As children with ADHD initially present to professionals with problems affecting behavior and learning, a comprehensive assessment is necessary to confirm a diagnosis according to criteria previously outlined. It is also necessary in order to exclude other conditions and to determine whether co-morbid conditions exist. This assessment also provides the basis for the formulation of appropriate management strategies for presenting problems and conditions which underlie them. Thus, a careful history of the individual case is required, followed by relevant examinations, reports and additional investigations if indicated by findings from the initial assessment.
Multiple sources of information on duration of symptoms as well as assessment in multiple settings, such as school, home and clinic, increase the accuracy and validity of assessments and are necessary if diagnoses are to meet the criteria set down in DSM-IV. Rating scales assist in obtaining systematic information from different settings and treatment responses. It should be taken into account, however, that there are difficulties in categorizing behavior as either “normal” or “abnormal”. Even when rating scales with cutoffs are used, assessments of behavior components are subjective.
The goals of the evaluation process are to establish the child’s diagnostic status and to translate assessment data into a potentially successful intervention plan. According to DuPaul (1991) the optimal methodology for addressing both goals of the evaluation process uses an assessment approach which incorporates:
Medical assessment of a child with behavior and learning difficulties should include a full medical history, relevant neurological examinations and family history. The assessment should focus on physical and mental health, and should exclude vision and hearing defects. Observation in the clinical consultation is important but ‘clinic hyperactivity does not reflect the pervasiveness of symptoms (Tripp and Luk 1997).
Further assessments may be necessary for selected children. Children who appear to have auditory problems require additional comprehensive audiological assessment, including auditory acuity, ‘speech-in-noise’ discrimination and auditory memory. Psychological assessment is recommended for children who appear to be developmentally intact and/or who exhibit a specific disability in learning. Speech pathology assessment is necessary for children with significant language difficulties, especially in younger children. Occupational therapy assessment is recommended, especially for younger children who have problems with motor function, handwriting and spatial and body awareness difficulties. In addition, physiotherapy assessment may be necessary for some children with gross motor dysfunction, for whom self-esteem and confidence are impaired with resultant negative effects on behavior and learning.
Children with behavior and learning difficulties should have a full assessment of developmental functioning. Assessment should focus on determining delays in development and difficulties with gross or fine motor abilities, visual motor function, auditory short-term memory, receptive and expressive language and attention. Assessment of such skills may identify factors contributing to behavior, and especially learning, difficulties (such as poor fine motor skills causing slow and labored handwriting).
Assessment of visual-perceptual difficulties may indicate problems with letter reversals, copying forms, confusion over spatial relationships and organisation of work. Assessment of sequencing problems may identify difficulties in following instructions, and with the concepts of time and relationships. The assessment of language – expressive, receptive syntactical, semantic and pragmatic ÷ may assist in the understanding of behavior and learning difficulties. Developmental assessment should include assessment of the maturity of the central nervous system (CNS) as evidenced by delayed laterality (tendency in activity to prefer to use one side of the body) or altered control of complex movement.
Although the symptoms of ADHD are commonly observed before the age of six in a child who is later identified as having ADHD, it is often not until the child enters formal schooling that the seriousness of these symptoms is first noted, typically by a teacher.
Children who have ADHD may experience academic functioning problems on several levels. They may fail to learn or they may underachieve in the classroom due to the inattention and impulsivity imposed by their attention deficits. They may be unable to demonstrate what they do know because of difficulties, for example, with taking tests in a group. They may also have learning disabilities in addition to their problem with ADHD.
The assessment of children who have ADHD should encompass a review of school observations and previous testing. These should cover estimates of intellectual capabilities, strengths and weaknesses and measures of academic achievement, including language development.
The problems of intellectual or academic functioning in most children with emotional or behavioral disorders appear to be significant, yet findings in this area are characterized by considerable variability (Forness, Kavale, Guthrie et al 1987). There is some evidence that modes of test presentation, such as timed versus untimed components, may be a critical factor in at least some of this variability (Scruggs and Mastropieri 1986).
Hyperactive children tend to perform less effectively than nonhyperactive children on tests of perceptual-motor functioning, measures of sustained attention (especially in situations in which the stimulus is unpredictable), and measures requiring delay of impulse (Campbell 1976). In these and other situations, hyperactive children tend to respond quicker and make more errors than do nonhyperactive children. Possibly their poorer performance reflects problems of attention and impulse control rather than problems of perceptual-motor control.
Individual testing plays a minimal role in the evaluation of ADHD. This has tended to ignore the other influences on the child, namely his or her home life and relationship with peers. In an ecological perspective, emphasis is placed on understanding the child and his or her behavior in the context in which the child lives, works and plays. One of the main aims of ecology is to understand the interrelationships between the various ‘systems’ in an ‘ ecosystem’. To date, no individually-administered test or group of tests has demonstrated an acceptable degree of ecological validity to be helpful in the diagnostic process (Barkley 1991). Further, those tests typically employed by school psychologists (eg Wechsler intelligence scales; see below) have not been found to reliably discriminate ADHD from normal children (Barkley, DuPaul, and McMurray 1990). For example, below average scores on the ‘freedom from distractibility’ factor (ie arithmetic, digit span, and coding subtests) from the WISC-III are not necessarily a diagnostic indicator of ADHD. Poor performance on this factor may be due to a variety of possible causes ( Wielkiewicz 1990). Further, children with ADHD often display appropriate levels of attention and behavioral control under task conditions which are highly structured and involve one-to-one interaction with a novel adult, and which are found in most testing situations (Barkley 1990a). In the individual test situation, examiners can be responsive to lapses in attention and, by making sure that they have the child’s attention before administering the test questions, can affect the result. Group intelligence tests may underestimate the ability level of hyperactive children; consequently, an intelligence quotient (IQ) obtained by hyperactive children on group-administered tests must be interpreted cautiously.
The Weschler scales (WISC-R and WISC-III) are the most widely used tests of intellectual ability (Sattler 1990). In addition to verbal, performance and full-scale IQ scores, the WISC-III scale produces factors of freedom from distractibility and speed of information processing. In combination with observations during testing and neurodevelopmental findings, these can be useful in supporting the diagnosis of ADHD. Many experienced clinicians will confirm that there are consistent and characteristic features of children with ADHD that are identifiable during cognitive assessments which, when matched with other observations and information, genuinely assist in making the diagnosis. Individuals with ADHD regularly exhibit difficulties with planning, organizing and setting about tasks in an orderly and constructive fashion. They also have problems with mental flexibility ÷ often demonstrated as perseveration (continuing an action for an excessively long time). Although some scores on new learning measures may be within normal limits, ADHD children often demonstrate problems with initial registration (first trials), and with benefiting from repeated presentations (plateauing). In regard to task vigilance, even when distractibility is low, ADHD children regularly tire quickly during the testing session and their symptoms of impulsivity and disorganization intensify as they tire.
The Kaufman ABC test produces sequential and simultaneous processing scores along with the mental processing composite score and an achievement score. The sequential processing score is sensitive to attention deficits. The Kaufman ABC also minimizes language demands, which may be an advantage, particularly with language-disordered children who have ADHD.
According to Sattler (1990), the major difficulties of hyperactive children lie in their ability to focus, sustain, and organize attention and to inhibit impulsive responses. These difficulties are likely to be reflected in their performance on some, but not all, psychological tests. The scores obtained by children diagnosed with ADHD on individual intelligence tests may be more variable than those of normal children. No particular patterns of scores (for example, verbal-performance discrepancies or lower abstract reasoning scores) on the Weschler scales or on other intelligence tests have been found to be associated with hyperactivity. However, individual psychological testing is important to establish particular patterns of abilities and deficits among children referred for ADHD. The tests can provide information with respect to levels of development of cognitive and attentional skills, as well as a description of cognitive strengths and weaknesses essential to appropriate treatment and intervention. Ehlers, Nyden and Gillberg et al (1996) discuss how characteristic cognitive profiles may differ between such differential diagnoses as ADHD, Asperger syndrome and autism, but individual heterogeneity is still marked.
Educational assessment will include observation of the behavior of the child in the classroom and playground as well as a review of the cumulative school-based assessments of academic progress. The appropriateness of the classroom setting and the educational tasks required of the child should be routinely assessed. Academic achievement tests are important for qualitative and quantitative measurement of academic abilities. A speech and language assessment component is particularly relevant as part of the evaluation.
A comprehensive assessment of a child with suspected ADHD should include the following elements:
Multiple sources of information should be utilized during assessment eg parents, care givers and relevant professionals, especially teachers.
A variety of rating scales and other measures have been developed over the years to test for the presence of behavioral conditions, among them ADHD. Each have their strengths and weaknesses and demonstrate the difficulty of defining ADHD and co-morbid conditions. The most important of these rating scales are described below.
Historically, the most widely used rating scales in the United States were developed by Conners initially to measure response to medication for ADHD (Goyette, Conners and Urlich 1978; Conners 1976; 1973, 1969). Conners (1976) described a 39-item teacher questionnaire and a 98-item parent symptom checklist, as well as a modified 10-item parent-teacher questionnaire. The teacher and parent-teacher scales were suitable for use in drug studies (with test-retest reliabilities over a one-month period ranging from 0.72 to 0.91 for the teacher questionnaire).
The teacher questionnaire contained five factors labelled “anxious” “fearful”, “hyperactivity” and “sociable” ” cooperative”.
The 10-item parent-teacher scale contained overlapping parent and teacher items and was particularly useful for repeated measures in drug studies. Goyette, Conners and Urlich (1978) replicated Conners’ factor structures (groups of symptoms derived from factor analysis) for the parent and teacher questionnaires and found a parent-teacher correlation of 0.49 for hyperkinesis index (parent-teacher scale). Conners recommended that the 10-item index described in the paper by Goyette, Conners and Urlich (1978) be used. The scale tends to select a mixed group of hyperactive and aggressive children.
Glow (1980; 1978a and b) carried out two large Australian normative studies of the Conners’ Teacher Rating Scale (TRS) and Conners Parent Rating Scale (PRS) with 1919 and 2475 subjects, respectively. In general, the factor scales obtained were similar to Conners’ scales, but possibly because of larger numbers, she obtained a hyperactive- inattentive scale loading on to a single factor. The correlation between parent and teacher scales, however, was modest (0.22 for the hyperactive-impulsive parent and hyperactive-inattentive teacher scales), suggesting that behavioral traits cannot necessarily be recognized independently of the context in which they are manifested.
Glow used a cutoff of at or above two standard deviations above the mean on the hyperactive-inattentive scale to rate prevalence of ADD. A commonly used cutoff point on the parent-teacher scale is a score of 16 or greater out of a possible 30 points. However, Werry and Sprague (1974) indicated that these types of rating scales are unstable across time, analogous to a ‘practice effect’, occurring predominantly between first and second ratings.
It would appear that behaviors of hyperactive children may vary across situations. A child may be hyperactive at school, but not at home or vice versa (Vaesen and van der Meere 1990; Szatmari, Offord and Boyle 1989a; Schachar, Rutter and Smith 1981; Lambert, Sandoval and Sassore 1978; Sandberg, Rutter and Taylor 1978).
Methodologically it may be difficult to distinguish which part of the variance between teachers and parents is explained by the child’s behavior pattern, and which by the individual response bias of teachers and parents themselves. Thus situational specificity may account, in part, for epidemiological differences between the United Kingdom and the United States studies. It is thus important that diagnoses should specify the setting in which they are made. The DSM-III-R (American Psychiatric Association 1987) allowed diagnosis at home or at school, whereas the DSM-IV and ICD-10 require that diagnoses be made in multiple settings.
It should also be noted that while the brief 10-item Conners parent-teacher scale (a combined scale designed for use by both parents and teachers) was once the most frequently used instrument for selecting hyperactive children for research, more recent research suggests that the scale assesses a mixture of hyperactive and conduct problems ( Barkley 1990a) and will select a mixed group of hyperactive/conduct-disordered children.
The Child Behavior Checklist (CBCL), Teacher Rating Form and Youth Self-Report (Achenbach 1991a,b; Achenbach and Edelbrock 1983) are probably the most carefully developed and standardized methods for addressing a broad array of psychopathological manifestations in children, and have been widely used, particularly in epidemiological studies.
Achenbach (1978) developed questionnaires based on extensive studies of children’s psychiatric symptoms, derived from case files. He published a series of child behavior profiles standardized separately for children of each sex at ages 4-5, 6-11, 12-16 years. The profiles are scored from the CBCL which was designed to obtain parent’s reports of their child’s competencies and problems in a standardized format. For example, the profile for boys aged 6-11 years, yielded nine behavior problem scales which were labelled: schizoid, depressed, uncommunicative, obsessive- compulsive, somatic complaints, social withdrawal, hyperactive, aggressive and delinquent. Second-order factors labelled “internalizing” and “externalizing” were obtained.
Recent versions have been published in the Manual for Child Behaviour Checklist (4-18) and Teachers’ Report Form and Manual for the Teachers Report Form and 1991 Profile (Achenbach 1991a,b)
Achenbach, McConaughy and Howell (1987) reported relatively low agreement between teachers and parents (similar to the variable correlations for Conners parent and teacher scales). This finding has important implications for situational specificity of ADHD symptoms, discussed above. The information obtained in the CBCL can be scored four different ways:
A problem with the scale is that it is designed for professional or semi-professional use, so that some items, such as “compulsions”, may make it difficult for parents to use.
Achenbach, Hensley, Phares et al (1990) have reported a comparative study of problems and competencies reported by parents of 2600 randomly selected Australian and American children, in the CBCL. Sydney children scored significantly higher on 82 problem items, but item scores correlated 0.92 between countries. Higher clinical cutoffs may be needed in Australia. Bond, Nolan, Adler et al (1994) studied 1009 Melbourne children, and reported results more similar to United States norms than in the Achenbach, Hensley, Phares et al (1990) Sydney study. The Western Australian Child Health Survey (Zubrick, Silburn, Garton et al 1995) has also developed norms using an abbreviated version of the Child Behavior Checklist.
Since the advent of the Diagnostic and Statistical Manual of Mental Disorders, 3rd edition revised and 4th edition (1987, 1994), rating scales based on criteria from these classification systems have begun to replace older scales.
A 14-item questionnaire (DuPaul 1991) based on DSM-III-R (American Psychiatric Association 1993) which was often used to monitor drug responses has now been superseded by an ADHD behavior checklist based on DSM-IV (American Psychiatric Association 1994) and devised by Barkley (1995a). The current available norms for this checklist, however, are based on adult norms. A very similar questionnaire based on DSM-IV has been designed by Levy and Wood for use in the Australian Twin ADHD Project (ATAP) (Levy, Hay, McLaughlin et al 1996). Rating scales that functioned well with DSM-III-R (Chen, Faraone, Biederman et al 1994), will require their relevance to DSM-IV to be determined.
A number of newer DSM-IV-based questionnaires and rating scales are currently marketed. These include:
The strengths and weaknesses of these newer scales will require future research.
The newer Behaviour Assessment Schedule for Children (BASC) (American Guidance Service 1992) may be closer to current conceptualizations of heterogeneity within ADHD. They consist of a teacher, parent and self-report scale, age range 4-18 years. A structured developmental history interview and a form for recording behavioral observations in the classroom are provided. The BASC items were derived from confirmatory factor analyses. Rowe Behavioral Rating Inventory (Rowe and Rowe 1995). This scale was initially designed to measure childrens reactions to synthetic food colorings. It incorporates parent and teacher ratings of behavior including attention and activity. The interesting feature of these scales is the fact that items are both positively and negatively worded and may therefore control for the skewed distributions that often occur with negatively-worded rating scales.
Barkley (1990a) reviewed behavior rating scales, including Conners’ PRS; Conners’ Abbreviated Symptom Questionnaire; Personality Inventory for Children; Werry-Weiss Peters Activity Rating Scale; Conners’ TRS; Conners’ Abbreviated Symptom Questionnaire; Iowa Conners’ TRS; ADD-H Comprehensive TRS; Swanson, Nolan and Pelham Rating Scale; Original Behavior Problem Checklist; Revised Behavior Problem Checklist; Preschool Behavior Questionnaire, and the Self-Control Rating Scale. The validation data and uses of these questionnaires are thoroughly discussed in Barkleys review.
Barkley pointed out that some of these scales are now outdated. At present, DSM-based rating scales, such as the DSM-III-R-based questionnaire (DuPaul 1991), are most useful for parents and teachers. The ADHD Behavior Checklist for Adults (Murphy and Barkley 1995) is a useful scale based on DSM-IV. G DuPaul (personal communication) is preparing DSM-IV norms for children. The Conners’ scale has been re-normed for DSM-IV and is now available as the Conners ADHD/DSM-IV Scales (CADS) (1997).
Appropriate rating scales should be used as part of the assessment for obtaining systematic information from different settings and to gauge treatment response.
Diagnostic interviews are semi-structured interviews that are most used in research and administered by professionals or trained lay interviewers. A number of interview formats have been described as follows:
1. Diagnostic interview for children (DISC-2C and DISC-2P, child and parent versions) (Shaffer, Schwab-Stone, Fisher et al 1993; Costello, Edelbrock, Dulcan et al 1984). This is a structured clinical interview which provides DSM-III-R-based diagnosis. Interrater reliability is very good (Shaffer, Schwab-Stone, Fisher et al 1993). For the diagnosis of ADHD, the Kappa coefficient is 0.55 (Schwab-Stone, Fisher, Piancentini et al 1993). DISC-2 also provides diagnosis for disruptive behavior disorders, Conduct Disorders and Oppositional Defiant Disorder.
2. Diagnostic interview for children and adolescents (DICA) (Herjanic and Reich 1982). This is a structured interview covering among other things disruptive behavior disorders, affective disorders, anxiety disorders and alcohol abuse. There is a version for 6-12 year-olds and for 13-17 year-olds. Interrater reliability was 95 per cent average.
3. Child assessment schedule (CAS) (Hodges, Kline, Stern et al 1987). This is an interview designed for 7-12 year-old children that can be administered by clinicians and lay interviewers. The reliability of algorithm- generated diagnosis was 0.43 for ADD.
Barkley (1991) has examined the ecological validity of laboratory tests for hyperactivity. In general, low correlations were reported between laboratory measures of attention, impulsivity and activity on the one hand, and parent and teacher ratings on the other. However, given the low correlations (situational specificity) between parent and teacher ratings, and the possible heterogeneity and co-morbidity of hyperactive samples, objective tests may ultimately prove as ecologically valid as rating scales (and may help to distinguish different groups). Few studies are currently available for comparing laboratory tests with direct observations in natural settings.
In recent years, there has been an increasing interest in vigilance (sustained attention) testing (Levy and Hobbes 1997). A number of commercial age-normalised vigilance tests have become available.
1. Conners Continuous Performance Test (CPT) (Conners 1992, 1994)
In traditional CPT tests, a respondent presses a button when a specific letter or series of letters appears on a screen. In the Conners’ test the respondent presses the appropriate key for any letter except the X. Norms are available from 4-18+ years.
2. TOVA (Test of Variables of Attention) (Teicher, Ito and Glod et al 1996; Greenberg and Waldman 1993; Greenberg 1987)
This test presents geometric shapes and measures errors of omission, commission and reaction times. Administration 25 minutes. Age range 6-15 years.
3. Vigil Continuous Performance Test (Psychological Corporation 1995)
This test is for use by neuropsychologists and clinical psychologists. Specificity and predictive value are not yet available. Administration time is 8 minutes per test. Age range 6-90 years.
The main problems with these tests are the variety of versions used, often with inadequate standardization and lack of comparability. Anastopoulos and Costabile (1995) proposed that the continuous performance test holds much promise as a clinic-based assessment, while other authors such as Barkley (1995a) questioned the impact of false positive and false negative results, indicating the need for sensitivity and specificity studies. Levy (1980) and Levy and Hobbes (1981) have reported Australian norms for the CPT.
These more detailed interviews and tests are not currently recommended as a necessary part of routine assessment of children who are suspected of having ADHD.
A number of neurophysiological techniques have been investigated, providing some information about underlying physiological processes.
It was proposed that statistical analysis of standardized quantitative electrophysiological features relative to a body of normative data might aid in the differential diagnosis of a variety of subtle brain dysfunctions (John, Prichep and Easton 1987; John, Prichep, Katz et al 1987; John, Karmel and Corning 1977). Comparisons are made to an age-appropriate normative database and results re-expressed in units of standard deviation and resulting scores re- imaged as significance probability maps (SPMs).
Brain maps are produced from 16 to 32 electrodes arranged in a grid pattern on the scalp giving a spatial resolution of about six centimeters. This is at best one quarter of the resolution required for accurate representation of cerebral electrical fields and providing no means of distinguishing between cerebral fields and artifact (Duffy 1988; 1986).
Sensitivity and specificity of these techniques have been investigated (Chabot, Merkin, Wood et al 1996), but at present they remain an experimental technique and are not necessary in routine clinical practice (Levy and Ward 1995).
Steady state probe topography (SSPT) is a novel methodology, utilizing the steady state visually evoked potential (SSVEP) elicited by a sinusoidal visual flicker. Analysis of the SSVEP at 64 scalp sites during a visual vigilance task allows high temporal resolution of brain electrical correlates of extended tasks, combined with high spatial resolution (Silberstein 1995; Silberstein, Schier, Pipingas et al 1990). The role in ADHD assessment and management is still being researched.
The positron emission tomography (PET) technique measures glucose metabolism in areas of the CNS involved in attention processes. However, the use of radioisotopes limits its clinical use in children. A PET study by Zametkin, Nordahl and Gross (1990) found lower prefrontal glucose utilization in parents of hyperactive children. However, a second study by Zametkin, Liebenauer, Fitzgerald et al (1993) in teenagers did not replicate the finding. However, NIMH studies in the United States by Matochik, Liebenauer, King et al (1994) and Matochik, Nordahl, Gross et al ( 1993), investigating the effects of acute and chronic treatment with stimulant drugs, showed that stimulants increased glucose metabolism in some areas and decreased it in others. PET studies using L-dopa are now being undertaken.
Lou, Henrickson and Bruhn (1984) used single-photon emission computed tomography (SPECT) to examine cerebral blood flow in ADHD children, and found hypoperfusion of caudate and frontal lobes. This technique is also still an experimental technique, not yet suitable for routine clinical use. O’Tauma and Treves (1993) have reviewed its use in investigating children’s behavior disorders.
Some work has been done on the brain anatomy of children using magnetic resonance imaging (MRI). Hynd, Semrud- Clikeman, Lorys et al (1991) have reported smaller than average right caudate areas and differences in anterior and posterior corpus callosum in ADHD children.
Further investigations such as brain imaging and neurophysiological tests are not recommended as part of the routine assessment of ADHD.
Medical assessment focuses on physical and mental health and contributes to diagnoses of intrinsic medical, psychiatric and other problems which account for behavior and learning difficulties. Developmental assessments provide information on development strengths and weaknesses. These evaluations may indicate a need for further assessments of vision, hearing, language and processing abilities.
Reviews of classroom observations and previous testing form the basis of educational assessment. No individual educational assessment method has an acceptable level of ecological validity to obtain a diagnosis of ADHD, although assessments, particularly tests such as the WISC-III, can be useful in supporting the diagnosis of ADHD. An array of academic achievement tests are useful in assessing academic difficulties and identifying children with behavior or learning difficulties.
Rating scales in both home and school settings can be used to obtain either trait or categorical diagnoses.
A number of semi-structured interviews and other investigations are available for use by researchers. A number of vigilance tests have been approved by the psychological test marketing bodies for use by trained professionals.
Neurophysiological investigation and brain imaging techniques are generally at a research stage of development and, according to the current state of knowledge, are not appropriate for routine clinical practice.
The combination of results from medical, developmental and educational assessments, including the use of rating scales, is the minimum required to establish or exclude the diagnosis of ADHD according to DSM-IV criteria, to exclude or diagnose other conditions which may have similar presenting features and to determine whether co-morbid conditions exist.