Is the Conners’ continuous performance test helpful for assessing attention deficit hyperactivity disorder in a clinical setting?

Abstract

Background

Despite lacking validation for Norwegian populations, the Conners Continuous Performance Test II (CCPT-II) is applied to almost one-third of children receiving an ADHD diagnosis. However, evidence of the CCPT-II’s ability to differentiate between children with and without ADHD is contradictory. Thus, this study examines how CCPT-II results correlate with ADHD symptoms reported by mothers and teachers in a sample representing ordinary child and adolescent mental health services and explores the extent to which the CCPT-II influences the diagnostic result.

Methods

Correlations between CCPT-II results and ADHD Rating Scale scores and a clinical diagnosis of ADHD were analysed in children aged 6–15 years (N = 69) referred to a child and adolescent psychiatric outpatient clinic.

Results

Total ADHD symptom scores rated by mothers correlated with hit reaction time (HRT) block change (.260), HRT inter-stimulus interval (ISI) change (.264) and CCPT-II overall index (.263), while hyperactivity subscale scores correlated with omissions (.285), HRT (.414) and variability (.400). In teachers’ ratings, total ADHD and both subscale scores correlated with commissions (.280–.382), while hyperactivity scores correlated with variability (.265). A higher number of commissions was the only significant difference in CCPT-II performance between children diagnosed with and children without ADHD.

Conclusions

Correlations between CCPT-II results and ADHD symptoms were all small to moderate. As such, CCPT-II results should be interpreted with caution, because they correspond to a limited degree with other sources of information.

Keywords:

• Continuous performance test
• attention deficit hyperactivity disorder
• conners continuous performance test
• ccpt-II

Introduction

Attention deficit hyperactivity disorder (ADHD) has over the last several decades become the most frequently diagnosed neurobehavioral disorder of childhood. In 2016, 8.4% of American children aged 2 to 17 years had been given an ADHD diagnosis [1]. Based on data from the Norwegian Patient Registry 2008–2011, by the age of 12 years, 3.8% of Norwegian children (5.4% of boys and 2.1% of girls) had been diagnosed with a hyperkinetic disorder [2]. Restlessness, impulsiveness and concentration difficulties may prevent many of these children from succeeding in school, as impairments to language, motor coordination, cognition and learning are often associated with the diagnosis [3]. A linkage study of Swedish national registers covering all 657,720 students graduating from primary and lower secondary school in Sweden in 2008-2013, including a large sample of students diagnosed with ADHD (n = 29 128, 4.4%) concluded that ADHD was associated with substantially lower school performance independent of socioeconomic background factors. Students with ADHD had an increased risk of not being eligible to higher education (Upper secondary school) at 37.6% compared to 10.7% in the student population unaffected by ADHD [4]. ADHD in childhood is highly associated with later receiving an official disability pension [5], and it is associated with a higher risk of developing a substance use disorder [6].

As much as 40% of children with ADHD are reported to have oppositional defiant disorder (ODD) and a greater risk of developing conduct disorders [7] and, at worst, a criminal career [8]. Between half and two-thirds of children with ADHD have at least one other psychiatric diagnosis, whether an anxiety disorder, mood disorder, sleep disorder, tic disorder or autism spectrum disorder, all of which are common comorbid conditions [3,9], representing extra challenges for many of these children. Furthermore, between 60% and 70% of children diagnosed with ADHD will continue to display the criteria of the disorder during their teen years and into adulthood [10]. In summary, ADHD has a lifelong impact on schooling, employment, family and social life. Studies of long-term outcomes of ADHD show that treatment may reduce the negative impact that untreated ADHD has on life functioning [4,11], making it crucial to diagnose the condition reliably and to offer available treatments and services. The high prevalence of comorbid disorders leads also to diagnostic challenges due to symptom overlap and complex relationships between the disorders mutually influencing each other [12].

However, diagnoses in practice differ remarkably in frequency between different nations when compared to research estimates, suggesting that the condition may be either over-diagnosed or missed in clinical practice [13]. In Norway, considerable variations in the prevalence of ADHD between counties, ranging from 1.4% to 5.1%, have been demonstrated, most likely due to variations in diagnostic practices [2,14]. A review of medical records (n = 549) showed that only about half of diagnoses were reliably documented in records. Among the other half, the researchers found a discrepancy between the information in the medical record and the diagnosis, indicating that the child did not meet the diagnostic criteria (38%), information is missing from the medical records (48%) or a lack of differential diagnostics (46%) [2].

National guidelines outline mandatory elements such as mapping and documentation of psychosocial, developmental, somatic and psychiatric history, status and symptoms to be included in in the assessment of ADHD in children and adolescents in all Norwegian Child and Adolescent Mental Health Services (CAMHS) [15]. However, the use and interpretation of assessment tools seem to vary. In addition, a diagnosis of ADHD is primarily clinical, based on observations and interpretations of characteristic behaviours and symptoms of inattention, impulsivity and hyperactivity. The above-described diagnostic difficulties underline a need for supplementary assessment tools designed to aid in the reliable identification of individuals fulfilling the diagnostic criteria.

Computerised continuous performance tests, which measure attentiveness, impulsivity and vigilance, have been used as supplementary tools to differentiate between children with and without ADHD. However, the evidence of their ability to support clinical decision-making is conflicting [16]. During these tests, the child is observed in a situation in which they are presented with a boring task to complete on a computer. The Conners Continuous Performance Test II (CCPT-II) [17], for example, lasts approximately 14 min, and it involves the participant pressing the space bar or clicking the mouse button when a specific letter or symbol appears on the screen. To test impulsivity, in between, another specific letter or symbol appears that should not be clicked. The child’s total performance, changes in performance over time, attitude and behaviour, including uneasiness and the occurrence of motor restlessness during the session, were registered and evaluated. Sustained attention, impulsivity and vigilance are analysed with such test parameters as hit rate, omissions, commissions, reaction time (RT) and reaction time variability (RTV). Primarily, the test is assumed to evaluate attention problems and treatment usefulness, where response patterns on the CCPT provide information that may sort the various deficits present. A factor analysis of the CCPT [18] indicated that the test measures five sub-functions of attention, namely focused attention, hyperactivity/impulsivity, sustained attention, vigilance and change in control. The same authors tested the validity of these factor constructs in a comparison with 282 individuals with ADHD, schizophrenia, affective disorders, brain injury, language disorders and normal individuals in whom four of the factors are valid, excluding vigilance. The authors concluded that CCPT results should not be reduced to one overall score of attention and that the concept of sustained attention should be used for measures of change as a function of time on task. Differing between five sub-functions of attention made the test more useful for differentiating between various clinical groups suffering from impaired attention [19].

The CCPT-II, and later the third revised version, the CCPT-III, has been widely used in ADHD research and clinical assessments of participants aged 6 years or older. For a couple of decades, it has been part of the assessment in many Norwegian CAHMS. Close to one-third of children receiving an ADHD diagnosis were tested with the CCPT-II as part of the diagnostic process [2]. Despite lacking validation for Norwegian populations, the CCPT-II seems to have a dominating position as the most frequently applied test in the assessment of ADHD in Norway. About two-thirds of neuropsychologists practising in Norway have used the CCPT-II [20]. Meanwhile, the National Guidelines for the Assessment of ADHD refer to the frequent use of CPTs, such as the CCPT-II, in clinical practice, and they advise that the results, while not diagnostic, may contribute to a comprehensive evaluation [15]. Despite its intensive utilisation as an assessment tool in Norway, little is known about the role of CCPT-II results in the assessment of ADHD.

The aim of the study was primarily to examine how the CCPT-II results correlate with core symptoms of attention deficiency, hyperactivity and impulsiveness, as reported in the ADHD rating scales by mothers and teachers of the patients in a sample representing ordinary CAMHS. We also aimed to explore to what extent the CCPT-II was decisive for the diagnostic result.

Materials and methods

Participants

The sample (n = 69) consisted of consecutively recruited children and adolescents aged 6–15 years, referred to a Child and Adolescent Psychiatric Outpatient Clinic in the Central Part of Norway for attention problems, hyperactivity, impulsive behaviour and/or behaviour problems over a two-year period (2003–2005). The referrals came from school psychology services (58%), general practitioners or other health care providers in schools and somatic wards (30%) and child welfare services (13%). About one-third of the children were living with one caregiver, the others with two caregivers and all but one family were native Norwegians.

The mean age was 9.9 years (SD 2.8) and 55 participants (79.9%) were boys. Of the entire sample, 67% received a diagnosis of hyperkinetic disorder/ADHD, specified as ADHD combined type in 35 children (76%), as ADHD, predominantly inattentive type in 3 children (7%) and as Hyperkinetic conduct disorder, i.e. ADHD associated with conduct disorder (F 90.1 in ICD 10) in 8 children (17%). The presence of comorbid disorders in the ADHD group was 44%, with conduct disorder being the most common (35%), followed by tic disorders (9%) and separation anxiety (7%). In the non-ADHD group, conduct disorder was most frequently diagnosed (n = 3), followed by separation anxiety, a specific phobia, OCD and tic disorder (n = 1, respectively). For demographic details of the participants, see Table 1.

Table 1. Sample characteristics.

	Children receiving an ADHD Diagnosis N = 46			Children not receiving an ADHD Diagnosis N = 23
	N	Mean	SD	N	Mean	SD	P
Age	46	9.8	3.1	23	10.0	2.4	0.163
ADHD-Rating Scale total mothers	44	30.1	10.9	22	25.1	11.0	0.097
ADHD-Rating Scale total teachers	42	31.1	11.0	22	17.8	10.5	0.000***
	N	%		N	%		P
Female	11	24		3	13		0.356
With comorbid disorder	20	44		7	30		0.433
>1 comorbid disorder	4	8.7		1	4.3		0.658

***p<.001 after Bonferrroni correction.

Diagnostic procedures

All children underwent a standard assessment procedure, while information from parents and teachers was obtained through an open clinical interview and by completing rating scales (see measures below). The Achenbach System of Empirically Based Assessment (ASEBA) [21] was filled in by patients over 11 years of age (Youth Self Report [YSR]), the caretakers at home filled in the Child Behaviour Check List (CBCL) and the class teachers of all patients filled in the Teachers’ Report Form (TRF). Because mothers were the available informants in all cases (in contrary to fathers), we chose to use only the mothers’ ratings for the analyses. All children (except one) underwent a medical examination and an assessment of intellectual functioning with an age-appropriate Wechsler test [22]. Preschool children were tested with the Wechsler Preschool and Primary Scale of Intelligence-Revised (WPPSI-R) and school-aged children with the Wechsler Intelligence Scale for Children-Revised (WISC-R) or the WISC-III, while all children were tested with the CCPT-II. An interdisciplinary team, including an experienced physician, psychiatrist or psychologist, analysed the assessment results, leading to a consensus on the diagnosis, according to the International Statistical Classification of mental and behavioural disorders 10th edition (ICD 10) [23]. A trained and experienced psychologist, who was not part of the assessment team, conducted the Kiddie-Schedule for Affective Disorders and Schizophrenia-Present and Lifetime version (K-SADS-PL) interview with caregivers to evaluate the ADHD diagnosis independently and to assess comorbidity. The independent rater had no access to other information, but the results of the K-SADS-PL were made available to the assessment team.

Measures

The ADHD Rating Scale IV (ADHD-RS) [24] presents a list of 18 ADHD core symptoms, according to the Diagnostic and Statistical Manual of Mental Disorders IV (DSM IV) [25], the presence and frequency of each of which are rated on a scale from 0 to 3 (0 = never or rarely, 1 = sometimes, 2 = often and 3 = very often). Further, there are two subscales, where nine of the 18 symptoms reveal attention problems and the remaining nine reveal symptoms of hyperactive/impulsive behaviour. Scores range from 0–27 for the subscales, while total scores for all 18 symptoms range from 0–54. Accordingly, the ADHD-RS has been shown to have satisfactory psychometric properties [26]. The ADHD-RS showed excellent internal consistency for both mothers’ (α = .92) and teachers’ ratings (α = .92) in the present sample.

The Kiddie Schedule for Affective Disorders and Schizophrenia-Present and Lifetime version [27] is a widely used semi-structured interview for the diagnostic assessment of DSM-IV psychiatric disorders and subsyndromal symptomatology in children and adolescents. The K-SADS-PL was used to confirm the diagnosis of ADHD according to the DSM-IV and to evaluate comorbidity.

The CCPT-II [17] is a computerised test that assesses a person’s attention capacity over a period of approximately 14 min. The test does not require specific skills, but the person tested must be able to recognise letters as targets or non-targets. During the test, letters are presented on a screen one at a time. All letters except the letter X are targets to which the test person must respond, while the Xs represent non-targets. Participants respond to a target by pressing the spacebar on the computer. The letters are presented in six blocks during the test, and the time between letters presented (the inter stimulus interval [ISI]) varies among one, two and four seconds [17]. In our survey, the CCPT-II test programme was installed on a stationary computer in a room as free from disturbing visual and auditive stimuli as possible. An observer from the staff was present in the room during the test. A definition of the different CCPT-II parameters is given in Table 2.

Table 2. Definitions of the parameters measured with the CCPT-II, modified after (17].

Omissions – errors occurring from the failure to respond to target letters (non-X’s)

Commissions - errors occurring when responses are given to non-targets (X’s)

Hit reaction time - overall (Hit RT) - the average speed in milliseconds of correct responses for the entire test over all six blocks. High T -scores reflect fast response times

Hit RT Standard Error – overall (Hit RT Std Error) - a measure of response speed consistency; higher values relate to greater inconsistency

Variability of standard error – the amount of variability the individual shows in 18 separate segments of the test in relation to his or her own overall standard error. The higher variability of standard error, the greater the inconsistency of the response speed

Detectability (d’) - a measure of the difference between the signal (non-X) and noise (X) distributions. d’ provides information about an individual’s ability to discriminate between targets and non-targets

Response Style indicator (beta) - differs between cautious individuals, those more seldom responding than most other individuals, and those more freely responding

Hit RT Block Change - measures change in reaction time during the test; high values indicate slowing in reaction times whereas low values indicate fastening reaction times as the test is progressing

Hit SE Block Change - changes in response consistency over the duration of the test. High values indicate a loss of consistency as the test progresses, while low values indicate improved consistency

Hit RT ISI Change – measures change in average reaction time at the different inter stimulus intervals (ISI), in the test there are 1, 2 or 4 seconds intervals

Hit SE ISI Change – measures change in the standard error of reaction times at the different Inter Stimulus Intervals (ISI). High values indicated that the response time became less consistent with increasing intervals, higher consistency at increasing intervals gave lower values

Overall Index – is a weighted sum of the measures listed above

The measure of Perseverations - slow responses, random responses, or repeated responses given without considering the forgoing stimulus or task requirement (was let out because only a few of the individual responses in our sample could be analysed)

Statistics

Group comparisons for continuous variables were evaluated with independent sample t-tests or Welch’s t-tests, while group comparisons for dichotomous data were evaluated with chi-squared tests, and the level of significance was set to p <.05 two-sided (Tables 1 and 4). Due to the lack of a normal distribution and numerous outliers, the potential correlations between mothers’ and teachers’ ratings of ADHD symptoms using the ADHD-RS and the patient’s CCPT-II performance parameters (number of omissions, number of commissions, hit rate, variability, detectability, response style, hit reaction time (HRT) block change, HRT ISI Change and CPT overall index) were assessed with a two-tailed Spearman’s Rho analysis (Table 3). Thresholds for statistical significance were set at .05 two-sided. A correlation of ≥ .50 was interpreted as large, .03–.49 as medium and .10–.29 as small [28]. Consensus between K-SADS-PL rating and final diagnosis is shown in a frequency table (Table 5).

Table 3. Unadjusted Spearman Rho correlations between mothers’ and teachers’ ADHD rating scales scores and CCPT-II parameters (n = 69).

ADHD-RS scores	CCPT-II Omissions	CCPT-II Commissions	CCPT-II Hit Reaction time	CCPT-II Variability	CCPT-II Detectability	CCPT-II Response style	CCPT-II Hit RT Block Change	CCPT-II Hit RT ISI Change	CCPT-II Hit RT ISI Change - SE	CCPT-II Overall Index
Inattention:
Mothers	.094	.083	−.069	.028	−.003	.042	.166	.129	−.044	.178
Teachers	.096	.382**	−.256	.004	−.155	.129	−.224	.157	.090	.124
Hyperactivity/Impulsivity:
Mothers	.285*	−.104	.414***	.400**	.082	.094	.271*	.264*	.217	.263*
Teachers	.151	.280*	.038	.265*	−.201	.169	−.123	.224	.246	.049
Total score:
Mothers	.218	.010	.182	.249	.033	.086	.260*	.264*	.217	.263*
Teachers	.159	.371**	−.128	.164	−.191	.129	−.224	.157	.090	.126

*p < 0.05, **p < 0.01, ***p < 0.001.

Table 4. Uncorrected differences between children with or without ADHD diagnosis and children’s CCPT-II performance.

	Children receiving an ADHD Diagnosis			Children not receiving an ADHD Diagnosis
	N	Mean	SD	N	Mean	SD	t-value	p
CCPT-II Omissions	40	21.4	20.4	22	14.2	14.0	1.641	0.106
CCPT-II Commissions	40	27.0	6.2	22	23.1	6.6	2.196	0.032*^a
CCPT-II Hit RT	40	400.1	105.9	22	425.8	75.1	1.104	0.272
CCPT-II Variability	40	29.3	24.3	22	24.7	20.3	0.792	0.431
CCPT-II Detectability	40	0.7	0.7	22	1.1	0.9	1.983	0.052
CCPT-II Block Change	40	0.0	0.0	22	0.0	0.1	1.646	0.105
CCPT-II Hit RT ISI Change	40	0.1	0.1	22	0.1	0.1	1.974	0.053

^aafter Bonferroni correction not significant.

Table 5. Diagnostic consensus between Kiddie-SADS rating and final diagnosis of ADHD (N = 69).

Final conclusion ADHD yes 46 (66.7 %) / Final conclusion ADHD no 23 (33.3 %)	Frequency (%)
Agreed ADHD yes, in Kiddie-SADS and in final conclusion	38 (55.1 %)
Agreed ADHD no, in Kiddie-SADS and in final conclusion	15 (21.7 %)
Disagreed ADHD yes in Kiddie-SADS, but ADHD no in final conclusion	8 (11.6 %)
Disagreed ADHD no in Kiddie-SADS, but ADHD yes in final conclusion	8 (11.6 %)

Ethics

Informed consent was obtained from the parents of all children, and the study was approved by the Regional Committee for Medical Research Ethics, Health Region IV and by the Norwegian Data Inspectorate (REK nr. 4.2005.713). Data handling was approved by the Norwegian Centre for Research Data (NSD nr. 12499).

Results

The ADHD symptom total scores rated by mothers correlated significantly with three CCPT-II performance parameters: HRT block change (ρ = .260), HRT ISI change (ρ = .264), CCPT-II overall index (ρ = .263) and the total number of commissions (ρ = .371) in teachers’ ratings (Table 3). In the inattention subscale, only commission errors showed a moderate positive correlation (ρ = .382) with inattention symptoms rated by teachers.

The ADHD-RS subscores of hyperactivity and impulsivity rated by mothers correlated with the following CCPT-II measures: omissions (ρ = .285), HRT (ρ = .414), variability (ρ = .400), HRT block change (ρ = .271), HRT ISI change (ρ = .264) and CCPT-II overall index (ρ = .263). The correlations were small to moderate, and in contrast to teachers’ ratings, where the only significant correlations between the ADHD-RS hyperactivity and impulsivity subscores and CCPT-II parameters were the number of commissions (ρ = .280) and variability (ρ = .265) (Table 3).

The only statistically significant difference in the CCPT-II performance between children diagnosed with and children not diagnosed with ADHD was found in the higher number of commissions for children diagnosed with ADHD (p<.05) (see Table 4). After Bonferroni correction this significance disappeared.

To delineate the influence of the K-SADS-PL rating on the final diagnosis, we analysed the agreement between the independent K-SADS-PL rater and the final diagnosis (Table 5). The total agreement was 76.8%, which is relatively high.

Discussion

The CCPT-II is designed to assess a wide range of processes related to a person’s attentional capacity, and there is considerable evidence that continuous performance tests may help to differentiate between children with and children without ADHD. The first meta-analysis, including 26 paediatric studies in 1996, concluded that children with ADHD made significantly more errors of omission and commission than normal children [29]. However, the CCPT’s role in differentiating individuals with ADHD from those not fulfilling the diagnostic criteria was questioned in later studies [30–32]. Meta-analyses have in part confirmed and refined these results [33,34]. In our study, the only significant difference in CCPT-II performance between children diagnosed with and children not diagnosed with ADHD was a higher number of commission errors in the group of children with ADHD. This finding may be unsurprising, as a large metanalysis with 47 between-group studies of CPT performance in children with ADHD [34] emphasised that artefacts and measurement unreliability account for a considerable proportion of the variance in CCPT-II results. Because the overall number of omissions and commissions is unable to address performance over time, these parameters will not assess the quality of sustained attention. Rather, such test parameters as HRT block change, HRT ISI change and variability are assumed to reflect performance throughout and among the various parts of the test, giving support to the assertion that a sustained attention deficit could only be claimed when error rates deteriorate over time.

In our sample, children’s CCPT-II performance showed only limited correlations with ADHD total symptom scores assigned by mothers and teachers. We found only small correlations between three CCPT-II parameters (HRT block change, HRT ISI change and CCPT-II overall index) and ADHD-RS total scores rated by mothers. For teachers’ ratings of the ADHD total scores, we found a moderate correlation with the number of commissions but no correlation with other CCPT-II measures. A similar pattern was reflected in the inattention subscale, where commissions again showed a moderate correlation with inattentiveness rated by teachers, but not by mothers. In contrast, the hyperactivity and impulsivity subscale scores rated by mothers showed small to medium correlations with several CCPT-II measures, including omissions, HRT, variability, HRT block change, HRT ISI change and CCPT-II overall index, but not commissions. This is somewhat surprising, as omission errors are thought to reflect inattention, while commission errors reflect impulsivity. However, this is in line with the findings of the above-mentioned metanalysis [34], which found the overall number of omissions and commissions unable to address performance over time. Teachers’ hyperactivity and impulsivity subscale scores showed a small correlation with the number of commissions and with variability, while the mothers’ reported symptoms correlated with the measures associated with performance over time. However, all correlations were only small to medium, and the number of omission errors showed only a small correlation with hyperactivity/impulsivity subscale scores by mothers and no correlation with teachers’ scores at all.

HRT and variability are measures associated with an individual’s ability to perform over time, though response time variation is assumed the parameter having the strongest impact on effect size regarding how children with ADHD differ from non-clinical groups [35]. Our results identify HRT and variability as the measures with the highest correlations to the mothers’ ratings of ADHD symptoms on the hyperactivity and impulsivity subscale. Variability also showed a moderate correlation with teachers’ hyperactivity and impulsivity subscale ratings, but we found no difference in these measures in relation to a diagnosis of ADHD compared to the group without ADHD. These findings are to some extent consistent with prior studies of CPT performance in children with ADHD [35].

A large epidemiological study (n = 817, age 9–17 years) found primary systematic effects of improved performance with older age for all CPT variables [36]. In 2.6% of the children (n = 21) who were diagnosed with an ADHD symptom profile based on the Child and Adolescent Psychiatric Assessment (CAPA) [37] and the addition of four items from the ADHD-RS that were not covered in the CAPA-interview, most of the CPT variables had no relationship to specific ADHD symptoms. Instead, the CPT parameters demonstrated relationships with multiple ADHD symptom domains. One CPT variable, mean HRT, was minimally related to ADHD symptoms overall but showed some specificity to symptoms of hyperactivity [38]. In our study, HRT had also the highest correlation (r = .414) with mother-rated hyperactivity subscale scores.

A large study with 379 children calculated the likelihood of predicting a best-estimate ADHD diagnosis based on measures from the Achenbach System of Empirically-Based Assessment (ASEBA), namely attention problems from the CBCL and the TRF versus the CCPT [39]. The parameter HRT Standard Error showed adequate diagnostic efficiency and unique contributions to the prediction of the ADHD combined type (but not the ADHD predominantly inattentive type) and was recommended for use in a supplemental diagnostic assessment in cases in which diagnostic probability is unclear. A meta-analysis of RTV in ADHD [35] concluded that children and adolescents with ADHD demonstrated greater RTV relative to typically developing groups, reflecting a stable feature of ADHD, but also of other clinical disorders. However, children with ADHD were only minimally more variable than the clinical control children and essentially indistinguishable from the clinical control groups, indicating that RTV lacks specificity and thus is a non-viable diagnostic marker of ADHD. In addition, not all individuals diagnosed with ADHD have executive deficits comprising RTV [40]. A systematic review with 60 studies and six commercially available CPTs, including the CCPT [16], summarised the current evidence base for the use of CPTs to support the diagnostic procedure, as well as medication management in children with ADHD. The studies showed mixed findings on the clinical utility of CPTs to aid diagnosis and assessment in clinical practice, warranting further research.

A recent Swedish retrospective study [40] of 80 children with and 38 children without ADHD compared two different CPTs, confirming this recommendation. In this study, dichotomised values of the CCPT-II confidence index (cut-off = 50) were analysed, showing incremental clinical utility in the diagnostic assessment of children with ADHD when teacher and parent ratings were inconclusive. The confidence index indicates the degree of fit to a clinical respondent’s profile, where higher confidence indexes indicate greater fit to the clinical profile, and very high values (>60%) indicate higher evidence for the clinical classification. Tallberg and colleagues found that the CCPT-II as a ‘stand-alone test’ did not demonstrate clinical utility, which they defined as a post-test probability of ADHD of greater than 85%. However, when parent and teacher ratings were inconclusive, adding the CCPT-II markedly increased (if positive) or decreased (if negative) the post-test probability.

In our study, the K-SADS-PL interview, performed by a trained independent rater who did not have access to other information was used to independently confirm the diagnosis of ADHD and to evaluate comorbidity. That said, the K-SADS-PL results were made available for the clinical team before the final diagnosis was established. We identified an agreement between the K-SADS-PL results and the final diagnosis of ADHD of 77%. Of course, the K-SADS-PL is not a completely independent source, as caregivers were informants for both the K-SADS-PL and the ADHD-RS. We found that ADHD-RS total symptom scores were clearly linked to the likeliness of receiving an ADHD diagnosis based on teachers’ ratings (p < .001), but the differences between groups in the total symptom scores rated by mothers were not significant, suggesting that additional information contributed to the final diagnosis. For the 23% of participants with a disagreement between the K-SADS-PL result and the final diagnosis (Table 5), the supplemental information from all sources, including CCPT-II results, likely influenced the final diagnosis. In our sample, 44% of the participants diagnosed with ADHD had one or more comorbid psychiatric conditions, and the corresponding percentage in the group without ADHD was 30%. The difference between groups was not significant; nevertheless, an assessment of co-occurring disorders is crucial. The presence of comorbid disorders in ADHD is common and may confound the assessment of ADHD due to overlapping or masking symptoms. Conversely, in the group without ADHD, we would expect a higher rate of other psychiatric disorders, as this can explain why they were referred for an evaluation of ADHD symptoms. In addition, interfering conditions, such as reading disorders, may present false positives on the CCPT-II [30]. Numerous psychiatric, somatic and developmental disorders are associated with ADHD [41], many of which require treatment in their own right. Interestingly, in a population-based study of children’s development and mental health, CCPT-II performance did not differentiate between three diagnostic groups, namely ADHD, ODD, and ADHD + ODD. Children with ODD (with or without comorbid ADHD) did not differ from children in the control group in any of the CCPT-II parameters [42]. Thus, we believe that an appropriate assessment of comorbidity, such as with the K-SADS-PL, may prevent diagnostic errors due to confounding comorbid symptoms. The K-SADS-PL interview is resource-intensive, it takes time, and a provider should be trained in its application. However, adding the K-SADS-PL (or an equivalent diagnostic interview) to the assessment battery gives the advantage of combining both a quality assurance of the ADHD diagnosis and a reliable comorbidity assessment, which is in line with the recommendation in the Norwegian clinical guidelines for the assessment and treatment of ADHD [15].

As the CCPT-II in its original design seem to fail in distinguishing reliably between children with ADHD and other clinical groups, several attempts to modify both its application and interpretation has been described. For example, by introducing the test to a virtual classroom [43, 44] or bringing task-unrelated movements (measured via actigraphy) into the CCPT, the test results seem to provide a far more detailed or accurate basis for interpretation [45]. Berger and Cassuto [46] run CCPT with ongoing auditive and visual disturbances, and adolescents with an ADHD diagnosis made significantly more omission errors with distractions present during the test than without distractors, while distracting stimuli did not affect the CPT performance of non-ADHD adolescents.

Strengths and Limitations

A strength of the study is that it was based on a fairly diverse sample of participants living with one or two caregivers, referred to a regular CAMHS outpatient clinic and consecutively assessed. Thus, they represented a broad and non-selected sample of children and adolescents with attention problems and impulsive and hyperactive behaviour from a distinct geographic area in the central part of Norway. Another strength was that the assessment instruments applied were standardised and well-validated [47]. In addition, confirmation (or disproval) of the diagnosis with the K-SADS-PL interview, performed by an independent rater, added another source of information and contributed to the total assessment quality. However, to perform the evaluation independently, this rater did not have access to all other information that was available for the team establishing the final diagnosis. Therefore, it was not surprising that their results and the K-SADS-PL rating were not completely overlapping.

Limitations include the fact that the data are based on the CCPT-II, which in some CAMHS has been replaced by the revised third version (CCPT-III), very much like the previous versions but including a modification to the paradigm, a new age range, new normative data, scoring changes and a reconceptualization of the assessed dimensions of attention. In addition, there are some improvements to the usability of the programme. However, no fundamental changes have been made to the assessment that may alter the conclusions in this paper. Nevertheless, the CCPT-II is still commonly in use in many CAMHS clinics throughout Norway.

Sixteen children did not meet the diagnostic criteria of any psychiatric disorder, despite of having been referred to the outpatient unit because of behaviour problems and or externalizing symptoms. This group may comprise children with time limited or situation dependent problems, but also children with subdiagnostic threshold symptoms. Still, they could have experienced an elevated symptom level, to a degree that could have influenced the results, e.g. made it more difficult for the CCPT-II to differentiate between the two groups. This could also have been a contributing factor to the non-significant group difference in mother-rated ADHD symptoms.

We had fewer patients with comorbid mood and anxiety disorders in our sample, than in a comparable population of children and adolescents with ADHD constituting a potential bias. The male preponderance with 80% of boys could be an explanation for this distribution. There are obviously other limitations to the generalisability of our results, primarily the low sample size and its homogeneity in terms of ethnicity, as all but one family were native Norwegians. The low sample size also prohibited age- and gender-specific analyses. Regardless, the gender distribution of the sample showed a male preponderance with approximately 80% boys and 20% girls; however, a ratio of four boys to one girl is frequently reported in clinical studies and may be representative of such a clinical population.

Conclusion

In our study, CCPT-II performance parameters reflected to a limited degree the symptoms of inattention, hyperactivity and impulsive behaviour, as rated by the mothers and teachers. Of all the CCPT-II parameters, a slightly higher number of commission errors which are considered as indicator of impulsivity in the group of children with ADHD was the only statistically significant difference in CCPT-II performance between children diagnosed with ADHD and the children not diagnosed with ADHD. The clinical implication is to interpret CCPT-II results with caution, because they correspond to a limited degree with other sources of information.

Acknowledgment

The authors wish to thank the young people and their parents participating in the study, as well as colleagues in the Department of Mental Health, Kristiansund Hospital.

Disclosure statement

No potential conflict of interest was reported by the author(s). BW has received royalties from book Publishers Gyldendal and Universitetsforlaget for co-authorship of books on OCD and Child and Adolescent Psychiatry. All other authors declare they have no competing interest.

Additional information

Funding

This work was supported by the Regional Resource Centre for Autism, ADHD and Tourette’s Syndrome, Norwegian University of Science and Technology, the Møre and Romsdal Hospital Trust and the Liaison Committee between the Central Norway Regional Health Authority and the Norwegian University of Science and Technology