Time of onset and/or diagnosis of ADHD in European children: a systematic review

Background Attention-Deficit/ Hyperactivity Disorder (ADHD) is one of the most common childhood neurobehavioral conditions. Symptoms related to this disorder cause a significant impairment in school tasks and in the activities of children’s daily lives; an early diagnosis and appropriate treatment could almost certainly help improve their outcomes. The current study, part of the Models Of Child Health Appraised (MOCHA) project, aims to explore the age at which children experience the onset or diagnosis of ADHD in European countries. Methods A systematic review was done examining the studies reporting the age of onset/diagnosis (AO/AD) of ADHD in European countries (28 European Member States plus 2 European Economic Area countries), published between January 1, 2010 and December 31, 2019. Of the 2276 identified studies, 44 met all the predefined criteria and were included in the review. Results The lowest mean AO in the children diagnosed with ADHD alone was 2.25 years and the highest was 7.5 years. It was 15.3 years in the children with ADHD and disruptive behaviour disorder. The mean AD ranges between 6.2 and 18.1 years. Conclusions Our findings indicate that there is a wide variability in both the AO and AD of ADHD, and a too large distance between AO and AD. Since studies in the literature suggest that an early identification of ADHD symptoms may facilitate early referral and treatment, it would be important to understand the underlying reasons behind the wide variability found. Trial registration PROSPERO registration: CRD42017070631. Supplementary Information The online version contains supplementary material available at 10.1186/s12888-021-03547-x.


Background
Attention-Deficit/Hyperactivity Disorder (ADHD), which is one of the most common childhood neurobehavioral conditions, has been characterized by continually increasing global prevalence rates over the past few decades [1]. A global consensus on the ADHD prevalence rate in children and adolescents has yet to be reached: meta-regression analyses have estimated the worldwide rate at between 5.29% [2] and 7.1% [3], but according to one comprehensive meta-analysis, the bestestimate prevalence rate of study based on case definition was 1.4%, (range: 1.1-3.1) [4].
Children with unmanaged ADHD often experience unnecessary impairments and detrimental long-term consequences leading to high personal and societal costs [8,9]. Early identification and effective management could significantly improve the functioning and overall quality of life of these children and their families. Healthcare professionals specialized in child psychology, including the American Academy of Paediatrics, advise screening for the disorder early as the preschool period [10] so that those affected can be treated precociously permitting them to achieve their full potential in school and at home [11].
Multiple factors may affect the perception of the disorder by family members and healthcare providers and thus the timing of its diagnosis and treatment [7]. Moreover, there are numerous factors intrinsic to childhood or adolescence that could affect the diagnosis of ADHD including gender, age, race, socioeconomic status, and severity of symptoms [12,13].
Parents play a central role in recognizing behavioural problems early in their children, their perception, awareness and acceptance of the disease, as their decision to accompany the child to a specialist [14]. Once parents decide to seek help, they need to be able to access specialised care for a timely and accurate diagnosis as well as optimal disease management strategies. Although there is an operationalized psychodynamic diagnostic process, no objective test is at yet available and substantial controversy exists regarding the challenge of formulating a correct diagnosis [15,16]. In fact, conflicting views continue to exist with regard to the symptoms and psychometric features leading to a diagnosis of ADHD diagnosis [17][18][19].
Many clinicians depend on and utilize the Diagnostic and Statistical Manual of Mental Disorders (DSM) [20] for guidance in making diagnoses, even if general diagnostic issues (e.g. model of diagnosis and level of impairment) need of better clarification [17] also using different diagnostic criteria, such as the International Classification of Diseases (ICD) and the Research Domain Criteria.
Although the heterogeneity in the methodology of diagnosing of ADHD has resulted in a high variability in prevalence rates around the world [21], differences linked to age at diagnosis (AD) or onset (AO) of ADHD need to be investigated.
Within the Models Of Child Health Appraised (MOCHA) project [22], which has been critically assessing the existing models of primary care for children in 30 European countries (28 European Member States plus 2 European Economic Area countries), Minicuci et al. [23] have been involved in investigating the AO and AD of ADHD.
The current work set out to examine the studies involving children with ADHD in European countries that report their age at onset or diagnosis.

Methods
Following a systematic review approach and a standardized method of Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) [24], we searched for studies that reported the AO or AD of ADHD. The review protocol was registered in the PROSPERO database (registration number CRD42017070631). The PRISMA checklist for this systematic review is presented in Additional file 1.
We searched the Medline (PubMed) database for studies in the literature examining ADHD onset or diagnosis published between January 1, 2010 and December 31, 2019. It was decided to limit the review to the last decade because it is the one in which the clinical guideline's recommendations previously produced by many parties had to be consolidated also with the fifth revision of the DSM started in 2000 and finished in 2013 [25]. The search terms used were: "ADHD", "Attention deficit", "Hyperactivity Disorder" and "Attention disorder" in the title or abstract, combined with "age", "onset" or "diagnosis" and "child" or "adolescent" in the text word (see Additional file 1 for details). Any study not in English were excluded. In order to include all studies reporting ADHD AO/AD, no exclusion criteria was applied to diagnostic criteria/tools used for participants' diagnosis in the studies reviewed. The diagnostic criteria or tools adopted in each included study, as well as the inclusion and exclusion criteria followed to select the study sample, have been examined at a later time with other relevant characteristics.
The abstracts of all the articles were read and the full version of the papers for those seemingly fulfilling the selection criteria were retrieved.
Studies were included in this review if they reported the AO or AD for ADHD and were conducted in or referred to data from a European country.
We utilized a standardized form for data extraction that included the following items: the authors' names, the year of publication, the country in which the study was performed, the journal in which the study was published, the type of study, the aim of the study, the year in which the study was performed, the types of persons composing the study sample (including age and sample size), the diagnostic criteria adopted for the diagnosis, and, of course, the AO/ADs. Two of the authors (IR and BC) screened all the articles; any differences in viewpoints that arose were resolved through discussion with the third author (NM).

Results
The initial PubMed search yielded 2276 studies (Fig. 1). After the abstracts were screened, a total of 1163 articles were excluded, mainly because the population studied and/or the geographic area (did not meet our criteria, in the former case for age, in the latter it referred to studies outside Europe). Out of the 1113 full-text articles reviewed, 49.1% were carried out outside Europe and 43.4% did not report AO or AD. Forty-four articles met our inclusion criteria for this review.

Study characteristics
The characteristics of the studies included are outlined in Table 1. Twenty-three articles were published in 2010-15 and 21 articles in 2016-19.
One article reported both the AD and the AO, 34 studies reported only the AD, and 9 reported only the AO. The majority of the studies included in this review were conducted in Sweden (7 articles) and in Germany (5 articles), followed by 3 countries publishing 4 articles each. Table 2 provides a full list of the 44 studies mentioned here, in the order of their publication date; its chronological number is also used throughout the text in all subsequent references to that article.

Diagnostic criteria
In the majority of the articles, the diagnostic criteria used to define ADHD symptoms or to formulate a diagnosis of ADHD was the DSM. One study conducted in Sweden [31] reported that the DSM criteria in the DSM-III-R [70] and in the DSM-IV [71] were used before and after 1994 respectively; while in the study of van Lieshout et al. [53] the DSM-IV and DSM-5 [72] were adopted. In eight papers the 4th edition (DSM-IV) was adopted, in five papers the "text revision" of the DSM-IV, namely the DSM-IV-TR [73], was used. The DSM-IV items of the Conners' teacher questionnaire were used with the Parental Account of Childhood Symptoms (PACS) interview in the study by Muller and colleagues [32].
The Composite International Diagnostic Interview (CIDI) version 3.0 was used to determine the presence of ADHD according to the DSM-IV criteria in the article by Tuithof and collaborators [36]. Developed by the World Health Organization, the CIDI is a fully structured, lay administered interview used worldwide that has been shown to be a reliable and valid instrument [74]. We also identified 15 papers using the ICD to define ADHD symptoms or to make an ADHD diagnosis. Among these papers, one article [47] used both the 9th [75] and 10th [76] editions; the remaining articles used the 10th edition.
In three articles, multiple sources of information were taken into consideration for the diagnosis of ADHD. The diagnostic criteria of the DSM-IV and the ICD-9/ ICD-10 were adopted and the results of Conner's questionnaire were considered in the six countries involved in the study by Hodgkins and collaborators [39]. In the article by Chen and collaborators [57], the individuals who were diagnosed with hyperkinetic disorder (ICD-9, ICD-10) or ADHD (DSM-IV) were defined as ADHD cases; in Bahmanyar et al. [38] the ICD-10 and DSM-IV were adopted.
The semistructured Schedule for Affective Disorders and Schizophrenia for School-Age Children, Present and Lifetime Version (K-SADS-PL) is a DSM-IV-based diagnostic interview procedure that was used in some of the articles to support ADHD diagnosis [41,53,59].
Two studies [65,66] used Read codes for an ADHD diagnosis. Read codes are clinical terminology developed in the UK by the National Health Service (NHS) based on clinical parameters and usage.
Read codes have become the de facto standard for coding diagnoses, operations, and procedure for all national data sets and statistics on hospital and community health services in the UK.
Lastly, the diagnostic criteria utilized were not specified in six articles. In three papers [43,45,58], parents/ caregivers of children and adolescents were asked if their children had ever been diagnosed with ADHD by a doctor or other healthcare professional; in a paper by Caci and colleagues [49] the physicians treating children with ADHD were asked to select patients to enrol in the study; finally, the patients were identified from the psychiatric cases and drug registers in two articles [48,64].

Age at onset
Eight out of 10 of the studies presenting information on the AO reported the mean, median or age range at symptom onset in the sample of children being studied [26,29,35,36,41,49,51,53]. The lowest AO was reported by a Dutch study examining a sample made up of 347 patients with combined ADHD, whose ages were between 5 and 19 years; the first ADHD symptom appeared at a mean age of 2.25 years [53]. The highest AO was reported in a study referring to children in Finland: it was 7.5 years in the children with only ADHD diagnosis and 15.3 years in the children with comorbid ADHD and disruptive behaviour disorder (DBD) [41].
The study by Muller and colleagues [32] reported the time of ADHD detection rather than the time of symptom onset which was analysed by comparing probands with combined ADHD with their siblings without ADHD diagnosis or to different subtypes of ADHD.
Polanczyk et al. [28] focused on the implications of extending the ADHD AO criterion from ages 7 to 12 years, since the variation would lead to a negligible increase in ADHD prevalence (0.1% in their cohort) by age 12.

Age at diagnosis
Thirty-two of the 35 studies presenting information on the AD of ADHD reported its mean, median, range or distribution, one study presented the peak of the ADHD incidence in males and females [63]; in two other studies, information on the AD was inferred through the          To explore the association between iodine intake from food in pregnancy (as a proxy for long-term iodine intake and 77,164 mother-child pairs were included in this study. ADHD diagnosis was registered in 1725 children (2.2%) by December 2015

ICD-10
The median AD was 8.2 years (IQR: 7.0, 9.5 years).    graphical representations of the cumulative incidence of ADHD diagnosis among siblings and cousins [57] and of the prevalence of ADHD diagnosis in an insured population [55]. The age range in the 24 studies reporting the mean AD was 6.2 to 18.1 years. The lowest mean value, which was reported by Andreou and Trott [37], concerned a group of 30 university students living in Greece who were diagnosed with ADHD during childhood (26 a combined form and 4 a hyperactive impulsive form).
Granström et al. [64] reported the highest mean AD in 26 individuals with Hirschsprung disease and 7390 controls taking any drug for the treatment of ADHD according to the Swedish Prescribed Drug Register.
Two studies reported the mean age for ADHD diagnosis for a specific group of countries [39,43]. Taking into consideration the same countries (i.e. France, Germany, Italy, the Netherlands, Spain and UK), the two studies identified Germany as the country with the lowest mean age and the UK [39] and the Netherlands [43] as the countries with the highest mean age.

Discussion
ADHD, one of the most common childhood neurodevelopmental disorders, is characterized by a pattern of inattention and/or impulsivity and hyperactivity, behaviours that can have a dramatic impact on children and on family life [77]. Since early identification of the disease is essential to optimize the quality of life of both the children themselves and their families, there is growing research interest in investigating the timing of diagnosis which can lead to prompt medical attention.
The current study set out to investigate age at the time of onset and/or diagnosis of ADHD in children living in European countries by examining the studies published between January 1, 2010 and December 31, 2019 reporting on the AO and AD of ADHD. The study's most important finding was that there is a wide variability in both.
Much of the variability could be attributed to discrepancies in study methods. Differences in study designs, ranging from case-control, cohort, to cross-sectional, could have affected the AO/AD as a cohort study could more accurately identify the AO and thus the incidence of ADHD with respect to a cross-sectional retroactive study basing its figures on parent reports.
The studies also show differences in sampling methods. Several studies, in fact, used convenience samples from clinic-based studies; others were based on registry data or medical records. It is reasonable to hypothesise that referred patient samples have lower AO/ AD compared to community cohorts given the differences in the severity of the disorder in these populations. The source of information (self-reported, parentreported, teacher-reported, doctor-reported) can also significantly influence the figures on the AO/AD registered by the different studies. In a multi-country crosssection study by Caci et al. [43] which assessed the degree to which ADHD impairs patients' everyday lives, the diagnosis was caregiver-reported and the mean AD was 7.0 years, ranging from 6.3 years in Germany to 7.6 years in the Netherlands. The diagnosis was obtained following the consultation of a mean of 2.7 doctors, ranging from 2.3 in the Netherlands to 3.2 in France, over a mean period of 20.4 months, ranging from 12.2 in Spain to 31.8 in the UK.
The articles by Genuneit et al. [45] and Pohlabeln et al. [58] presented the parent-reported AD: in the first, 44% of the children included in the sample were diagnosed before the age of 8 years and the others between 9 and 11 years; in the second article the percentage of children diagnosed between 9 and 11 years fell to 17%. Although the age range as well as the source of information (parent-reported) of these two articles was the same, the differences in AD could probably be explained by the presence of a comorbidity, that is Atopic Eczema in the case of the first sample.
The presence of comorbidities is an exceedingly important consideration when ADHD diagnosis is being discussed. ADHD symptoms can overlap with those of other disorders, including autism spectrum disorder, disorders of mood and conduct, oppositional defiant disorder, learning difficulties, impaired motor control, poor executive functions (working memory, planning, organisation, and time management), communication difficulties, sleep disorders, tics/Tourette syndrome, epilepsy, and anxiety disorders, that commonly coexist with ADHD.
Socanski et al. [42], who compared the ADs of ADHD in a group of children with epilepsy and a control group, uncovered a statistically significant different in the mean ages: the children with epilepsy had a mean AD of 8.2 years, while those without epilepsy had a mean AD of 9.4 years (p-value = 0.07). This result suggests that children with comorbidities related to ADHD have a greater probability of being diagnosed with ADHD at a younger age, presumably because they already have access to some kind of healthcare services and are being monitored by medical specialists.
Thus, the characteristics of enrolled populations (exclusion/inclusion criteria) and the sample size enormously undermine the evaluation and comparison of studies as well as a formal summary of the results. The application of meta-analysis is also impossible due to the lack of publication by all the studies of the absolute values.
The criteria used to diagnose the condition is another important factor that could affect the heterogeneity in the studies considered.
The two main diagnostic systems currently being used are the ICD-10 and the DSM-V [72]. Both systems require that symptoms be present in several settings, for example school/work, home life and leisure activities, and that the onset of symptoms be evident in early life, although this criterion has not yet received a consensus among specialists and has changed over the last decades. For the DSM-V, onset is expected to occur by the age of 12 years; for the ICD-10 and the DSM-IV by the age of 7 years. Since AO itself is one of the criteria of a diagnosis of ADHD in the studies included in our analysis, the AD falls within the ages determined by the diagnostic criteria.
Several research teams have been concerned about the implications of increasing the age of onset to 12 years [78]. Some investigators seem favourable to adopting the new criteria since it has increased the number of ADHD patients receiving help [79]. Other researchers believe that parents' inability to recall the AO prior to 7 years might give false negative results and reduce some of the diagnostic relevance connected to recalling the AO [80].
A revision of AO criteria should in any case be based on studies assessing the performance of different diagnostic criteria in the population [79].

Strengths and limitations
To our knowledge, this study represents the first systematic review of the AO and AD of ADHD in European countries. Although we did adhere to PRISMA guidelines [24] to ensure methodological rigour, the study does have a number of potential limitations.
The first is that studies not published in English as well as those not available in PubMed were not taken into consideration.
We would also like to point out that the 44 articles included in this systematic review refer to studies conducted in 13 European countries.
Despite these limitations, and those methodological of analysed studies, the review offers new insights into the timing of the onset and diagnosis of ADHD.

Conclusions
One of the key functions of primary care is to recognize the symptoms of an illness at an early stage. As far as childhood illnesses are concerned, neurodevelopmental disorders are relatively common and increasing in Europe. Early diagnosis makes it possible to contemplate and implement opportune treatment strategies thus reducing, in this case, some of ADHD's adverse current and future consequences in the child and family. This study provides a preliminary overview on the timing of the onset and the diagnosis of ADHD in children living in European countries. The long term validity and heterogeneity of the classification systems used to guide diagnoses and the factors behind the social, cultural and genetic differences affecting the timing of identification of the syndrome need further analysis. The fact that Germany has a much earlier AO and AD with respect to the UK and the Netherlands is just one example of differences that need to be clarified. Studies in the literature suggest that identifying ADHD symptoms early on can facilitate early referral and treatment, and thus limit its cost in personal and societal terms [81,82]. To optimize the quality of the service and of the care delivered is the task of both policymakers and clinical experts. To guarantee an equal standard of care for all children and adolescents with ADHD is a pressing need to reduce the times to complete the diagnostic path, and promptly star with appropriate therapy [83]. However, further studies are necessary to uncover the underlying reasons for the large variability observed in both AO and AD, and reducing the distance between the onset and the diagnosis of ADHD.
Funding This work was supported by the MOCHA project, that was funded by the European Commission through the Horizon 2020 Framework [grant agreement number: 634201] but this work was produced subsequently without funding. The funding body was not involved in the design of the study, the collection, analysis, and interpretation of data, and in writing the manuscript.

Availability of data and materials
The data used to support the findings of this study are included within the article.