting from the use of different clinical and demographic subgroups, differences in disease prevalence and severity among various populations sampled for research, and the use of different behavioural and symptom measures (Whiting et al., 2004). The GDG wishes to emphasise that psychiatric nosology is a dynamic and developing field and changes are to be expected as more data are accrued over time.
5.4. METHODOLOGY
To ensure that a transparent, structured approach was taken, the GDG agreed to use one similar to the Washington University Diagnostic Criteria (Feighner et al., 1972). The methodology used to create the Washington University Diagnostic Criteria has been widely accepted for this purpose, and similar approaches have been taken to validate diagnostic categories for the Research Diagnostic Criteria, the DSM and the ICD. The approach involves setting out criteria for validating a particular disorder and seeing how far a particular set of phenomena are consistent with those criteria. Using these criteria as a framework this chapter sets out to answer the following questions:
To what extent do the phenomena of hyperactivity, impulsivity and inattention, which define the current DSM-IV-TR and ICD-10 criteria for ADHD and hyperkinetic disorder, cluster together in the general population and into a particular disorder that can be distinguished from other disorders and from normal variation?
Is the cluster of symptoms that defines ADHD associated with significant clinical and psychosocial impairments?
Is there evidence for a characteristic pattern of developmental changes, or outcomes associated with the symptoms, that define ADHD?
Is there consistent evidence of genetic, environmental or neurobiological risk factors associated with ADHD?
Studies were selected for inclusion in this review if they met the SIGN quality assessment criteria for systematic reviews and cohort studies. For diagnostic and factor analytic studies the GDG established a set of criteria approved by NICE: (1) the study addresses an appropriate and clearly focused question (or hypothesis) and (2) the sample population being studied are selected either as a consecutive series or randomly, from a clearly defined study population.
A literature search was conducted for existing systematic reviews and meta-analyses on CINAHL, EMBASE, MEDLINE, PsycINFO, which were considered to be the best level of evidence. The initial search found 5,516 reviews of which nine were relevant to the questions about ADHD and application of the Washington University Diagnostic Criteria. Where insufficient evidence was found from previous systematic reviews, a search for primary studies was carried out (see Appendix 16).
In addition to the review of the literature, a consensus conference was held to bring together experts in the field who held a range of views and could address the concept of ADHD from different perspectives. This provided an opportunity to debate the key issues surrounding the use of the diagnostic category and thereby to assist the GDG with the task of deciding what should trigger the use of the guideline and for whom the guideline is intended. A summary of the consensus conference is provided in Section 5.14.
5.5. REVIEWING THE VALIDITY OF THE DIAGNOSIS: SUMMARY OF THE EVIDENCE
The first issue to be addressed is: To what extent do the phenomena of hyperactivity, impulsivity and inattention, which define the current DSM-IV-TR and ICD-10 criteria for ADHD and hyperkinetic disorder, cluster together in the general population and into a particular disorder that can be distinguished from other disorders and from normal variation?
The evidence addressing this issue is divided into three main questions:
- 5.5.1.
Do the phenomena of hyperactivity, inattention and impulsivity cluster together?
- 5.5.2.
Are ADHD symptoms distinguishable from other conditions?
- 5.5.3.
Are the phenomena of hyperactivity, inattention and impulsivity distinguishable from the normal spectrum?
5.5.1. Do the phenomena of hyperactivity, inattention and impulsivity cluster together?
No evidence was found from the systematic search of reviews that was of direct relevance to this question. This is because, despite a large primary literature, no systematic reviews in this area have been undertaken. Therefore a systematic search of factor-analytic and cluster-analytic studies was carried out. Additional factor-analytic and cross-sectional studies were identified by the GDG (Appendix 17.1). None of these studies met the SIGN inclusion criteria that require an appropriate reference standard for diagnostic measures, but most did meet the extension to the SIGN criteria approved for this review, since the aim of the question was to evaluate whether the phenomena of hyperactivity, inattention and impulsivity cluster together in the population, rather than to assess the accuracy of diagnostic tests.
The inclusion criteria for factor- and cluster-analytic studies were defined as follows: (i) that the study addresses an appropriate and clearly focused question (ii) that the sample being studied was selected either as a consecutive series or randomly, from a clearly defined study population.
Evidence
Many factor analyses indicate a two-factor model: ‘hyperactivity-impulsivity’ and ‘inattention’. This has been replicated in population-based studies (Lahey et al., 1994; Leviton et al., 1993; Wolraich et al., 1996) and clinical samples (Bauermeister et al., 1992; Lahey et al., 1988; Pelham et al., 1992).
In an early study, ‘hyperactivity-impulsivity’ was reported as a single factor, where the factor ‘hyperactivity’ was defined as ‘impulsive, excitable hyperactivity’ (Dreger et al., 1964).
More recent factor-analytic studies based on DSM-IV criteria support previous findings that the phenomena of inattention and hyperactivity-impulsivity form distinct symptom clusters in children (Molina et al., 2001; Amador-Campus et al., 2005; Zuddas et al., 2006) and young people (Hudziak et al., 1998).
Looking specifically at children identified as having a behavioural problem, Conners (1969) found ‘hyperactivity’ and ‘inattention’ as separate and distinct factors. The factor structure of adolescent self-report behavioural data was investigated by Conners and colleagues (1997): six factors were identified, including ‘hyperactivity’ and ‘cognitive problems’. The ‘hyperactivity’ factor included characteristics such as being unable to sit still for very long, squirming and fidgeting and feeling restless inside when sitting still. The ‘cognitive problems’ factor consisted of having trouble keeping focused attention, having problems organising tasks and forgetting things that were learnt. In a further study by Conners and colleagues (1998) similar findings were reported. An attentional problem factor was found that overlapped with the DSM-IV criteria for the inattentive subtype of ADHD, with a similar overlap between the factor items for hyperactivity and the DSM-IV criteria for hyperactivity-impulsivity.
Some studies have identified three factors, with ‘hyperactivity’ and ‘impulsivity’ as two distinct factors in addition to ‘inattention’, in both population (Gomez et al., 1999; Glutting et al., 2005) and clinical samples (Pillow et al., 1998). However, Gomez and colleagues (1999) showed that the model fit for the three-factor solution was only marginally better than the two-factor model. In the study of Pillow and colleagues (1998) of boys with ADHD, the impulsive and hyperactive symptoms formed a single factor when oppositional-defiant and conduct disorder items were also included in the factor analysis.
Werry and colleagues (1975), however, found that hyperactivity, impulsivity and inattention formed a single factor using both population control and ‘hyperactive’ samples.
Latent class analysis (LCA) identifies clusters of symptoms that group together. Using this approach, Hudziak and colleagues (1998) found that hyperactivity-impulsivity and inattentive symptoms cluster together as a ‘combined’ type latent class, as well as separate hyperactive-impulsive and inattentive latent classes. The latent classes map closely to the DSM-IV criteria, with DSM-IV combined type ADHD falling entirely within the severe combined type latent class, whereas individuals with the DSM-IV inattentive subtype fell either within the severe inattentive or the severe combined type latent classes.
The clustering of hyperactivity, impulsivity and inattention appear to be stable across a number of countries. Ho and colleagues (1996) found separate robust dimensions for ADHD symptoms, antisocial and neurotic behaviour in a sample of 3,069 Chinese schoolboys. Correlations among different dimensions were similar to those reported in European and US samples. Taylor and Sandberg (1984) compared data from 437 English schoolchildren with published data from the US and New Zealand. They identified a factor of hyperactivity-inattention that was distinct from conduct disorder. The comparisons supported the view that English schoolchildren were similar to their contemporaries in the US and New Zealand with differences in prevalence rates between different countries accounted for by discrepancies in diagnostic practice.
In adult population samples a two-factor model has been identified (DuPaul et al., 2001; Smith & Johnson, 2000) as well as a three-factor model (Kooij et al., 2005). Glutting and colleagues (2005) assessed university students aged 17 to 22 using parent-rated information in addition to self-rated data. They reported slightly contrasting findings within each set of data: exploratory and confirmatory analysis showed that DSM-IV ADHD symptoms generated a three-factor model in the self-report data and a two-factor model in the parent-informant data.
Although most studies show separate factors for inattention and hyperactivity-impulsivity, these are highly correlated in children (Gomez et al., 1999) and adult samples (Kooij et al., 2005).
There may be age-dependent changes in the factor structure. Bauermeister and colleagues (1992) found that there was a single attention/impulsivity-hyperactivity factor in pre-school children, and separation into two factors in school-age children. Nearly all the studies of school-age children reported two factors. In contrast, the study from Glutting (2005) using college students aged 17 to 22 found three factors, with the separation of hyperactive and impulsive symptoms. Similarly Kooij and colleagues (2005) using adult samples identified three separate factors.
Summary
There was strong evidence for clustering of inattentive and hyperactive-impulsive symptoms in both population and clinical samples. Evidence for one-, two- and three-factor models was found, with most studies supporting a two-factor model. Most studies found two correlated factors for hyperactivity-impulsivity and inattention, while others were able to distinguish between hyperactivity and impulsivity and a few found one combined factor for all three domains. There is some evidence that the number of factors identified depends on the age of the sample, with nearly all studies of school-age children reporting two factors. These findings have been observed in both population and clinical samples and in a number of different cultural settings. LCA in population samples detects clustering of symptoms into groups that are similar but not identical to DSM-IV subtypes for ADHD.
5.5.2. Are ADHD symptoms distinguishable from other conditions?
No systematic reviews were identified in the literature that addresses this question. The GDG considered that the most important and controversial distinction to be made was between ADHD and oppositional-defiant and conduct disorders. These are also the most commonly reported coexisting conditions in children and young people diagnosed with ADHD and define a set of behaviours that might be difficult to distinguish from ADHD. It was therefore decided to restrict a formal literature search to identify studies that indicate whether a distinction can be made between ADHD, oppositional-defiant and conduct problems. Additional references were identified by the GDG members (see Appendix 17.1).
Evidence
ADHD and oppositional-defiant and conduct problems Most of the studies using factor-analytic approaches for the analysis of ADHD symptoms report separate factors for hyperactivity-impulsivity, inattention and oppositional-defiant or conduct problems. These include most of the studies reviewed in the previous section on the factor structure of ADHD symptoms (for example, Bauermeister et al., 1992; Conners, 1969; Conners, 1997; Ho et al., 1996; Pelham et al., 1992; Taylor & Sandberg, 1984; Werry et al., 1975; Wolraich et al., 1996). These studies are highly consistent in being able to separate the items that describe oppositional-defiant and conduct problems from hyperactivity-impulsivity and inattention. Although the behavioural items fall into separate dimensions there are significant correlations between the various behavioural factors.
Two studies using LCA came to different conclusions. Frouke and colleagues (2005) conducted a diagnostic study of 2,230 Dutch pre-adolescents from the general population. LCA revealed that ADHD symptoms clustered together with symptoms of oppositional defiant disorder and conduct disorder. A further study from the Netherlands of disruptive behaviour in 636 7-year-old children (van Lier et al., 2003) came to similar conclusions. LCA identified three main classes of children with: (i) high levels of oppositional defiant disorder and ADHD; (ii) intermediate levels of oppositional defiant disorder and ADHD with low levels of conduct problems; and (iii) low levels of all disruptive problems. No classes were identified with only ADHD, oppositional defiant disorder or conduct problems.
In contrast, King and colleagues (2005a) identified five distinct groups using a cluster analysis, which like LCA identifies discrete groups of symptoms clusters: ADHD with inattention (ADHD-I), ADHD with hyperactivity-impulsivity (ADHD-H/I), ADHD with both hyperactivity/impulsivity and inattention (ADHD-C), ADHD-C with oppositional defiant disorder, and ADHD-I with oppositional defiant disorder. For both the inattentive symptoms and combined inattentive/hyperactive-impulsive symptoms they found clustering either with or without symptoms of oppositional defiant disorder.
Latent dimension modelling by Ferguson and colleagues (1991) looking at children with ADHD and conduct disorder suggested that these could be seen as independent dimensions, although they are highly inter-correlated. Having said that, the two often occurred independently of each other and only partially shared aetiological factors.
ADHD can be a precursor of other problems. When ADHD and disruptive behavioural problems coexist, the history usually suggests that symptoms of ADHD appear first before the development of disruptive behavioural problems. A follow-up of a community sample of children with ADHD symptoms but no oppositional behaviour between the ages of 7 and 17 found that children with ADHD symptoms could develop oppositional behaviour at a later stage, but that the reverse pathway from oppositional behaviour to ADHD was uncommon (Taylor et al., 1996).
Population twin studies find that symptoms of ADHD are distinct from but share overlapping genetic influences with conduct problems (Thapar et al., 2001; Silberg et al., 1996; Nadder et al., 2002). Multivariate twin modelling suggests that while the genetic influences on conduct disorder are largely shared with those that influence ADHD, there are in addition important environmental factors shared equally that influence the risk for conduct problems but not ADHD (Thapar et al., 2001). In nearly all twin studies of ADHD there is evidence for the influence of unique environmental factors but not shared (familial) environment; whereas for conduct problems, twin studies find evidence of shared environmental influences. Nadder and colleagues (2002) conclude that the co-variation of ADHD and oppositional defiant disorder/conduct disorder is the result of shared genetic influences with little influence from environmental factors. There are, however, substantial additional influences from shared environmental factors on oppositional defiant disorder/conduct disorder, especially when they are not accompanied by ADHD (Silberg et al., 1996; Eaves et al., 1997).
ADHD and other coexisting conditions Population twin studies find that symptoms of ADHD are distinct from but share overlapping familial and genetic influences with other neurodevelopmental traits including reading ability (Gilger et al., 1992; Willcutt et al., 2000; Willcutt et al., 2007), general cognitive ability (Kuntsi et al., 2004), symptoms of developmental coordination disorder (Martin et al., 2006) and symptoms of pervasive developmental disorders (Ronald et al., 2008).
ADHD is reported to coexist with personality disorder in young offenders (Young et al., 2003). A prison survey found that 45% of incarcerated young adults had a previous history and persistence of ADHD symptoms (Rosler et al., 2004). The distinction between ADHD and personality disorder in adults raises important nosological questions and remains poorly investigated.
Dysthymia, depression and anxiety symptoms and disorders are frequently associated with ADHD in adults. In the US National Comorbidity Survey, adults with ADHD had increased rates of mood disorders, anxiety disorders, substance misuse disorders and impulse control disorders (Kessler et al., 2006). The causal links between ADHD and these coexisting symptoms, syndromes and disorders remains poorly investigated.
Summary
In the majority of factor-analytic studies, ADHD symptoms (inattention, hyperactivity and impulsivity) are found to represent separate but correlated factors from oppositional behaviour and conduct problems. This suggests that they exist as separate dimensions or traits.
When symptom clusters were considered using statistical approaches that aim to identify symptoms that group together, ADHD symptoms were found to group with oppositional behaviour in two studies that used LCA; but in another study using a cluster-analytic approach, two groups of children with ADHD symptoms were identified, one group where ADHD symptoms occurred with oppositional behaviour and a separate group where ADHD symptoms were not accompanied by oppositional behaviour. The GDG concluded that on the basis of these findings, symptoms of ADHD and oppositional and conduct problems represent distinct but correlated sets of behaviours that often coexist. The relationship of ADHD symptoms and oppositional and conduct problems cannot be clearly defined on the basis of statistical analysis of child behaviour that makes use of cross-sectional data alone.
One study using longitudinal data suggested that ADHD represents a separate condition that is a risk factor for the development of oppositional and conduct problems, since ADHD came first and was associated with the future development of oppositional/conduct problems, whereas the reverse situation of oppositional/conduct problems leading to ADHD did not occur. There was, however, no other similar study with which to compare this result.
Twin studies suggest overlapping genetic influences on ADHD and conduct problems, but there are also shared environmental influences on oppositional defiant disorder/conduct disorder that do not act on ADHD. Twin studies of ADHD and oppositional defiant disorder/conduct disorder show different patterns of twin correlations suggesting the existence of shared environmental influences on oppositional defiant disorder/conduct disorder but not on ADHD. This suggests that some aspect of the environment shared by children in the same family increases the risk for oppositional defiant disorder/conduct disorder but not the risk for ADHD; this indicates a separation between the two at the level of aetiological risk factors.
The correlation between ADHD and several neurodevelopmental traits (cognitive ability, reading ability, developmental coordination and pervasive developmental disorders) is due largely to the effects of shared genetic influences. For this reason ADHD may be viewed as one component of a general propensity to neurodevelopmental problems that arises from shared aetiological influences.
In adults, coexisting symptoms, syndromes and disorders are frequently found to exist alongside the core ADHD syndrome, but their distinction from ADHD and the reasons for high rates of coexistence are not well addressed in the current literature.
5.5.3. Are the phenomena of hyperactivity, inattention and impulsivity distinguishable from the normal spectrum?
No systematic reviews were identified that were of direct relevance to this question. The previous search for primary studies revealed two factor-analytic studies relevant to this question. The GDG identified further factor-analytic and quantitative genetic studies that addressed this question (see Appendix 17).
Evidence
Many studies have found a strong correspondence between quantitative measures of ADHD symptoms and the categorical diagnosis (Biederman et al., 1993; Biederman et al., 1996; Boyle et al., 1997; Chen et al., 1994; Edelbrock et al., 1986). These studies show that children with ADHD appear to be at one extreme of a quantitative dimension of ADHD symptoms in the population and that on this quantitative dimension of symptoms there is no obvious bi-modality that separates children with ADHD from children who do not have ADHD.
Twin studies using individual differences approaches (reviewed in Thapar et al., 1999; Faraone et al., 2005) and De Fries-Fulker (DF) extremes analysis (Gjone et al., 1996; Levy et al., 1997; Willcutt et al., 2000; Price et al., 2001) estimate similar magnitudes for the proportion of genetic, shared environmental and non-shared environmental influences on ADHD symptoms in general population twin samples. These studies indicate that aetiological influences on ADHD symptoms are distributed throughout the population and there is no obvious threshold or cut-off between people with high levels of ADHD symptoms and the continuous distribution of symptoms throughout the population. These studies do not take impairment into account, but only investigate the proportion of genetic and environmental influences on ADHD symptom counts.
Using LCA, ADHD symptoms can be divided into multiple groups, distinguished on the basis of three symptom groupings: inattention, hyperactivity-impulsivity and the combination of these two symptom domains. In addition, the symptom groups are separated on the basis of low, medium and high levels into distinct severity groups. Twin data from female adolescents in Missouri and children in Australia both found a similar pattern of familial segregation for the latent classes suggesting that familial influences can distinguish between ADHD and the normal range of behaviour (Rasmussen et al., 2004). These data provide evidence for the distinction of ADHD into inattentive, hyperactive-impulsive and combined subtypes and suggest that ADHD might be distinguishable from the normal range on the basis of familial risks for the observed symptom clusters.
Summary
Most analytic approaches are unable to make a clear distinction between the diagnosis of ADHD and the continuous distribution of ADHD symptoms in the general population. Twin studies suggest that the genetic and environmental influences on groups with high levels of ADHD symptoms are of the same magnitude as those that influence ADHD symptom levels in the normal range. It is not yet known whether the same specific factors are involved, but the studies using DF analysis suggest that there are at least some overlapping genetic influences on ADHD symptoms and the continuity of ADHD symptoms throughout the population.
Twin studies have in most cases defined ADHD on the basis of symptom criteria alone. It is not yet known whether the results would be different if full diagnostic criteria, including impairment, were to be applied. In contrast, LCA can distinguish groups with high, moderate and low levels of ADHD symptoms and suggests that these groups can be distinguished on the basis of familial risks. The current literature does not address the difference in interpretation of the latent class and quantitative approaches.
The GDG concluded that on the basis of current evidence, ADHD was similar to other common medical and psychiatric conditions that represent the extreme of dimensional traits, such as hypertension, obesity, anxiety and depression. The disorder can therefore only be defined on the basis of high levels of symptoms and their association with significant clinical impairments and risk for development of future impairments.
5.6. IS THE CLUSTER OF SYMPTOMS THAT DEFINES ADHD ASSOCIATED WITH SIGNIFICANT CLINICAL AND PSYCHOSOCIAL IMPAIRMENTS?
There were no systematic reviews that addressed this question. A search for cohort studies was carried out and additional primary studies were identified by the GDG members (see Appendix 17).
5.6.1. Evidence
Academic difficulties
Follow-up studies of people diagnosed with ADHD in childhood have consistently indicated impairment in their academic functioning. Children and young people with ADHD have been shown to have greater impaired attention, less impulse control, and greater off-task, restless and vocal behaviour (Fischer et al., 1990). They also have higher rates of both specific and generalised learning disabilities, poor reading skills (McGee et al., 1992) and speech and language problems (Hinshaw, 2002) when compared with healthy controls. These impairments often lead to grade retention (Hinshaw, 2002), to a lower probability of completing schooling when compared with children who do not have ADHD (Mannuzza et al., 1993), suggesting potential long-term ramifications for vocational, social and psychological functioning into adulthood (Biederman et al., 1996; Young et al., 2005a & b; Wilson & Marcotte, 1996).
An important question about educational impairment of children with ADHD is whether, given an appropriate educational environment, this is determined primarily by the presence of high levels of ADHD symptoms or the association with coexisting behavioural conditions such as conduct disorder or learning disabilities. Wilson and Marcotte (1996) found that the presence of ADHD in young people increased the risk for lower academic performance and poorer social, emotional and adaptive functioning, but that the additional presence of conduct disorder further increased the risk for maladaptive outcomes. In another study the association of conduct disorder with academic underachievement was found to be because of its comorbidity with ADHD (Frick et al., 1991).
Family difficulties
Impaired family relationships have been reported in families of children with ADHD. Follow-up studies indicate that mothers of children and young people with ADHD have more difficulty in child behaviour management practices and in coping with their child’s behaviour (August et al., 1998), and display higher rates of conflict behaviours, such as negative comments, social irritability, hostility and maladaptive levels of communication and involvement (August et al., 1998; Fletcher et al., 1996).
Family impairment also permeates the parents’ lives. Parents of children with ADHD report having less time to meet their own needs, fewer close friendships, greater peer rejection, less time for family activities, factors which together might lead to less family cohesion and a significant effect on the parents’ emotional health (Bagwell et al., 2001).
Coexisting conduct and emotional problems may drive the association between maternal expressed emotion (negativity, resentment and emotional over-involvement) and ADHD (Psychogiou et al., 2007).
Social difficulties
Girls with ADHD tend to have fewer friends (Blachman & Hinshaw, 2002) and more problems with peers and the opposite sex (Young et al., 2005a & b). Hyperactive children with or without conduct problems have higher rates of problems with peers and higher rates of social problems because of lack of constructive social activities (Taylor et al., 1996). In a study by Ernhardt and Hinshaw (1994) it was reported that a diagnosis of ADHD significantly predicted peer rejection; having said that, aggressive and non-compliant disruptive behaviours were important and accounted for 32% of the variance in peer rejection.
Antisocial behaviour
Antisocial behaviour is more prevalent in children and young people with ADHD than non-ADHD groups. Some studies show increased rates of antisocial acts (for example, drug misuse) in comparison with children who do not have ADHD (Barkley et al., 2004; Mannuzza et al., 1998).
Follow-up studies have also shown that people with high levels of ADHD symptoms had significantly higher juvenile and adult arrest rates than normal control boys (Satterfield & Schell, 1997). Young adults with a diagnosis of ‘hyperactivity’ in childhood were more likely to have a diagnosis of antisocial disorder (32% versus 8%) and drug misuse (10% versus 1%) than were healthy controls at follow-up (Mannuzza et al., 1991).
ADHD is also a risk factor for psychiatric problems including persistent hyperactivity, violence and antisocial behaviours (Biederman et al., 1996; Taylor et al., 1996) and antisocial personality disorder (Mannuzza et al., 1998).
In a prospective follow-up of 103 males diagnosed with ADHD, the presence of an antisocial or conduct disorder almost completely accounted for the increased risk for criminal activities. Mannuzza and colleagues (2002) reported that antisocial disorder was more prevalent in children with pervasive and school-only ADHD. Lee and Hinshaw (2004), however, reported that the predictive power of ADHD status to adolescent delinquency diminishes when key indices of childhood externalising behaviour related to ADHD are taken into account.
Boys with ADHD and high defiance ratings show significantly higher felony rates than healthy controls (Satterfield et al., 1994). However, ADHD diagnosed in childhood increases the risk of later antisocial behaviour even in the absence of oppositional defiant disorder or conduct disorder (Mannuzza, 2004).
Adolescent and adult problems
A 10-year prospective study of young people with ADHD found that the lifetime prevalence for all categories of psychopathology were significantly greater in young adults with ADHD compared with controls. This included markedly elevated rates of antisocial, addictive, mood and anxiety disorders (Biederman et al., 2006b).
In adolescence and adult life, symptoms of ADHD begin to associate with other diagnoses that are seldom made in childhood. Adolescent substance misuse, in particular, seems to be more common in people with the diagnosis of ADHD (Wilens et al., 2003), though it is not yet clear whether it is the ADHD per se that generates the risk or the coexisting presence of antisocial activities and peer groups.
Both cross-sectional epidemiological studies and follow-up studies of children with ADHD show increased rates of unemployment compared with controls (Biederman et al., 2006b; Kessler et al., 2006; Barkley et al., 2006). Adults with ADHD were found to have significantly lower educational performance and attainment, with 32% failing to complete high school; they had been fired from more jobs and were rated by employers as showing a lower job performance (Barkley et al., 2006). The survey from Biederman and colleagues (2006b) showed that 33.9% of people with ADHD were employed full time versus 59% of controls.
An increased rate of road traffic violations and driving accidents in adults with ADHD has been documented by several authors (Reimer et al., 2007; Barkley and Cox, 2007; Thompson et al., 2007; Jerome et al., 2006; Fischer et al., 2007).
5.6.2. Summary
ADHD symptoms are associated with a range of impairments in social, academic, family, mental health and employment outcomes. Longitudinal studies indicate that ADHD symptoms are predictive of both current and future impairments. Impairments also result from the presence of coexisting problems including conduct problems, emotional problems and overlapping neurodevelopmental disorders. Adults with ADHD are found to have lower paid jobs and lower socioeconomic status and have more car accidents. Impairment is an essential criterion when considering the diagnosis of ADHD. The presence of high levels of ADHD symptoms is associated with impairment in multiple domains; it is not possible, however, to delineate clearly a specific number of ADHD symptoms at which significant impairment arises.
5.7. IS THERE EVIDENCE FOR A CHARACTERISTIC PATTERN OF DEVELOPMENTAL CHANGES, OR OUTCOMES ASSOCIATED WITH THE SYMPTOMS, THAT DEFINE ADHD?
The search for systematic reviews and meta-analyses identified one review that was of relevance to this question. Additional reviews and primary studies were identified by the GDG members (see Appendix 17).
5.7.1. Evidence
There is evidence for continuity of ADHD symptoms over the lifespan. Faraone and colleagues (2006) analysed data from 32 follow-up studies of children with ADHD into adulthood. Where full criteria for ADHD were used approximately 15% of children were still diagnosed with ADHD at age 25. In addition, the meta-analysis found that approximately 65% of children by age 25 fulfilled the broader definition of DSM-IV ADHD ‘in partial remission’, indicating persistence of some symptoms of ADHD associated with continued clinically meaningful impairments.
Relative to controls, levels of overactivity and inattention are developmentally stable (Taylor et al., 1996). Longitudinal studies of children with ADHD show similar rates of ADHD in adolescence (Biederman et al., 1996; Faraone et al., 2002; Molina & Pelham, 2003).
Population twin studies have also addressed the stability of ADHD symptoms throughout childhood and adolescence. Rietveld and colleagues (2004) reported that parent ratings of attentional problems were moderately stable from age 3 to 7, and greater stability from age 7 to 10. They further showed that such stability appeared to be mediated largely by overlapping genetic influences such that most, but not all, genetic influences at one age influenced ADHD at another age. Price and colleagues (2005) reported similar findings with correlations around 0.5 between ADHD symptoms at ages 2, 3 and 4. This stability was estimated to be mediated 91% by genetic influences. Kuntsi and colleagues (2004) extended these data to age 8, and found similar moderate stability between the data for ages 2, 3 and 4 and the data for age 8. Larsson and colleagues (2004) completed a similar longitudinal twin study of 8 to 13 year olds and found fairly high stability between the two ages. They further concluded that this stability was the result of shared genetic effects. Change in symptoms between childhood and adolescence was thought to be because of new genetic and environmental effects that become important during adolescence.
5.7.2. Summary
There is evidence for the persistence of ADHD symptoms from early childhood through to adulthood. Longitudinal studies confirm that ADHD persists into adulthood but developmentally appropriate criteria have yet to be developed for ADHD in adults. Using child criteria, approximately 15% of children with ADHD retain the diagnosis by age 25 but a much larger proportion (65%) are in partial remission, with persistence of some symptoms associated with continued impairments. The profile of symptoms may alter with a relative persistence of inattentive symptoms compared with hyperactive-impulsive symptoms. The evidence base for this conclusion is poor, however; it is based on the analysis of developmentally inappropriate measures of hyperactivity-impulsivity in adults.
The GDG concluded that there is currently insufficient evidence to warrant a different diagnostic concept in childhood and in adulthood. Having said that, it is envisaged that improved definitions that take into account developmental changes will develop as further evidence is accrued. Familial and genetic influences in ADHD symptoms appear to be stable through childhood and early adolescence, but there is a lack of data on the factors that modify the course of ADHD into adulthood.
5.8. IS THERE CONSISTENT EVIDENCE OF GENETIC, ENVIRONMENTAL OR NEUROBIOLOGICAL RISK FACTORS ASSOCIATED WITH ADHD?
The literature search identified eight systematic reviews and meta-analyses. GDG members identified additional reviews and primary studies (see Appendix 17). When interpreting this section it is important to note that associations do not imply causal associations and may represent epiphenomena of ADHD rather than causal processes.
5.8.1. Evidence
Cognitive experimental studies
Willcutt and colleagues (2005) reviewed 83 studies that had administered executive functioning measures and found significant differences between ADHD and non-ADHD groups where the former showed executive function deficits. The size of the difference between children with ADHD and unaffected controls, while significant, was moderate rather than large. The term executive function refers to a set of higher cognitive and emotional mental functions involved in the control and regulation of behaviour and performance. This includes concepts such as cognitive inhibition and initiation, self-regulation and motor output. The neural mechanisms by which the executive functions are implemented is a topic of ongoing debate in the field of cognitive neuroscience. It is not yet clear whether impairments in the performance of executive tasks is because of primary deficits in the brain processes underlying executive functions, or whether the performance deficits are secondary to more general processes.
Differences in executive functioning between ADHD and non-ADHD groups have also been reported in adults (Hervey et al., 2004; Boonstra et al., 2005; Schoechlin & Engel, 2005; Woods et al., 2002). The results of studies of ADHD in adults suggest a wide variety of general and specific performance on cognitive-experimental tasks that are similar to those seen in children with ADHD. The review from Hervey and colleagues (2004) did not point to impairments in one area of cognitive performance, but rather impairments across a range of cognitive functions.
The interpretation of cognitive-experimental studies in ADHD remains controversial, but most authorities agree that both executive and non-executive processes are disrupted in people with ADHD. Although work has largely focused on the executive functions, there is an interest in non-executive processes (Rhodes et al., 2006; Berwid et al., 2005). A recent meta-analysis of the stop-signal paradigm concluded that there are significantly slower mean reaction times, greater reaction time variability and slower stop signal reaction times in children with ADHD relative to controls (Alderson et al., 2007). The pattern of findings suggested a more generalised impairment of attentional and cognitive processing rather than a primary deficit of behavioural inhibition alone. Recently it has emerged that intra-individual variability is one of the more consistent associations with ADHD in both children and adults (Klein et al., 2006).
In an adoptive study conducted by Sprich and colleagues (2000), higher rates of hyperactivity were found in the biological parents of children with ADHD compared with their adoptive parents.
Neuroimaging studies
In an attempt to provide a robust summary of available functional magnetic resonance imaging (fMRI) studies, Dickstein and colleagues (2006) performed a quantitative meta-analysis of task-based imaging studies using 13 fMRI studies and four positron emission tomography (PET)/single-photon emission computed tomography (SPECT) studies that had published stereotactic space coordinates. The meta-analytic data showed reduced activation in regions in the left pre-frontal cortex, the anterior cingulate cortex, the right parietal lobe, the occipital cortex and in the thalamus and claustrum. When only response inhibition studies were included in the analysis, a more restricted network was identified, which included the right caudate (part of the striatum). The analysis also identified certain regions where the ADHD groups tended to show hyperactivation: these included parts of the left pre-frontal cortex, the left thalamus and the right paracentral lobule. The extent of neural networks remains uncertain since the available data were limited by the narrow selection of tasks. A major limitation was the small number of suitable datasets and the unavoidable inclusion of studies that differed in the specific aspects of design and quality.
A systematic review of available fMRI studies in ADHD reached several conclusions (Paloyelis et al., 2007). First, in tasks that examined brain activation during successful inhibitory control, there were large inconsistencies among studies in the direction of group differences. Group differences were also spread across many different brain regions, but the frontal lobes were predominantly involved. For this reason no firm conclusions can be drawn on the association of brain activation changes during response inhibition tasks in ADHD. Second, in analyses that examined inhibition errors, as well as in tasks that tapped attention processes, motor function and working memory, the ADHD group almost exclusively showed lower brain activity; in the attentional tasks this was mostly over temporal and parietal areas; in motor function tasks mostly over frontal areas. Third, among the different brain regions, the most consistent findings as regards direction of activation were observed in the striatum. In all but one study significant group differences were observed in which the ADHD group showed lower activity in the striatum. The only study where increased activation was observed had used a sample of young people of whom only half met full criteria for ADHD at the time of testing. Fourth, the review included a summary of findings from people with ADHD who had not used stimulant or other medication. These studies suggest that altered brain activation patterns in children with ADHD are not due to the effects of long-term stimulant treatment. Pliszka and colleagues (2006) was the only study to compare individuals with ADHD on long-term medication with those that were drug naïve as well as healthy controls. The study found no differences between the treated and untreated ADHD groups on most comparisons. Where some differences were found the treated group was more similar to controls than the untreated group.
A systematic meta-analytic study of brain structural changes in ADHD analysed all brain regions reported by all the studies found (Valera et al., 2007). The study found global reductions in brain volume in ADHD cases compared with controls. Regions most commonly assessed and showing the largest differences included cerebellar regions, the splenium of the corpus callosum, total and right cerebral volume and right caudate. Several frontal regions examined in only two studies also showed significant differences. It was not possible to include or exclude the role of medication in the observed changes to brain volume and structure.
Molecular genetic studies
A systematic meta-analysis of molecular genetic association for associated markers in or near to the dopamine D4 (DRD4), dopamine D5 (DRD5) and dopamine transporter (DAT1) genes, found strong evidence for the association of DRD4 and DRD5 but not DAT1 (Li et al., 2006). Although there are many other individual and meta-analytic studies of genetic findings in ADHD, Li and colleagues (2006) compiled most of the available data for three of the best-studied findings to date, and found significant levels that were in excess of that expected from scanning the entire human genome: 8 × 10−8 for DRD5 and 2 × 10−12 for DRD4. A significance level close to 5 × 10−8 is widely accepted to indicate a true association after adjusting for the number of potential false positive findings in a scan of the entire human genome (for example, Risch & Merikangas, 1996). Other reported genetic associations with ADHD, including DAT1, do not reach this level of significance in the literature and cannot be confirmed or refuted at this time. The level of risk associated with DRD4 and DRD5 is small with odds ratios in the order of 1.2 to 1.4. This level of risk is similar to that seen for genetic influences in common medical conditions such as diabetes (Altshuler & Daly, 2007). As with all other types of risk factor associated with ADHD, the individual genetic variants associated with the disorder are neither sufficient nor necessary to cause it, but contribute a small increase to the overall risk for ADHD.
Quantitative genetic studies
A systematic review of 20 population twin studies found an average heritability estimate of 76%. In most cases, heritability in these studies is estimated from the difference in the correlations for ADHD symptoms between identical and non-identical twin pairs, as reported by parents and teachers: with the correlation for identical twin pairs in the region of 60 to 90% and for non-identical twin pairs being half or less than half of this figure in most studies (Faraone, 2005). Under the equal environment assumption for the two types of twin pairs, heritability can be estimated as twice the difference in the two sets of correlations.
The assumption of ‘equal environment’ for identical and non-identical twins can be questioned. If it were not valid, then the estimated effect of genetic influences would decrease and that of shared environmental influences would increase. Even if this were to be the case, however, it would not argue against the validity of the disorder. It is not in doubt that twins’ scores are highly correlated – the level of ADHD symptoms in one child predicts that in the other. This tendency to run in families supports the idea that it is a coherent syndrome, whether the reasons are genetic or environmental.
Sibling correlations (the similarity between two siblings) can arise from either shared environmental or shared genetic influences. The equal environment assumption impacts on the estimate of the proportion of the familial risk that is due to genes or shared environment (for example, Horwitz et al., 2003). Because the estimated heritability of ADHD is less than 100% we know that environmental influences are likely to cause differences in siblings and contribute to why one child in a family might have ADHD while another child does not (so-called unique environmental effects). High heritability and low shared environmental factors estimated by twin studies do not exclude an important additional contribution of the environment, acting through mechanisms of gene-environment interaction (Moffitt et al., 2005) or gene-environment correlation (Jaffee & Price, 2007). Much more work is needed to understand the complex interplay of genetic and environmental influences on the risk for ADHD.
Evidence for genetic influences also comes from adoption research. One study showed increased rates of ADHD among the biological parents of non-adopted children with ADHD when compared to adoptive parents of children with ADHD and biological parents of non-adopted children who did not have ADHD (Sprich et al., 2000). To date there are no published studies that compare the adoptive and biological parents of adopted children.
Physical environmental risk studies
Schab and Trinh (2004) completed a systematic meta-analysis of the effect of exposure to food additives on ADHD symptoms. They identified 15 studies that met initial inclusion criteria and estimated an effect size of around 0.2, but many of the studies were either of a non-ADHD sample or sample sizes were very small (n < 10) and/or were not properly randomised. The authors report associations between the use of food additives and ADHD, but given the limitations of the studies included it is difficult to establish a clear conclusion.
More recently in the UK, Stevenson and colleagues (McCann et al., 2007) completed a double-blinded placebo-controlled crossover trial of food additives in 3-year-old and 8/9-year-old children. This study confirmed the association between food additives (artificial colours, sodium benzoate, or both) on increased levels of ADHD symptoms in the child populations studied. These studies indicate short-term toxic effects of food additives on the level of ADHD symptoms in children whether they have ADHD or not and might contribute towards significant impairment in some cases. There is no indication that food additives cause long-term effects on child development.
Linnet and colleagues (2003) completed a systematic review of the evidence for association between prenatal exposure to nicotine, alcohol, caffeine and psychosocial stress. They concluded that exposure in utero to the consequences of tobacco smoking is associated with an increased risk for ADHD. In contrast contradictory findings were found for the risk from prenatal maternal use of alcohol and no conclusions could be drawn from the use of caffeine. Studies of psychosocial stress indicated possible but inconsistent evidence for an association with ADHD.
Talge and colleagues (2007) completed a systematic review of studies that indicate the association of antenatal maternal stress on aspects of child development including ADHD symptoms, emotional and cognitive problems, anxiety and language delay. These effects appear to be independent of postnatal depression and anxiety. Two studies identified an increase in ADHD symptoms in children between the ages of 4 and 15 (O’Connor et al., 2002; van den Bergh and Marcoen, 2004). The effect size of the association was marked. Van den Bergh and Marcoen estimated that 22% of the variance in symptoms of ADHD was accounted for by maternal anxiety during pregnancy. O’Connor and colleagues (2002
O’Connor and colleagues (2003) found that women in the top 15% for symptoms of anxiety at 32 weeks’ gestation increased the risk of symptoms of ADHD, conduct disorder, anxiety or depression by 5 to 10%. Prenatal maternal stress is therefore associated with an increase in ADHD symptoms but is not specific to ADHD. The mechanisms involved in this association are poorly understood.
Non-physical environmental risk studies
As stated in the section on associated impairments, impaired family relationships have been reported in families of children with ADHD. Follow-up studies indicate that mothers of children and young people with ADHD have more difficulty in child behaviour management practices and coping with their child’s behaviour (August et al., 1998), and display higher rates of conflict behaviours, such as negative comments, social irritability, hostility and maladaptive levels of communication and involvement (August et al., 1998; Fletcher et al., 1996).
Persistent problems with inattention and overactivity have been documented in a sample of institution-reared children adopted from Romania before the age of 43 months. The syndrome of inattention and overactivity was strongly associated with early institutional deprivation lasting 6-months or more, with higher rates in boys than girls, and was strongly associated with conduct problems, disinhibited attachment and executive function impairments (Stevens et al., 2008; Rutter & O’Connor, 2004).
In general, the diagnosis of ADHD is distributed unequally across different levels of deprivation and is mediated by social class and ethnicity (Bauermeister et al., 2005; Cunningham & Boyle, 2002). Maltreatment has been associated with higher rates of ADHD in addition to oppositional behaviour and post-traumatic stress disorder (Famularo et al., 1992). McLeer and colleagues (1994) found very high rates of ADHD (46%) among children with a history of sexual abuse.
Adversity in the form of familial risk factors has also been shown to be associated with ADHD (Biederman et al., 1995). In a sample of clinical cases of ADHD, exposure to parental psychopathology and exposure to parental conflict were used as indicators of adversity, and their impact on ADHD and ADHD-related psychopathology and dysfunction in children was assessed. The analyses showed significant associations between the index of parental conflict and several of the measures of psychopathology and psychosocial functioning in the children confirming the role of adversity on the risk for ADHD and its associated impairments.
Work by Rutter and colleagues (1975) revealed that it was the aggregate of adversity factors (severe marital discord, low social class, large family size, paternal criminality, maternal mental disorder and foster care placement) rather than the presence of any single factor that led to impaired child development (Rutter et al., 1975). Based on this work, Biederman and colleagues (1995), using a sample of 140 ADHD and 120 normal control probands and using Rutter’s indicators of adversity, investigated whether family-environment risk factors were associated with ADHD. A positive association was found to exist between adversity indicators and the risk for ADHD as well as for its associated psychiatric, cognitive, and psychosocial impairments, supporting the importance of adverse family-environment variables as risk factors for children with ADHD.
5.8.2. Summary
There is consistent evidence from family, twin and adoption studies of both genetic and environmental influences on ADHD symptoms throughout the population. Under the equal environment assumption, twin studies indicate that sibling similarity for ADHD symptoms results mainly from genetic influences. Some supportive evidence is given by adoptive research. Unique environmental influences play a role in bringing about differences in ADHD symptoms within families. Environment may also play an important role in ADHD acting through mechanisms of gene-environment interaction and correlation. Environmental measures associated with ADHD have been identified, including maternal use of tobacco during pregnancy and prenatal maternal stress. Other associated environmental measures include early deprivation, maltreatment and sexual abuse, family factors including severe marital discord, low social class, large family size, paternal criminality, maternal mental disorder and foster care placement. Some dietary components have been shown to increase the level of ADHD symptoms in children and are expected to contribute to increased levels of ADHD symptoms in all children. These may give rise to increased symptoms and impairments in a sub-group of individuals who go on to develop ADHD, although this has yet to be clearly demonstrated.
The causal relationships between environmental measures and ADHD are not well understood. In most cases it is not known whether specific associated environmental variables represent direct risks for ADHD, or indirect risks acting through correlated environmental or genetic factors, or are passively correlated with the ADHD symptoms themselves.
The GDG concluded that specific genetic variants associated with small increases in the risk for ADHD have been identified within the dopamine D4 receptor gene and close to the dopamine D5 receptor gene. These are the only two genetic findings where convincing levels of evidence have accrued as demonstrated by the recent meta-analytic study from Li and colleagues (2006). Other genetic findings require further data before they can be included or refuted as true associations with ADHD.
Analysis of ADHD versus non-ADHD groups has identified consistent changes in brain structure, function and performance on neurocognitive tests; however differences from controls are not universal, do not characterise all children and adults with a clinical diagnosis of ADHD, and do not usually establish causality in individual cases. The degree to which the observed heterogeneity in the associations with neurobiological and psychological measures represent multiple aetiological contributions to a common causal pathway, or independent contributions to multiple causal pathways, is not yet understood. It may also be the case that these associations represent epiphenomena of the ADHD syndrome and play no direct causal role.
5.9. LIMITATIONS
In line with methodology agreed with NICE, the approach adopted initially was to identify all available systematic reviews and meta-analytic studies that related to the questions on validity of the diagnosis. While this was possible for much of the neurobiological, genetic and environmental data, there were few systematic reviews in other areas such as the factor- or cluster-analytic studies. Where systematic reviews were not available for the studies of ADHD symptoms and studies that investigated the differentiation of ADHD from oppositional-defiant and conduct problems, a systematic review of the primary literature was conducted. For the interpretation of factor and cluster analytical approaches it is important to recognise the limitations that arise from the high variability in quality of these types of exploratory statistical analyses papers. Factor- and cluster-analytic methods require a certain degree of unstructured judgments to be made by researchers, rarely produce reproducible results and in the majority of cases were underpowered. Despite this, as outlined in the evidence, a reasonable level of reproducibility in the findings was observed.
For other sub-questions addressed in this section, the systematic evidence was supplemented with expert opinion, drawing on evidence known to members of the GDG. Additional evidence was obtained following a review of the initial draft of this chapter by independent experts (see Appendix 16 for their commentary). The lack of specific reference standards for the diagnosis of ADHD led to an adaptation of the SIGN criteria to ensure sufficient quality of the data used to derive recommendations for this guideline. The revised criteria agreed by the GDG members were as follows: (1) the study addresses an appropriate and clearly focused question (or hypothesis); (2) the sample population being studied is selected either as a consecutive series or randomly, from a clearly defined population.
When considering the Washington University Diagnostic Criteria (Feighner et al., 1972) for validity of a psychiatric disorder, the question of whether there are characteristic responses to pharmacological, psychological, educational and other interventions for ADHD was excluded from this section, because the response of ADHD to these interventions is considered in detail elsewhere in this guideline. The related question of the specificity of the response to therapeutic interventions for ADHD was surprisingly difficult to determine on the basis of available published evidence. For example, behavioural, educational and pharmacological treatments can all alter the behaviour of children whether they have ADHD or not.
In relation to the use of stimulants we were unable to identify studies that investigated their effects on mental health disorders other than ADHD. The GDG identified a literature on the misuse potential of stimulants, indicating that methylphenidate and dexamfetamine increase ratings of subjective activity, alertness (wakefulness) and energetic and high feelings (for example, Stoops et al., 2004), but there were no direct comparisons with the effects of people fulfilling diagnostic criteria for ADHD. One paper was identified that addressed the effects in a normal population; it did not meet the quality control criteria for the evidence sections of this chapter, but it is mentioned here because of its potential importance. The authors reported the response to dexamfetamine and placebo in a group of 14 pre-pubertal boys who did not fulfil criteria for ADHD (Rapoport, 1978). When amphetamine was given, the group showed a decrease in motor activity and reaction time and improved performance on cognitive tests that was similar to that seen in other studies of children with ADHD. The very small numbers used in this study and lack of further similar studies means that caution must be taken in drawing firm conclusions from this one study. Nevertheless, the similarity of the response observed in children without ADHD to that reported in children with the disorder provides further evidence that the aetiological processes in ADHD are similar to those that influence levels of ADHD symptoms throughout the population.
The question of a paradoxical effect of stimulants on people with ADHD has been raised but is not well studied. For example, do stimulants have an impact on the same processes and in the same way in all people, whether they have ADHD or not? Or is there a different pattern of effects in people with high levels of ADHD symptoms compared with people with low levels? The GDG concluded that the critical question for these guidelines is whether stimulants and other non-pharmacological interventions effectively treat the impairments associated with high levels of ADHD symptoms. The effectiveness and cost benefits of these interventions are addressed in other sections of this guideline.
5.10. SUMMARY OF VALIDATION OF THE DIAGNOSIS OF ADHD
The diagnosis of ADHD is difficult and somewhat controversial for a number of reasons. Of particular concern has been the rapid increase in the recognition and treatment of children with ADHD and the very high prevalence rates reported in some studies, leading some people to question the validity of the disorder. In common with most mental health conditions there is no definitive biological test for ADHD; diagnosis depends on the observation of clusters of symptoms in three main behavioural domains according to the DSM-IV and ICD-10 criteria. In order to examine the validity of the diagnosis, the Washington University Diagnostic Criteria (Feighner et al., 1972) were applied to demonstrate whether there are well-defined clinical correlates, characteristic course and outcome, neurobiological underpinnings and associations with genetic and environmental factors. The review above identified clinical, genetic, environmental and neurobiological factors associated with ADHD or correlated with levels of ADHD symptoms in the general population that were sufficient to validate the diagnostic construct of ADHD.
One of the key issues addressed in the review was the question of whether ADHD represents a discrete clinical entity or the extreme end of a continuum of normal behaviour. Indeed, the debate between a categorical diagnostic view and a dimensional approach is longstanding in psychological and sociological research. The diagnosis of many common psychological conditions, such as anxiety and depression represents a line drawn at one end of a continuum of a population characteristic that is continuously distributed throughout the population; the threshold for diagnosis being drawn at a point where significant impairment arises.
The review concluded that on the basis of current evidence, ADHD is best conceptualised as the extreme of a continuous trait that is distributed throughout the population; the distinction from normality being made by the presence of high levels of ADHD symptoms when they are accompanied by significant impairments. This highlighted the importance of defining what amounts to a significant impairment and ensuring that impairment is fully evaluated when applying the diagnostic criteria.
5.11. DEFINING SIGNIFICANT IMPAIRMENT
The GDG wished to define more precisely the level of impairment indicating when the guidelines should be triggered. The GDG recognised the breadth of views on what amounts to a significant impairment. The existence of polarised views in this debate, and the implication for both under-and over-diagnosis, means that a balanced and pragmatic view is required that takes into account concerns on both sides. For example the GDG recognised that people with hyperkinetic disorder (ICD-10) do not always receive a diagnosis and treatment despite the presence of marked impairments, while on the other hand in some cases stimulants have been used to boost academic performance in the absence of more pervasive and enduring impairments. The following criteria were discussed and agreed by a consensus within the group:
The GDG wishes to emphasise the importance of significant impairment in defining the difference between a set of mental health problems and a mental health disorder. An appreciation of this difference is helpful in preventing over-diagnosis. In addition, the diagnosis of ADHD should not be applied to justify the use of stimulant medication for the sole purpose of increasing academic performance, in the absence of a wider range of significant impairments indicating a mental health disorder.
Many mental health problems, including those with ADHD features, are transitory and related to psychosocial stresses. They often clear up spontaneously or do so after a basic-level intervention by, for example, parents and teachers. In contrast, a mental health disorder implies something far more serious. Without a specialist professional or a higher level of intervention by others to ameliorate the problems, there are likely to be long-term adverse implications for the person affected as well as problems in the short and medium term. It is therefore important that the assessing clinician considers whether the clinical presentation is indicating a threat to general development and psychosocial adjustment that would be more likely than not to occur if expert help or some other significant intervention was not to take place. This would apply to the current presentation and also the longer-term outlook.
The GDG concluded that impairment should be pervasive, occur in multiple settings and be at least of moderate severity. Significant impairment should not be considered where the impact of ADHD symptoms are restricted to academic performance alone, unless there is a moderate to severe impact in other domains: these would include self-esteem, personal distress from the symptoms, social interactions and relationships, behavioural problems, and the development of coexisting psychiatric syndromes.
5.12. POSITION STATEMENT ON THE VALIDITY OF ADHD
On the basis of the evidence reviewed above the GDG drew the following conclusions:
Symptoms that define hyperactive, impulsive and inattentive behaviours are found to cluster together.
Hyperactivity, inattention and impulsivity cluster together both in children and in adults and can be recognised as distinct from other symptom clusters, although they frequently coexist alongside other symptom clusters.
Symptoms of ADHD appear to be on a continuum in the general population.
ADHD is distinguished from the normal range by the number and severity of symptoms and their association with significant levels of impairment.
The importance of evaluating impairment and the difficulty in establishing thresholds on the basis of symptom counts alone needs to be addressed. It is not possible to determine a specific number of symptoms at which impairment arises.
There is evidence for psychological, social and educational impairments in both children and adults with ADHD.
ADHD symptoms persist from childhood through to adulthood in the majority of cases. In a significant minority the diagnosis persists and in the majority, subclinical symptoms continue to be detectable and are associated with significant impairments.
In adults the profile of symptoms may alter with a relative persistence of inattentive symptoms compared with hyperactive-impulsive symptoms.
There is evidence of both genetic and environmental influences in the aetiology of ADHD. The extent to which there is diversity in the aetiology of the disorder is not known. Current evidence indicates the presence of multiple risk factors of minor effect.
The complex interplay between genes and environment is not well understood. Environmental risks may interact with genetic factors, be correlated with genetic factors or have main effects. Similarly genetic factors may interact or correlate with environment or have main effects. There will be a different balance of factors in individual cases.
There is evidence of genetic associations with specific genes, environmental risks and neurobiological changes in groups of children with ADHD. However, no neurobiological, genetic or environmental measure is sufficiently predictive to be used as a diagnostic test.
The diagnosis remains a descriptive behavioural presentation and can only rarely be linked to specific neurobiological or environmental causes in individual cases.
Hyperkinetic disorder (ICD-10) is a narrower and more severe subtype of DSM-IV-TR combined type ADHD. It defines a more pervasive and generally more impairing form of the disorder. Both concepts are useful (
Santosh et al., 2005).
There was limited evidence to support a different concept of ADHD in children and adults. Age-related changes in the presentation are recognised, however. Theses changes are not yet reflected in the current diagnostic criteria.
All current assessment methods have their limitations. There is evidence of the need for flexibility and for a consideration of levels of impairment in assessments and when deriving appropriate diagnoses.
5.13. CONSENSUS CONFERENCE
In addition to a review of published evidence on the question of validity, a consensus conference was held to bring together experts in the field with a range of views, in order to debate the key issues of the use of ADHD as a diagnostic category. The aim was to provide a range of contemporary perspectives that would assist the GDG with the task of deciding what should trigger the use of the guideline and for whom the guideline is intended (see Chapter 3). The speakers delivered a 15-minute presentation addressing the key questions relating to the validity of the ADHD diagnosis set out by the GDG, followed by questioning from the GDG members and a subsequent discussion of the presentation among members of the GDG. Each presenter was subsequently asked to provide a summary of their presentation and these are presented in Appendix 16.
The consensus conference involved presentations from professionals who came from a range of backgrounds and with differing perspectives on the validity and aetiology of ADHD. The range of views contributed to highlight the importance of an interdisciplinary approach to the diagnosis and treatment of children and young people with ADHD. The conference did not consider diagnosis and treatment of adults with ADHD.
Here some of the issues that were raised, and the areas of controversy arising from differences in the perceptions of the speakers at the consensus conference, are discussed. Some of the complex areas of controversy relate to broader sociological and philosophical issues representing two conceptual paradigms, broadly characterised as medical–scientific and social–scientific. The latter perspective casts doubts on the utility and legitimacy of ADHD as a diagnostic category by emphasis on: the problematic nature of the meaning of ADHD, the social determinants of the behaviours that come to be labelled as ADHD, and the spectrum of human behaviour that results in indistinct boundaries of many medical diagnostic categories. While it is important to acknowledge the validity of the social scientific paradigm and its body of literature, in the context of the development of practical clinical guidelines, it is not possible to offer alternative processes for clinical assessment or treatment. It is accepted that the research literature reflects the dominant medical scientific paradigm and hence the nature of the evidence base.
The evidence presented at the consensus conference indicated that there was a high degree of unanimity about there being a group of people who could be seen as having distinct and impairing difficulties and who should trigger the use of this guideline. While recognition of a particular group was agreed upon, uncertainty about the breadth of diagnosis was discussed, namely, whether the use of a narrow (ICD-10 hyperkinetic disorder) versus a broad (DSM-IV ADHD) diagnosis should be used. The problems of using a narrow diagnosis are: (i) the under-recognition of people that are in need of help and (ii) the lack of connection with the research literature, which is based mainly on the broader definition of DSM-IV ADHD. It was established that the main differences between people falling into narrow or broad diagnoses are the breadth of symptoms (requirement for both inattentive and impulsive-hyperactive behaviour versus only one domain being sufficient), more or less stringent criteria for situational pervasiveness and the requirement for no major comorbidity (apart from oppositional defiant disorder or conduct disorder) under ICD-10. Both groups present similar problems of impairment. Overall there was general agreement that both the use of broad DSM-IV ADHD diagnosis and narrow ICD-10 hyperkinetic disorder criteria were useful.
It should be emphasised that the current definitions of ADHD are descriptions of a behavioural syndrome with associated mental phenomena, and do not implicate specific causal pathways. Validation of the cluster of symptoms that contribute to the diagnosis of ADHD occur at the level of their association with impairments, familial risks, genetic risks, environmental risks and the association with measures of changes in cognitive function and brain structure and function. Few direct causal inferences have yet been established, however. For example the associations with changes in cognitive and brain function may represent epiphenomena of ADHD rather than imply a causal process. Environmental measures associated with ADHD may not themselves represent direct risk factors, but may be correlated with more proximal environmental or genetic risks. A common conceptualisation is that both intrinsic and extrinsic processes are involved in generating the cluster of behavioural symptoms that we call ADHD. Extrinsic factors, such as parental coping and consistency, might exacerbate problems of behavioural control in a child with intrinsic difficulties in regulating core processes such as attention and activity level. The child’s difficult behaviour may further exacerbate the difficulties in providing consistent parenting. Parental behaviour itself will also be influenced by both genetic and environmental factors, further increasing the complexity of the aetiological relationships involved. Given the complexity of this question, the GDG does not seek here to put forward a particular causal model, but wishes to emphasise the role that both genes and environment play on both intrinsic and extrinsic factors in generating the clinical syndrome of ADHD.
One of the major issues of controversy in the UK setting is the very high and variable prevalence rates reported in the literature. For example, recent prevalence figures range from 6.8 to 15.8% for DSM-IV ADHD (Faraone et al., 2003) while the British Child and Mental Health Survey reported a prevalence of 3.6% in male children and less than 1% in females (Ford et al., 2003). Reasons for this are discussed in Faraone and colleagues (2003) who conclude that prevalence rates derived from symptom counts alone, or from ratings in one setting, were higher than those that took into account functional impairment and pervasiveness. For example Wolraich and colleagues (1998) estimated prevalence to be 16.1% on the basis of symptom counts, but 6.8% when functional impairment was taken into account. A study in the UK that specifically addressed the role of impairment found that among 7 to 8 year olds, 11.1% had the ADHD syndrome based on symptom count alone (McArdle et al., 2004). In contrast, 6.7% had ADHD with Children’s Global Assessment Scale (C-GAS: measuring impairment) scores of less than 71; 4.2% had C-GAS scores of less than 61. When pervasiveness included both parent- and teacher-reported ADHD and the presence of psychosocial impairment, prevalence fell lower to 1.4%. The literature on prevalence therefore indicates that the rate of ADHD is sensitive to the degree of impairment associated with the symptom criteria and the degree to which the disorder shows situational pervasiveness.
All the speakers acknowledged the importance of functional impairments in relation to diagnosis. In other words, the diagnostic threshold should be based on pragmatic grounds such as impairment and the need for treatment. There was also agreement that defining suitable thresholds for impairment is difficult, since different people hold a range of views on what amounts to significant impairment. The fear was expressed that too broad a definition would lead to the over-diagnosis of children as a way of justifying the use of stimulant medication to enhance academic performance, in the absence of a wider range of pervasive and enduring impairments. Given the sensitivity of the prevalence rates of ADHD to definitions of impairment, this could potentially lead to very high numbers of children being treated when educational or psychological interventions might be sufficient, or where the level of impairment does not warrant a therapeutic intervention at all. The GDG concurred with this view, but were equally concerned to ensure that the thresholds for the diagnosis were not so restricted as to leave children with ADHD (who by definition have significant impairment) undiagnosed and therefore untreated.
The level and types of behaviour that define impairment remain a contentious issue and are to some extent dependent on the cultural and environmental context. For this reason expert clinical advice is required to evaluate the level of impairment to ensure that: the child’s view is taken into consideration and not just that of the child’s parents and teachers; that everyone’s perspective is taken into account; and that cultural factors are considered.
Considering when use of this guideline should be triggered, the GDG concluded that it would be difficult to be prescriptive for any individual case, but that measurement of impairment linked to the symptoms of ADHD is a key component of the decision. Significant problems can arise at various levels, including personal distress from symptoms of the disorder, difficulties in forming stable social relationships and emotional bonds, difficulties with education and long-term risk for negative outcomes such as emotional problems, antisocial behaviour and addiction disorders. The GDG concluded that those responsible for initiating diagnosis and treatment must take into account the severity of the disorder in terms of clinical and psychosocial impairments. When monitoring treatment response, evidence of improvement in such impairment is critical and should be monitored in addition to the narrow focus on changes in reported levels of ADHD symptoms.
One of the areas of controversy highlighted in the consensus conference was the degree of impairment and severity of ADHD needed to trigger the diagnosis and, related to this, treatment with medication. Concern was expressed that the diagnosis automatically leads to treatment with medication and this is not always desirable when the breadth of the definition includes people who might gain substantial benefit from education or psychosocial interventions alone. Having said that, even the most ardent supporters of non-pharmacological interventions in ADHD recognised the importance of pharmacological treatment in the most severe cases. In this context the participants in the consensus conference made an important contribution by raising the important question of suitable thresholds for ‘significant impairments associated with ADHD symptoms’ and hence the proportion of children fulfilling criteria for the disorder and triggering use of the guideline. The related issue is the importance of considering the full breadth of effective interventions (including educational, social and psychological support and pharmacological treatment), depending on the severity of the disorder, the extent of impairment and needs of each individual case.
One conclusion is that the acceptable thresholds for impairment are partly driven by the contemporary societal view of what is an acceptable level of deviation from the norm. Impairment in ADHD should not be based only on the views of others because people with ADHD, particularly older adolescents and adults, have strong subjective experience of the impact of their condition on their functioning.
The GDG did not consider that the diagnosis should be reserved only for the most serious cases, however, since the broader concept of ADHD is important in triggering educational and behavioural support in addition to pharmacological approaches. The GDG concluded that defining appropriate thresholds of impairment associated with the disorder was important, but that treatment implications might be different for individuals falling above or below particular thresholds.
Confirmatory factor-analytic studies clarify that ADHD symptoms represent a distinct set of symptoms and behaviours that co-vary together in both clinical and control populations. However, these cross-sectional studies are far less informative than longitudinal studies, which can clarify the predictive outcomes of early ADHD. Having said that, there are a few studies that provide suitable data on the relative outcomes of ADHD and other disruptive disorders such as oppositional defiant disorder, which are important in delineating specificity in the outcomes related to ADHD. The available evidence suggests that when considering the link between ADHD and conduct problems, ADHD comes first and conduct problems develop later. In contrast there is no evidence that conduct problems in the absence of ADHD lead to the later development of ADHD. The small number of suitable longitudinal outcome studies highlights an important area for future research.
The aetiology of ADHD remains another area of controversy. In the view of the GDG this largely stems from the complex nature of ADHD and the many factors involved in aetiology. Major identified risk factors associated with the disorder include having a first-degree relative with ADHD and prenatal maternal stress. These are likely to be proxy markers of processes that are themselves expected to be highly complex, however. At the level of specific factors such as individual genes or direct environmental stresses, the increased risk of ADHD is expected to be small. There is an ongoing debate about the degree to which ADHD represents a homogeneous disorder, with multiple risk factors of small effect contributing to the disorder, or whether ADHD represents the syndromic end-point of multiple different processes. Further research is required to provide a full understanding of the complex aetiology involved.
One important question raised by the consensus conference was the interpretation of family, twin and adoption studies and the relative contributions between genetic and environmental influences indicated by these studies. The argument against important genetic influences is not strong unless one questions the conventional interpretation of twin and adoption data. The findings from twin studies are not, however, controversial because they have been replicated many times. The main finding is that parent and teacher reports of ADHD symptoms show high correlations of around 70 to 80% in monozygotic (identical) twins, and around 20 to 40% in dizygotic (non-identical) twins (Thapar et al., 1999). The usual interpretation of these findings is that the large difference in monozygotic and dizygotic correlations results from genetic influences. The alternative argument that the equal environment assumption is incorrect would not alter the basic conclusion that ADHD tends to run in families and is therefore a familial disorder, since the level of ADHD symptoms in one child is highly predictive of the level of ADHD symptoms in their siblings. It is therefore non-controversial that ADHD is familial and this in itself is strong evidence that the construct is sufficiently delineated to show clear familial effects.
Interestingly there are limited data from twin studies using ADHD cases (for example, concordance rates for the clinical disorder), so the literature mainly uses extremes analysis of rating scale data for ADHD symptoms and does not take into account other important aspects of the clinical disorder such as pervasiveness and impairment. Similarly there is a lack of twin data in adult populations.
Adoption studies also indicate that genetic as well as environmental influences increase the risk for ADHD. All adoption studies show that adopted children with ADHD are more similar to their biological parents than to their adoptive parents. These studies, except for one (Sprich et al., 2000) are, however, limited by small sample size and in most cases the interviewers were not blind to psychiatric or adoptive status; the studies have therefore not been used as evidence of validity in this chapter.
There was broad agreement that environmental influences play an important role in the aetiology of ADHD. However, the nature of the specific risk factors and the mechanisms involved are poorly understood and remain an area of controversy. Twin studies indicate that unique environmental effects are expected to cause differences between siblings and would explain in part why one child in a family has ADHD while another child from the same family does not. Environmental risks may be the sole or main cause of ADHD in some cases, for example, where there is extreme deprivation in early childhood (Rutter & O’Connor, 2004; Rutter et al., 2007). One important question is whether the evidence of genetic influences in ADHD can be reconciled with the view that environmental influences play a critical role in development of the disorder. In fact, high heritability is consistent with the existence of environmental risks for ADHD that are very common, and for this reason explain little of the observed variance in ADHD symptoms in the population. Environmental risks may also be modified by genetic risks (gene-environment interactions) or correlated with genetic risks (gene-environment correlation). The complexity of the interplay between genes and environment in the risk for ADHD is not well understood and for this reason is one of the main focuses for contemporary research. The GDG considered that polarised positions in this debate are not helpful since the contemporary understanding of complex behavioural disorders emphasises the interplay between nature and nurture.
The GDG wishes to stress that the role of genetic influences in ADHD does not exclude an important role for environmental influences for several reasons. Individual differences in genetic risk factors are likely to alter the sensitivity of an individual to environmental risks. Either genetic or environmental risks alone may play a prominent role in individual cases. Reducing environmental risks would be expected to reduce the risk for ADHD under most models of gene-environment interplay in the contemporary literature.
The GDG also wishes to emphasise that the extent to which the disorder results from genetic influences has no direct bearing on the choice of treatment and in particular, does not provide sufficient justification alone for the use of pharmacological interventions. For example, traits such as obesity or diabetes are influenced by both genetic and environmental factors, yet individual changes in lifestyle as well as the use of medication in some (but not all) cases is indicated. In ADHD, educational, social, psychological, and pharmacological treatments all need to be considered and could be important in improving levels of impairment and preventing the development of negative long-term outcomes. The evidence base for treatment of ADHD is dealt with in other sections of this guideline.
5.14. SUMMARY FROM REVIEW OF THE DIAGNOSIS
On the basis of this review, the GDG summarised the evidence for the diagnosis of ADHD upon which the guideline recommendations are made:
ADHD is a valid clinical condition that can be distinguished from coexisting conditions and the normal spectrum.
ADHD is distinguished from the normal spectrum by the co-occurrence of high levels of ADHD symptoms when they are associated with significant clinical, psychosocial and educational impairments. These impairments should be enduring and occur across multiple settings.
There is no specific biological test for ADHD, so the diagnosis must be made on the basis of a full developmental and psychiatric history, observer reports and examination of the mental state.
In the absence of a biological test for the diagnosis of ADHD or hyperkinetic disorder, validity is based on the association of ADHD symptoms with genetic, environmental, neurobiological and demographic factors; and the association of high levels of ADHD symptoms with impairments in multiple domains.
Hyperkinetic disorder (ICD-10) identifies a sub-group of people with ADHD with severe impairment in multiple domains.
ADHD commonly persists throughout childhood and into adult life, either as the full diagnostic criteria or in partial remission, where it continues to cause significant clinical and psychosocial morbidity.
5.15. IMPLICATIONS FOR PRACTICE
5.15.1. General principles for the diagnostic process
The aim of this section of the guideline is to provide a commentary and further recommendations on the implementation of the diagnostic process. As reviewed above there is sufficient evidence that ADHD is a valid diagnostic category to apply to relevant children, young people and adults. The GDG concluded that on the basis of current evidence ADHD is a complex disorder resulting from multiple genetic and environmental risk factors, representing the extreme and impaired tail of a normally distributed trait in the population. The disorder is recognised by the presence of a high level of pervasive and enduring problems with attention, overactivity and impulsiveness when they lead to a significant degree of clinical, psychosocial and/or academic impairments.
The current operational criteria for ADHD (DSM-IV-TR) and hyperkinetic disorder (ICD-10) are highly reliable when they are applied by trained individuals following the careful evaluation of reported behaviours and symptoms, and when the criteria define a group with clear clinical implications. The diagnosis depends on the evaluation of two necessary components, both of which are required to trigger the use of this guideline. The first is the presence of the symptom cluster of age-inappropriate levels of inattentive, hyperactive and impulsive behaviours; and the second is the presence of significant clinical and psychosocial impairments. Other key criteria include onset during childhood and situational pervasiveness. Behaviours and symptoms that are restricted narrowly to one environmental setting only (for example, school), or one set of impairments (for example, educational attainment alone) would not be considered sufficient grounds to make the diagnosis.
The implementation of the diagnostic and treatment process should be within the framework of a structured stepped pathway as described in Chapter 6. Within this framework a flexible approach to assessment should be adopted that enables an evaluation of individual and family needs, drawing on the experience and expertise of the individual clinician and other professionals involved, and taking into account different perspectives using an interdisciplinary approach.
5.15.2. Implementation of the diagnostic criteria
Diagnostic criteria are constantly evolving in the light of new information. The GDG reviewed the current diagnostic criteria and made recommendations that reflect the current state of knowledge and clinical practice. Below is a list of common questions with the summary statements upon which the recommendations are based.
A. Should ADHD be recognised in the presence of pervasive developmental disorders/autism spectrum disorders?
ICD-10 unequivocally says this is not permitted and DSM-IV-TR states that, ‘symptoms should not occur exclusively in the course of a pervasive developmental disorder’ (APA, 2000); yet pervasive developmental disorders once established are in most cases always present.
The evidence that core symptoms of ADHD occur together with those of pervasive developmental disorders/autism spectrum disorders is strong and therefore the GDG recommends that for effective practice ADHD should be recognised on the basis of core symptoms of ADHD, even when pervasive developmental disorders/autism spectrum disorders are present.
Summary statement: ADHD can be diagnosed in the presence of pervasive developmental disorders.
B. Should ADHD be recognised in the presence of general learning disability?
Both DSM-IV-TR and ICD-10 state that symptoms of ADHD must be developmentally inappropriate. This means that the levels of ADHD symptoms should be inappropriate and impairing in comparison with other people at the same developmental stage taking into account both age and general cognitive ability. DSM-IV-TR states that symptoms should be ‘excessive for mental age’. The GDG recognised the importance of an appropriate developmental comparison group and recommends that adjustment is made for mental age.
For example a mental age of 5 in a 10 year old should have the same standard of what is expected for impulsiveness and inattention as a mental age of 5 in a 5 year old. However, derivation of ‘mental age’ through standardised cognitive assessment does not always correlate with emotional and behavioural age. Professionals undertaking clinical evaluation should have expertise in both ADHD and learning disability, and awareness of the normal range of behaviour in the equivalent peer group of comparable age and general cognitive ability.
Summary statement: ADHD can be recognised in the presence of a general learning disability, with behavioural symptoms compared to a group of similar mental age.
C. How should impairment be judged?
The GDG agreed that the presence of impairment associated with the core behavioural symptoms of ADHD is critical to recognising the disorder; but difficulties arise since impairment is itself a continuum.
Moderate impairment is a requirement for the diagnosis of ADHD. Moderate ADHD in children and young people is taken to be present when the symptoms of hyperactivity/impulsivity and/or inattention, or all three, occur together, and are associated with at least moderate impairment, which should be present in multiple settings (for example, home and school or a healthcare setting) and in multiple domains where the level appropriate to the child’s chronological and mental age has not been reached: self-care (in eating, hygiene, and so on); travelling independently; making and keeping friends; achieving in school; forming positive relationships with other family members; developing a positive self-image; avoiding criminal activity; avoiding substance misuse; maintaining emotional states free of excessive anxiety and unhappiness; and understanding and avoiding common hazards. The level of impairment could also be estimated by using a predetermined level on a global adjustment scale (for example, a score of less than 60 on the C-GAS). In later adolescence and adult life, the range of possible impairments extends to occupational underachievement, dangerous driving, difficulties in carrying out daily activities such as shopping and organising household tasks, in making and keeping friends, in intimate relationships (for example, excessive disagreement) and with child care.
Severe ADHD corresponds approximately to the ICD-10 diagnosis of hyperkinetic disorder and the GDG took this to be present when hyperactivity, impulsivity and inattention are all present in multiple settings and when impairment is severe (that is, it affects multiple domains in multiple settings).
The GDG considered that impairment needs to be considered relative to a comparable peer group since this represents the potential of each individual. For example, relative academic impairment would include a child with a chronological age of 7, a mental age of 10, but an academic achievement age only of 7. Importantly, impairment should be pervasive and enduring, affecting several aspects of an individual life. This would mean that impaired academic achievement alone would not be sufficient to trigger the diagnosis, but would be sufficient where this were accompanied by significant impairments in other areas such as emotional or social development (see Section 5.6).
Summary statement: Impairment should be pervasive and enduring, affecting several aspects of an individual life.
D. Should the age of onset before 7 years be strictly applied?
The GDG recognised the inadequacy of the current age of onset criteria, which would exclude individuals with typical ADHD with an apparent onset after the age of 6 years. Symptoms may not be recognised in young children and impairments may not be pronounced. This is likely to be particularly true where the predominant symptoms are those of inattention rather than impulsive or overactive behaviour and because it can be the later development of coexisting problems that draws attention to the difficulties that a particular child is having. Recent evidence indicates that the level of impairments are similar for individuals with onset before and after age 7 years leading the GDG to consider that ADHD should be diagnosed in some cases where onset is dated between the ages of 7 and 12 years (Applegate et al., 1997).
Summary statement: ADHD should be diagnosed in some cases where onset is dated between the ages of 7 and 12 years.
E. Should some kinds of aetiology be excluded?
The GDG recognised that ADHD is a complex heterogeneous disorder with a range of different aetiologies, including environmental, genetic and non-genetic neurobiological factors. The DSM urges the distinction of ADHD from ‘children from inadequate, disorganised or chaotic environments’ (APA, 2000).
The GDG considered that there is not yet sufficient data to include or exclude individual cases on the basis of aetiology. For example exposure to chaotic environments might be one potential cause of ADHD, and prenatal exposure to alcohol another. The GDG therefore recommends that the diagnosis of ADHD should be distinguished from other behavioural disorders on the basis of the pattern and type of behaviours, rather than on the basis of specific aetiologies. This is an important point since the diagnosis might be excluded in the presence of a severe environmental risk such as child abuse. The view that child abuse is the cause of behavioural problems, while likely to be important in an individual case, should not lead to the exclusion of the individual from these guidelines if they fulfil the diagnostic criteria for ADHD.
Summary statement: In the current state of knowledge, ADHD should be considered whenever diagnostic criteria are fulfilled, regardless of the presence of any specific aetiological factors.
F. Should the same definitions be used for both genders?
Epidemiological studies typically apply the same definitions to boys and girls, and typically find a male preponderance – most commonly about 3 to 1 (Schachar & Tannock, 2002). The gender ratio for children attending ADHD clinics is typically higher than in community surveys, raising the possibility of under-recognition in females. The outcome in adolescence seems to be no better for girls than has been reported for boys (Young et al., 2005a & b).
In adult life, the male-female ratio for ADHD appears to be approximately equal (Kooij et al., 2005), again raising the possibility that the high gender ratios in childhood may be partly a result of under-identifying the problem in girls, or of a different presentation of symptoms in girls.
The evidence does not allow for a clear scientific consensus, so the practice is still to apply diagnostic criteria regardless of gender. Research is needed, however, to clarify the nature and prognostic implications of different presentations in boys and girls.
Summary statement: In current knowledge, the same diagnostic criteria should be applied to males and females.
G. Can the diagnosis be made from rating scales only?
Despite reasonably high sensitivity and specificity from rating scales, the GDG took the view that diagnosis of ADHD should not rely on rating scale measures alone. Rather, it is important to complete a full evaluation including diagnostic clinical interviews with parents, children (especially older children and adolescents) and other corroborative evidence such as school reports. The use of rating scale data alone will generate both false positive and negative diagnoses and would remove the critical element of an in-depth appraisal of the entire clinical picture including onset, cause, associated developmental and mental health exacerbating and causal factors.
Summary statement: The diagnosis of ADHD should only be made after a full clinical and psychosocial evaluation, and never on the basis of rating scale data alone.
H. Can the diagnosis be made on the basis of observation alone?
Direct observation of an individual with ADHD, particularly older adolescents and adults, for short periods of time during assessment sessions may not demonstrate any obvious features of the condition. This should not exclude the diagnosis where there is a clear account of inattentive, impulsive or hyperactive behaviours in usual situations. The reason is that some people with ADHD can regulate their behaviour for short periods of time and because ADHD behaviours are typically reduced in situations where a person is engaged in an important task. The GDG advises that diagnosis should only be made on the basis of a full assessment.
Summary statement: The diagnosis of ADHD should not be made on the basis of observational data alone.
I. How should social, cultural and economic circumstances and factors be taken into account in making the diagnosis of ADHD?
At a general level, diagnoses of ADHD are distributed unequally by relative level of deprivation, mediated by social class and ethnicity (Bauermeister et al., 2005; Cunningham & Boyle, 2002; Dahl et al., 1991; Timimi, 2006). While these factors are not thought to cause the behavioural symptoms of ADHD, such immediate environmental circumstances may have a role to play in mediating the experience of symptoms and impairment (Isaacs, 2006). Relative deprivation increases the likelihood that a child will be subject to various environmental risk factors, potentially increasing the risk of ADHD and associated disorders (Hartl et al., 2005; Lahti et al., 2006; Neuman et al., 2007; Rodriguez & Bohlin, 2005). Additionally the ethics and beliefs of those responsible for the daily care of children have a role to play in their perception of symptoms and impairment (Couture et al., 2003; Curtis et al., 2006; Epstein et al., 2005; Rey et al., 2000; Singh, 2003; Wolraich et al., 2003). This being so, some attempt should be made to investigate and if possible either discount or take account of the immediate environmental circumstances of the child.
If existing evaluations of the social, cultural and economic circumstances have already been made through multi-agency collaboration then this information may be readily available at the time of referral (Burgess, 2002; San Roman, 2007). However, if these investigations have not been carried out by the relevant services (for example, social services, health visiting services or school health services), or if for some reason this information has not been made available, then they should be made part of the medical assessment.
There is a growing literature on the measures that can be taken to help the child with ADHD in the school and at home and as a minimum it should be ensured that such measures have been taken (Hughes & Cooper, 2006; Lloyd et al., 2006; Merrell & Tymms, 2002; Prosser, 2006). Regardless of socio-cultural circumstances, psychiatric diagnosis and treatment will have a significant impact on these circumstances, and this needs to be acknowledged by the individual and family concerned (Singh, 2004, 2005). The active participation of the child or young person should be sought at all stages of the diagnostic process (Wright et al., 2006).
Summary statement: Social, cultural and economic circumstances should always be evaluated by an expert and whenever possible by a multidisciplinary team.
5.16. DIFFERENTIATING ADHD IN ADULTS FROM OTHER COEXISTING CONDITIONS
5.16.1. Personality disorders
There is currently considerable nosological confusion that stems from the early onset and persistence of ADHD behavioural symptoms that therefore appear as stable traits or personality characteristics rather than symptoms. The difference in definition between a trait and a symptom is that symptoms represent a change from a normal pre-morbid state, such as the onset of adult depression or psychosis, whereas traits are considered to be enduring characteristics. Current psychiatric training in adult mental health tends to focus on the distinction between symptoms and traits and gives rise to a nosology that does not fit well with the concept of ADHD. First, because of the trait-like quality of ADHD phenomena, significant psychopathology often goes unnoticed or is regarded as a personality characteristic, resulting in a different set of treatments and expectations for the clinical course and outcome compared with ADHD. Second, because ADHD phenomena are sometimes associated with persistent disruptive and oppositional behaviour or development of poor interpersonal skills, it is often assumed that this represents an ingrained and therapeutically resistant set of behavioural traits. Further confusion stems from the definition of cluster B personality disorders, like antisocial, borderline and emotionally unstable personality disorder, which include symptoms such as mood instability, impulsivity and anger outbursts that are commonly seen to coexist in adults with ADHD.
The diagnostic issue is to recognise when there is evidence for ADHD, that is whether the operational criteria were fulfilled in childhood and whether ADHD symptoms that started in childhood have persisted and continue to bring about significant impairments. While the diagnostic focus should be on the main symptoms that define inattention, hyperactivity and impulsivity it is also important to remember that mood instability and impulsivity are commonly seen in adults with ADHD. Care must be taken to distinguish between uncontrolled, impulsive, oppositional and antisocial behaviours that arise in the context of a specific ADHD syndrome from those that do not. For this reason it is often useful to make particular enquiries about symptoms that are more specific to ADHD such as short attention span, variable performance, distractibility, forgetfulness, disorganisation, physical restlessness and over-talkativeness rather than focus only on the occurrence of maladjusted and disruptive behaviours.
5.16.2. Mood disorders
Depression
A volatile and irritable mood is frequently seen in adult ADHD and is not usually the consequence of coexisting depression or bipolar disorder. The overlap of mood symptoms does mean that care must be taken to exclude the possibility of a major affective disorder and that mood lability does not occur solely within the context of such disorders. Attending to the time-course of the symptoms and psychopathology can help to distinguish the two. Early onset, chronic trait-like course, frequent mood swings throughout the day, no recent deterioration or severe exacerbation frequently accompany ADHD, whereas extreme low or high moods, sustained mood change for long periods of time and recent onset are more indicative of a primary affective disorder. Some individuals previously diagnosed with atypical depression, cyclothymia or unstable emotional personality disorder will have a primary diagnosis of ADHD.
Bipolar disorder
Traditionally, the distinction between ADHD and bipolar disorder has been fairly easy to make. Bipolar disorder has been associated with euphoria, grandiosity and a cycling course, with each episode lasting for several days at least. ADHD, by contrast, has been regarded as a persisting disability in which euphoria is not particularly a feature. The goal-directed over-activity of mania is usually seen to be in contrast with the disorganised and off-task activity of ADHD. Individuals with ADHD often have difficulty sleeping but unlike mania or hypomania they complain about their lack of sleep and often feel exhausted during the day. In general individuals with ADHD report that they cannot function effectively and this is often associated with chronic low self-esteem, very different from the feelings of heightened efficiency seen in mania. In ADHD thoughts are often described as ‘on the go’ all the time, but unlike mania or hypomania, these are experienced as unfocused, muddled and inefficient and there is no subjective sense of improved efficiency of thought processes.
There has, however, been a broadening of the concept of bipolar disorder, to include cases where the mood change is not euphoria but irritability or chronic mixed affective states, and where the cyclical nature consists of many changes within a single day (indistinguishable from a volatile, labile mood). This leads to a very considerable similarity in formal definitions between this so-called ultradian version of bipolar disorder and ADHD. An unstable and over-reactive mood is very commonly seen in ADHD, even though it is not part of the diagnostic definitions, and the development of an oppositional disorder, in which frequent tantrums are common, can be described as an ‘irritable’ state and therefore contributes to a bipolar diagnosis.
One of the main questions relates to the validity of a diagnostic concept broadly-defined as bipolar disorder, or whether mood instability/irritability in the presence of ADHD may be more adequately described by a new dimension, such as mood dysregulation. Until the relevant empirical data become available, the classic definition of mania should be maintained: a diagnosis of bipolar disorder requires euphoria, grandiosity and episodicity, and the differential between ADHD and bipolar disorder remains explicit.
5.16.3. Anxiety disorders
Individuals with ADHD commonly report high levels of anxiety on rating scales. However, a more detailed enquiry about the psychopathology shows that in some cases the ADHD syndrome mimics some aspects of anxiety. Individuals with ADHD may have difficulty coping with social situations because they are unable to focus on conversations; difficulty travelling because they are unable to organise the journey; and difficulty shopping because they may become irritable waiting in queues and because they may forget things and be highly disorganised. Problems with simple everyday tasks that most people take for granted are a source of considerable concern and are often accompanied by avoidance of stressful tasks and poor self-esteem. In combination with ceaseless mental activity, these legitimate concerns and responses may take on the appearance of a mild to moderate anxiety state, although lacking the systemic manifestations of anxiety disorders. An important distinction is to consider whether the symptoms have a similar onset and time course to ADHD or whether they arise episodically and in response to stressors, which is characteristic of anxiety.
5.16.4. Psychotic disorders
Severe inattention may rarely mimic the thought disorder symptoms seen in some psychoses, such as derailment, tangential thought processes, circumstantiality and flight of ideas. Careful monitoring of both psychotic symptoms and ADHD symptoms is advised but it may be difficult to distinguish residual symptoms of a major mental illness from persistence of ADHD symptoms.
5.17. RECOMMENDATIONS
5.17.1. Diagnosis
- 5.17.1.1.
A diagnosis of ADHD should only be made by a specialist psychiatrist, paediatrician or other appropriately qualified healthcare professional with training and expertise in the diagnosis of ADHD, on the basis of:
a full clinical and psychosocial assessment of the person; this should include discussion about behaviour and symptoms in the different domains and settings of the person’s everyday life, and
a full developmental and psychiatric history, and
observer reports and assessment of the person’s mental state.
- 5.17.1.2.
A diagnosis of ADHD should not be made solely on the basis of rating scale or observational data. However rating scales such as the Conners’ rating scales and the Strengths and Difficulties questionnaire are valuable adjuncts, and observations (for example, at school) are useful when there is doubt about symptoms.
- 5.17.1.3.
For a diagnosis of ADHD, symptoms of hyperactivity/impulsivity and/or inattention should:
meet the diagnostic criteria in DSM-IV or ICD-10 (hyperkinetic disorder),
9
andbe associated with at least moderate psychological, social and/or educational or occupational impairment based on interview and/or direct observation in multiple settings, and
be pervasive, occurring in two or more important settings including social, familial, educational and/or occupational settings.
As part of the diagnostic process, include an assessment of the person’s needs, coexisting conditions, social, familial and educational or occupational circumstances and physical health. For children and young people, there should also be an assessment of their parents’ or carers’ mental health. (Key priority)
- 5.17.1.4.
ADHD should be considered in all age groups, with symptom criteria adjusted for age-appropriate changes in behaviour.
- 5.17.1.5.
In determining the clinical significance of impairment resulting from the symptoms of ADHD in children and young people, their views should be taken into account wherever possible.
5.17.2. Post-diagnostic advice for parents
- 5.17.2.1.
Following a diagnosis of ADHD, healthcare professionals should consider providing all parents or carers of all children and young people with ADHD self-instruction manuals, and other materials such as videos, based on positive parenting and behavioural techniques.
5.18. RESEARCH RECOMMENDATIONS
- 5.18.1.1.
Grounds for diagnosis of ADHD in adults
What is the prevalence of inattention, impulsivity, and hyperactivity/restlessness in males and females in the adult population? How far do the core symptoms of inattention, impulsivity and hyperactivity/restlessness cluster together? To what extent are the core symptoms comorbid with other forms of mental disturbance? To what extent are the core symptoms associated with neuropsychological and social impairment? This would be best conducted as an epidemiological survey.
Why this is important: There is evidence that ADHD symptoms can persist into adulthood and cause impairment, but there are no clear conclusions about the level of ADHD symptoms in adults that should be considered as grounds for intervention, or whether the symptoms take a different form in adulthood. The costs to society and to the affected people and their families make it pressing to know whether, and how far, services should be expanded to meet the needs of this group.
- 5.18.1.2.
Influences determining the impact of symptoms on impairment and on the risk of later disorder
For people of all ages and both genders with ADHD, what are the influences determining the impact of symptoms on their functioning (‘impairment’) and on the risk of later disorder? Symptomatology and its impact should be based on reliable assessments from several sources, and the outcomes should be specific to the effect caused in major social and developmental domains. The possible influences to be measured as moderators of the relationships between symptoms and impairment should include: gender and developmental level (in case different symptom criteria should be applied for different groups), the timing of any recognition and intervention (to estimate benefits and risks of early diagnosis and treatment) and potentially modifiable environmental circumstances (such as family atmosphere, peer group, and socioeconomic adversity). Additional research should examine the same relationships in short-term longitudinal designs to include a predictive element.
Why this is important: The research is needed in view of currently varying practice in the application of diagnostic criteria and unsatisfactory knowledge about the levels of symptoms and impairment that should indicate whether treatment is required. Such research is also needed to guide practitioners on the clinical features to target as part of comprehensive management.
- 5.18.1.3.
The extent to which neuropsychological tests can be used to guide psychological interventions
For children and young people with ADHD, what is the extent to which neuropsychological tests can effectively be used to guide psychological interventions? Standardised tests should be developed, normed and applied to functions such as response inhibition, delay-of-reward gradients and aversion to delay. Educational recommendations based on individual profiles of these and established executive function tests should be compared with standard advice for their acceptability to teachers, their implementation in practice, and the effects on child behaviour and learning in the classroom.
Why this is important: Scientific investigation has established robust associations between the behaviours of ADHD and deviations in performance on neuropsychological tests. These results however remain in the research arena only, partly because of a shortage of norms for the tests (required for diagnosing individuals) and partly because of uncertainty about the benefits to be obtained from prescriptions for remedial intervention based upon them.
- 5.18.1.4.
The prevalence of ADHD in young people and adults in substance misuse and/or forensic populations; and how individuals in these specific populations might best be treated
It has been claimed that there are much higher rates in these populations compared with the normal population, but this is not based on good evidence because many of the studies are methodologically flawed, for example by being based on rating scale screens only, and not controlling for a history of conduct disorder. Surveys should be mounted, using not only rating scales, but also clinical identification with interviews and source informants. There should also be an assessment of the efficacy, in these groups, of the ADHD treatments already recommended for ADHD in the community. Randomised controlled trial design is recommended with outcome measures including not only those of ADHD itself but also those relevant to the target populations (for example, offending and substance misuse).
Why this is important: It is important that individuals with ADHD are identified and receive treatment in these settings as this may have a positive impact on their quality of life, increase the effectiveness of other forensic rehabilitation activities and treatments provided to them, contribute to a reduction in antisocial behaviour and offending and increase public safety. Treatment of ADHD symptoms may improve treatment engagement and readiness more generally and provide service benefits by shortening length of stay within forensic secure services.
- 9
The ICD-10 exclusion on the basis of a pervasive developmental disorder being present, or the time of onset being uncertain, is not recommended.
