Application of clinical criteria
Assessment starts with self-reported symptoms. The physician should perform an in depth diagnostic interview to look for the characteristic psychopathology by careful questioning about childhood and current behavioural symptoms. Although the patient appears to be the best informant, comparison with parent and partner reports in order to provide more information on severity and pervasiveness of symptoms, is desirable . Both DSM-IV and ICD-10 criteria recognise that symptoms of ADHD and HKD persist beyond childhood into adulthood, yet neither set of criteria takes account of age-dependent changes in terms of the number and severity of symptoms, or changes in the way that the symptoms of ADHD might present in adults. Rather, they stipulate the exact same criteria as that applied in children. However developmental changes occur that include an increasing role for inattention and other 'executive function' difficulties in the impairments related to the demands of adult life . The DSM-IV criteria suggest that adults who demonstrate only some of the symptoms of ADHD should be given a diagnosis of "ADHD, in partial remission". This diagnosis, however, seems to underplay the significance of impairments seen in adults who no longer meeting the full DSM-IV criteria but show persistence of some impairing symptoms from childhood. In other words adults seem to outgrow the criteria rather than the disorder.
We therefore conclude that a definition of remission based on the number of criteria that have to be met in childhood does not seem appropriate in clinical practice with adults, especially as symptoms described by both DSM-IV and ICD-10 criteria typically apply to children and are not adjusted for developmental age [13, 35, 172, 173]. Furthermore, restriction of the diagnosis to ICD-10 criteria that focused on a severe form of the combined subtype during childhood will lead to underdiagnosis of adults impaired by ADHD symptoms, especially in women. Symptoms of ADHD in adults should therefore be judged with reference to developmentally appropriate norms. Preliminary research suggests that using the current DSM-IV criteria that stipulate six of nine symptoms of either inattention or hyperactivity-impulsivity, a lower threshold of four of nine of these symptoms in either domain is sufficient to identify impairing levels of ADHD symptoms in adults [17, 174]. We therefore recommend that future criteria for ADHD are appropriately adjusted by taking into account age-related changes to symptoms and their relationship to impairment.
Further work is now required to clarify whether broadening the criteria to include those with four or more symptoms in either domain is sufficiently specific within adult mental health populations, since by lowering the threshold other acute psychiatric disorders may generate ADHD-like symptoms that reach this lower threshold. This is however mainly a problem for cross-sectional screening for ADHD, since in clinical practice the diagnosis is based on eliciting a history of ADHD symptoms that start in childhood or early adolescence and are persistent (trait-like) and impairing over time and can therefore be differentiated from adult onset disorders. The other approach under investigation is the identification of alternative descriptions of behavioural impairments seen in adults with ADHD, such as the ecological executive function deficits described by several authors in recent years [175, 176], which may show greater sensitivity and specificity for the diagnosis in adults.
As discussed, the expression of ADHD in adults is to some extent different from that in children and the diagnostic descriptions of some of the features need to be adapted to adult expressions of the disorder. For example, physical overactivity in children could be replaced in adulthood by constant mental activity, feelings of restlessness and difficulty engaging in sedentary activities. Furthermore a range of characteristics closely associated with the core ADHD syndrome often lead to some of the impairments typically seen in adults with ADHD. These include symptoms such as disorganization, short-fuse, temper, mood lability and sensitivity to stress [35, 116]. Assessment of symptoms is further complicated by the fact that adults have more ways to adapt and/or compensate for problems with attention, hyperactivity and impulsivity. For example adults with high IQ, high socio-economic status or lower levels of comorbid disorders may have better compensatory strategies.
Another difference in the evaluation of ADHD in adults is the usual reliance on self-report of symptoms rather than informant accounts of behaviour by parents and teachers. Although the validity of a retrospective diagnosis of ADHD in adults may be questioned in some cases, this is also true for other psychiatric diagnoses that are dependent on descriptions of current and past behaviour. Self-reports of past and current symptoms can however be reliable if the patient has good insight into the condition . Nevertheless, caution is needed as retrospective recall of childhood symptoms may be compromised in adults with ADHD, with under reporting found to be a common problem due to difficulties with accurate recall [9, 178]. While a diagnosis based only on self-report is possible, such an approach may lead to underdiagnosis of ADHD and it may be more reliable to use information from informants as well . Reliance on self-report alone may also risk over diagnosis in some cases, but until now, no evidence has emerged that this is the case . For these reasons it is recommended that whenever possible corroborating information is obtained from a living parent of older relative for childhood behaviour and a partner, relative or close friend for current behaviour and symptoms.
Another important criterion to evaluate is the age of onset. The DSM-IV criterion that some symptoms and impairment should be evident before the age of 7 years is difficult to assess accurately in a retrospective diagnostic assessment in adults. In fact, research has failed to validate to this criterion since it has been found that the clinical syndrome that defines ADHD has similar clinical predictions in terms of course, treatment response and associated impairments regardless of whether symptoms with impairment started before the age of 7 years or later [181–187]. This may be particularly important in relation to the inattentive type which is characterised by a later onset of impairment [172, 174, 188], as the inattentive symptoms may have been missed in the pre-school years and impairment may not have been noticed or reported until the age of secondary school.
A further problem is recall bias by parents, since after decades they may not accurately recall when symptoms or impairments started. Recent evidence found that on average parents typically report a later age of onset by around 5-years, even where earlier age of onset was known from previous health care records [170, 189, 190]. One approach to circumvent these problems is to refer to school reports which may provide more accurate and relevant information with respect to the age of onset.
For these reasons it is proposed to use a broader onset criterion, up to early to mid adolescence [186, 191] and this approach has been recommended in the recent National Institute of Health and Clinical Excellence in the UK . Clinical judgement should be used in making the diagnosis if symptoms before the age of 7-years cannot be recalled. In most cases a clear report of at least some symptoms associated with impairment is expected by early to mid-adolescence. However, even if there is no clear recall of childhood symptoms, ADHD should be considered when the typical syndrome that defines ADHD is present and there is evidence of lifetime persistence of symptoms indicative of ADHD and associated impairments. This may for instance, be the case in some inattentive but talented people in whom high general cognitive ability and a structured environment may have helped them to cope during childhood and adolescence. The problem of inattention may remain unrecognised until they tried to live independently from their parents and were faced with the organisational and attentional demands of higher education or employment .
In addition to the evaluation of symptoms, another important criterion for the clinical diagnosis of ADHD is the presence of significant levels of impairment associated with the symptoms. This is critical because the symptoms of ADHD are found to be continuously distributed throughout the population and there is no natural boundary between affected and unaffected individuals . As with symptoms of anxiety and depression, ADHD symptoms are experienced by most people at times. The disorder is therefore distinguished from the normal range by the severity (extreme nature) and persistence of symptoms, and their association with significant levels of impairment and risk for the development of co-occurring disorders. Criteria relate in part to cultural expectations and this may explain why diagnostic and treatment rates of ADHD are higher in the US than many European countries. From a mental health perspective it is however important to define the impairments from ADHD at a level that most people would consider needs some form of medical, psychological or educational intervention and represents a mental health problem . Furthermore, the diagnosis should not be applied to justify the use of stimulant medication to enhance performance in the absence of a wider range of significant impairments indicating a mental health disorder. Impairments include problems with the following: self-esteem, personal distress from the symptoms, social interactions and relationships, behavioural problems, and the development of comorbid psychiatric syndromes.
Current evidence clearly defines ADHD as a clinical syndrome associated with impairments in multiple domains including academic difficulties, impaired family relationships, social difficulties and increased rates of conduct problems. In adults with ADHD increased rates of antisocial, drug use, mood and anxiety disorders are reported in both cross-sectional and longitudinal follow-up studies; in addition to increased rates of unemployment, poor work performance, lower educational performance, increased rates of traffic violations and accidents and criminal convictions: reviewed in NICE, .
Finally, the high familial risk among first degree relatives, in the order of 20% or more [20, 194], means that a strong predictor for the ADHD in adults is having a parent, sibling or child with ADHD. Family history of ADHD should therefore raise the index of suspicion and provides further supportive evidence when evaluating individuals for ADHD.
The assessment process
Diagnosis is based on a careful and systematic assessment of a lifetime history of symptoms and impairment. It is not just based on a single clinical impression gained during consultation. Central to this process is the assessment of childhood-onset, current symptoms of ADHD and the presence of symptoms and impairment in at least two domains (school, work, home, interpersonal contacts). Associated features should be evaluated including mood lability, temper outbursts and comorbid disorders (see differential diagnosis). It is important to take a full medical history of psychiatric and somatic treatments, as well as a family history of psychiatric and neurological problems. It is useful to ask the patient about the pattern of symptoms, typical of ADHD and its co-morbidities in his/her family, taking into account familial factors and the high heritability of the symptoms. Patterns of comorbidity in children and adults with ADHD have been identified and include: mood, anxiety, sleep, conduct and substance use disorders as well as personality disorders. Some care must be taken to distinguish between symptoms that often co-occur with the core syndrome of ADHD (e.g. mood instability, ceaseless mental activity, avoiding situations such as waiting in lines when frustration may occur) from those of a separate comorbid condition (e.g. bipolar disorder, major depression, anxiety, personality disorder). Comorbidity being the rule rather than the exception, the evaluation of co-occurring symptoms, syndromes and disorders must always be part of the clinical assessment of adult ADHD [148, 150].
ADHD is also associated with increased rates of neurodevelopmental traits and disorders including autism spectrum disorder , dyslexia  and impaired motor coordination ; which are thought to arise from overlapping genetic influences. Such neurodevelopmental comorbidities are less well studied in adults with ADHD, but they are commonly observed in clinical practice and may lead to continued impairments following successful treatment of ADHD symptoms with medication. It has also been the practice in some European countries to look for soft neurological signs during the assessment of adult ADHD, in analogy with the diagnosis of DAMP in childhood (Deficits in Attention, Motor control and Perception), that often accompanies ADHD [198, 199]. However there is limited data on the DAMP syndrome in adults.
Comorbid substance use disorder (SUD) deserves special attention due to the high rates of ADHD within SUD populations. A bidirectional link between ADHD and SUD is reported [142, 200] with ADHD symptoms over represented in SUD populations  and SUD in ADHD populations . From twenty five published studies that screened for substance misuse in ADHD samples prevalence was estimated to be around 45% to 55%. Alcohol and cannabis are the most frequently abused substances in these populations  followed by lower rates of cocaine and amphetamine abuse . In contrast, from ten studies that screened for undiagnosed ADHD in SUD populations estimates of ADHD ranged from 11%  to 54% . The causes for such comorbidity are likely to be complex including altered reward processing in ADHD, increased exposure to psychosocial risk factors and self treatment. Sometimes patients with ADHD describe self-treatment with unprescribed stimulants, such as parents trying their child's medication. In some other cases, paradoxical reactions to drugs of abuse are reported by patients who feel calmer, better able to concentrate and are less impulsive. While consuming drugs with a stimulant action such as cocaine or amphetamine is described in a few cases, it is more common for adults to describe a general reduction in symptoms from alcohol and cannabis [206, 207].
Although there are many challenges in identifying undiagnosed ADHD in SUD settings, systematic screening is feasible . This is particularly important since SUD patients with comorbid ADHD often present with severe forms of SUD  characterised by early onset, extended duration of SUD, greater impairment and a shorter transition from substance use to dependence . ADHD has been found to increase suicide risk in SUD adolescents . In SUD treatment outcomes, methadone maintenance patients in the USA with significant ADHD symptoms in the two weeks prior to admission were less likely to achieve abstinence .
Close supervision is advised when treating ADHD patients with SUD with stimulants and in cases where diversion or abuse is a particular concern, atomoxetine may be selected as the first line treatment. Research in ADHD populations suggests that only a small minority divert or misuse their medication  and certain long acting formulations with a low abuse potential can be used, although studies of adolescents suggest that 75% of patients who do misuse medication have comorbid SUDs .
Instruments for screening and diagnosis
There are many screening instruments and diagnostic interviews available, some of which have been translated into different languages. Commonly used rating scales for screening include the ADHD Rating Scale, based on the DSM-IV criteria , and the six item World Health Organisation Adult ADHD Self-Report Scale (ASRS) Symptom Checklist (online available without limitation and in many different languages at http://www.hcp.med.harvard.edu/ncs/asrs.php. The ASRS includes questions for each of the 18 DSM-IV items, re-worded to better represent the presentation of the ADHD items in adults. The 6-item short version was selected on the basis of sequential logistic regression, to optimize concordance with the clinical classification; and was estimated to have a sensitivity of 68.7% and specificity of 99.5% with total classification accuracy of 97.9%, evaluated using population survey data . Specificity of this and other screening tools may be lower within clinical samples with high rates of other mental health disorders; and positive screens should always be followed by full diagnostic evaluations based on clinical interview data. The ASRS, for example, has been investigated in a sample with substance use disorders and the sensitivity found to be higher (87.5%) and specificity lower (68.6%) than that reported in the previous study .
To guide the diagnostic assessment process, other self-report questionnaires are available, such as the Brown ADD Scale Diagnostic Form (BADDS)  that measures only behaviours relating to executive functioning and inattention; the Conners' Adult ADHD Rating Scale that includes the DSM-IV criteria and has different versions for patients and for significant others (CAARS) ; and the Wender Utah Rating Scale (WURS)  that includes also symptoms of other, often comorbid disorders . The total symptom scores of the ADHD Rating Scale and the CAARS may be used to screen for ADHD and to evaluate treatment outcome.
For the main diagnostic assessment use of a structured diagnostic interview is advised, such as the Conners Adult ADHD Diagnostic Interview for DSM-IV (CAADID) . An alternative under development is the Diagnostic Interview for ADHD in adults (DIVA) . Although several of the rating scales and interviews are available in different languages there is a need for validating these translations for use throughout Europe.
Currently there are no neurobiological or neuropsychological tests for ADHD with sufficient sensitivity and specificity to serve as an individual diagnostic test . Functional imaging seems promising although more research is needed to establish its value [224–226]. Neuropsychological tests (e.g. the CANTAB tests, the Stop Signal Reaction Time, IQ or computerized tests of executive functions and speeded reaction time responses) may complement diagnostic assessments and can provide an objective index of cognitive functions in the patient with ADHD . There is a recent surge in interest in the use of cognitive-electrophysiology that may provide data that is more sensitive to the diagnosis than cognitive-performance data alone [228, 229]. Patients with ADHD may suffer additional cognitive deficits, which may contribute to functional impairment, such as in learning, reading, writing difficulties, in addition to impairments related to the autistic traits. As there are no established norms for learning disorders in adults, it may be difficult to distinguish these disorders from ADHD-related learning deficits. Further research is required to provide a full understanding of the cognitive impairments associated with ADHD in adults.
Despite the poor predictive value of cognitive performance tests, some experts conceptualise ADHD as primarily a deficit of executive functions. While this may not always be seen in cognitive performance deficits, impairments are usually seen in the way that people with ADHD manage daily tasks [175, 176]. Many problems reported by adults with ADHD are thought to reflect executive dysfunctions, including impaired self-organisation, attentional and emotional regulation, sustained effort and alertness; that may only been seen in performance deficits in every day life and not under test conditions. This has led to efforts to provide more sensitive behavioural descriptions of performance deficits seen in ADHD that in general reflect problems of self-control and self-regulation . These new studies are likely to lead to improved and more sensitive symptom and behavioural checklists for ADHD in adults.
It is important for the diagnosis of ADHD, as well as the correct targeting of treatments, to identify comorbid conditions such as mood, anxiety, psychotic, organic and substance use disorders; in addition to personality, tic and autistic spectrum disorders. Because adults with ADHD often exhibit low self-esteem, low mood, affective lability and irritability, these symptoms may sometimes be confused with dysthymia, cyclothymia or bipolar disorder and with borderline personality disorder. Furthermore daily mood changes in ADHD are very common, and represent a poorly regulated but essentially normal range of moods, rather than the more severe extremes of depression and elation seen in bipolar disorder; and it is argued that chronic mood instability should be considered part of the core syndrome of ADHD [118, 175, 231].
ADHD and borderline personality disorder seem to share impulsivity, affective instability, anger outbursts and feelings of boredom [35, 117, 218]. In the ADHD patient, impulsivity and anger is usually short-lived and thoughtless, rather than driven; conflict relationships, suicidal preoccupation, self-mutilation, identity disturbances and feelings of abandonment are usually less intense than in borderline personality disorder. However, the differences may not be clear-cut because in both disorders symptoms are chronic and trait-like. Importantly, individuals presenting with a diagnosis of personality disorder who present with the ADHD syndrome that started in childhood will in many cases benefit from pharmacological treatments for ADHD [232, 233]. Recent trials in ADHD patients indicate that besides ADHD symptoms, mood instability improves with both stimulant treatment and atomoxetine [234, 235]. As the order of treatment will depend on the presence of and severity of co-morbidities, evaluation of co-morbid disorders is a key component of the ADHD assessment using appropriate clinical diagnostic approaches.