Abstract
ADHD is a common childhood-onset neurodevelopmental disorder that affects the child’s psychosocial, educational, and neuropsychological functioning, self-esteem, and quality of life from childhood to adulthood (Biederman et al., 2012; Bradshaw & Kamal, 2017; Harpin, Mazzone, Raynaud, Kahle, & Hodgkins, 2016; Wehmeier, Schacht, & Barkley, 2010). Community-based studies indicated ADHD prevalence in school-age children varying from 0.2% to 27% (Polanczyk & Jensen, 2008). In a systematic meta-regression analysis, the pooled worldwide prevalence of ADHD was determined as 5.29% and as 6.48% if only children were assessed (Polanczyk, de Lima, Horta, Biederman, & Rohde, 2007). A recent review of 86 studies using the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association [APA], 1994) revealed that the prevalence of ADHD ranges from 5.9% to 7.1% (Willcutt, 2012). Studies indicated that the prevalence rates vary mainly depending on methodological differences and there is no evidence to suggest an increase in the prevalence rates of ADHD across the last three decades (Faraone, Sergeant, Gillberg, & Biederman, 2003; Gathje, Lewandowski, & Gordon, 2008; Polanczyk et al., 2007; Polanczyk, Willcutt, Salum, Kieling, & Rohde, 2014; Rohde et al., 2005).
ADHD affects boys more commonly and the male: female ratio ranges from 1:1 to 3:1 in community samples (Skounti, Philalithis, & Galanakis, 2007). Most of the children with ADHD have another comorbid psychiatric condition. Studies indicated that learning disability, oppositional defiant disorder (ODD), anxiety disorders, conduct disorder (CD), tic disorders, and affective disorders are highly comorbid with ADHD (Larson, Russ, Kahn, & Halfon, 2011; The Multimodal Treatment Study of ADHD [MTA] Cooperative Group, 1999).
Data on prevalence, comorbidities, and sociodemographic features of ADHD have a critical role in the planning of child mental health services for a country. The literature search revealed only few studies published on ADHD epidemiology in Turkey. Diagnosis of ADHD was based on parent and teacher rating scales in two studies about the prevalence of the disorder (Ersan, Dogan, Dogan, & Sumer, 2004; Gul, Tiryaki, Kultur, Topbas, & Ak, 2010). Rating scales and clinical assessment have been conducted in the other research for the diagnosis (Ercan et al., 2013). Studies on epidemiology of ADHD in Turkey revealed that prevalence rates are between 8.1% and 13.3% and displayed a male predominance (Ercan et al., 2013; Ersan et al., 2004). As distinct from most of other research, ADHD-H was more common than ADHD-I in the Turkish population (Ersan et al., 2004; Gul et al., 2010). In the unique study that assessed comorbid disorders in ADHD, Ercan et al. (2013) revealed that comorbid disorders were common in primary school children.
The aims of this study were (a) to describe the prevalence of ADHD in Turkish school-age children from the general population using a two-stage epidemiological study design, (b) to examine comorbid psychiatric disorders in children with ADHD, and (c) to evaluate sociodemographic features related to ADHD and comorbid conditions.
Method
Sampling Process
This study was performed in Denizli, a city located in western Turkey with a population of 942,278 according to the Turkish Statistical Institute data of 2011 (Turkish Statistical Institute, 2014). The sampling universe included 67,149 pupils aged 6 to 14 years in a total of 106 primary schools in the provincial center of Denizli. The sample size needed for this study was determined to be 2,105 pupils according to the estimated average prevalence of ADHD of 6.5 ± 1% (Polanczyk et al., 2007), using the computer package EPI INFO Version 6.4. We used a multistage sampling technique to obtain a representative sample. In the first stage, 421 classrooms comprising first to eighth grade were selected out of 2,055 using a stratified random sampling technique. In the second stage, five pupils per class were selected using a random sampling technique according to the position in their class register. This procedure gave a total sample of 2,105 pupils.
Assessment Procedures
Both the Medical Ethics Committee of the Pamukkale University Faculty of Medicine and the Ministry of National Education Denizli Provincial Directorate approved the study protocol.
A two-stage design was followed for the study. Screening phase was performed during the period from February 2011 to May 2011 to provide enough time for each main teacher to observe the pupil’s behavior. The field investigators appraised all school coordinators and main teachers of selected classes about the purpose and design of the study. An information letter defining the aims and procedure of the study and inviting them to participate was first sent to the parents through their children’s teachers. The letter also indicated that they would be invited for an interview at the clinic if they agreed to participate in the study. Together with the information letter, (a) an informed consent form for parents, (b) a questionnaire including sociodemographic variables for parents, and (c) screening scales for both parents and teachers were also sent. Pupils who had a signed informed consent, a completed questionnaire, and two screening scales were included in the screening phase.
First phase: Screening procedure
In the screening stage, we obtained data on ADHD symptoms from parents and teachers using the Turgay DSM-IV Disruptive Behavior Disorders Rating Scale (T-DSM-IV-S). The T-DSM-IV-S was developed by Turgay (1994) and translated and adapted into Turkish by Ercan, Amado, Somer, and Cikoglu (2001). The T-DSM-IV-S is based on the DSM-IV Attention Deficit and Disruptive Behavior Disorders diagnostic criteria and assesses inattention (9 items), hyperactivity impulsivity (9 items), opposition-defiance (8 items), and conduct disorder (15 items). Parents and teachers rate each symptom on frequency (namely, 0 = not at all, 1 = just a little, 2 = much, 3 = very much). A rating of “much” or “very much” is considered as positive for each item. In this study, children with at least six of the nine positive inattentive and/or hyperactive-impulsive items in either teacher or parent scales were considered as “screen-positive.”
Second phase: Diagnostic procedure
All screen-positive children and their parents were invited to participate in the second stage of the study to detect ADHD and comorbid psychiatric disorders. Parents were contacted by phone to schedule the clinical interview. The diagnostic assessment of participants was performed by using the Schedule for Affective Disorders and Schizophrenia for School-Age Children–Present and Lifetime version (K-SADS-PL; Kaufman et al., 1997). The K-SADS-PL is a semi-structured diagnostic interview schedule based on DSM-IV criteria, which investigates a wide range of psychiatric disorders. For ADHD diagnosis, in addition to DSM-IV “A” criteria, K-SADS-PL also examines “duration,” “age of onset,” and “functional impairment” criteria. The reliability and validity of the Turkish version of K-SADS-PL have been established by Gokler et al. (2004). To avoid an interrater bias, the same experienced child and adolescent psychiatrist certified in the use of the K-SADS-PL (G.U.) carried out all assessments. Both parents and children were interviewed individually in the same interview schedule. In case of a discrepancy between parent and teacher rating scales, a telephone survey was conducted by the main teacher. Presence of some impairment from the symptoms in two or more settings was required for ADHD diagnosis. To assess the impairment criteria, we gathered information about child’s social and school functioning from parents. Disorders not covered by K-SADS-PL (e.g., learning disorders, pervasive developmental disorders, mental retardation) were evaluated based on DSM-IV criteria.
Data Analysis
Data were analyzed using the Statistical Package for the Social Sciences 11 (SPSS.11) program. Continuous variables were presented as M ± SD or median with an interquartile range (IQR), and categorical variables were described as numbers and percentages. All continuous variables were tested for normality and homogeneity of variance. As continuous variables were not normally disturbed, Mann–Whitney U test was carried out for comparisons. Pearson’s chi-square analysis was performed to test for differences in the proportion of categorical variables. p values less than .05 were accepted to be statistically significant.
Results
T-DSM-IV-S and sociodemographic questionnaires were sent to the parents and teachers of the 2,105 randomly selected cases (Figure 1). Data were analyzed only if both the parent and teacher scales and sociodemographic questionnaires were correctly completed and written informed consent was signed. Of the total 1,718 returned (81.6%), we excluded 149 whose parents refused to participate and 61 who had incomplete data. Parent and teacher scales of 1,508 cases were returned and considered valid. The sample consisted of 700 boys (46.4%) and 808 girls (53.6%) ranging in age from 6 to 14 years (10.6 ± 2.2 years). As none of the private schools participated in the study, all participants were public school pupils.

Flow diagram of study.
Of these 1,508 children, 210 (13.9%) were classified as “screen-positive” as they had at least six of nine DSM-IV inattentive and/or hyperactive-impulsive symptoms according to the parent or teacher T-DSM-IV-S questionnaire (Figure 1). One hundred twenty-seven (18.1%) boys and eighty-three (10.3%) girls were indicated as having ADHD symptoms. There was a statistically significant gender difference (p < .001).
Among the 210 screen-positive children, 80 (38.1%) had only an ADHD-positive parental report, whereas 73 (34.8%) had only an ADHD-positive teacher report. For 57 (27.1%) children, both parent and teacher scales indicated ADHD symptoms.
All screen-positive children and their parents were invited to participate in clinical assessment. The clinical assessment of 69 children (32.9% of those invited) could not be completed as their parents refused to participate in or did not attend the scheduled assessment interviews three times (Figure 1). Gender and family characteristics of children who did not participate (n = 69) in clinical assessment were not different from those who were evaluated with K-SADS-PL (n = 141). The mean ages of participants and non-participants were 10.5 ± 2.0 and 11.3 ± 2.2, respectively (p = .01). Data concerning these 69 non-participants were excluded from the subsequent statistical analyses due to their suspicious diagnostic status.
The final sample included 776 (53.9%) girls and 663 (46.1%) boys with an age range of 6 to 14 (10.6 ± 2.2). Most of the pupils (96.5%) were living in a nuclear family with a mean size of 4.4 ± 1.0 (2-11) persons. Nearly half of the parents (49.6%) reported that they had a middle household income. More than half of the fathers (56.5%) and mothers (68.7%) had an education below a high school degree.
According to clinical assessment, 115 participants met DSM-IV criteria for ADHD, corresponding to a prevalence of 8% (95% confidence interval [CI] = [6.6, 9.4]). The non-ADHD group consisted of 1,298 screen-negative children and 26 children who had been defined as screen positive first but were later determined as not having ADHD upon clinical evaluation (total N = 1,324). Table 1 shows the sociodemographic characteristics of the children with and without ADHD. The analysis of the demographic data revealed that the prevalence was significantly higher in boys than in girls (odds ratio [OR] = 2.08; 95% CI = [1.40, 3.08]). The prevalence was 10.9% (N = 72; 95% CI = [8.5, 13.5]) in boys and 5.5% (N = 43; 95% CI = [3.9, 7.1]) in girls.
Sociodemographic Characteristics of Participants.
Note. IQR = interquartile range.
As shown in Table 1, comparison of ADHD and non-ADHD children revealed no differences in terms of economic status and number of household members. Children from an extended family in which a child lives with his or her parent(s) and grandparents or uncles/aunts had significantly higher prevalence rates compared with children from intact nuclear family and single-parent family (p < .001). The highest rate was found in children with non-educated or primary education-level (primary + secondary schools) mothers as compared with those with high school or university-graduate mothers (8.5%, 7.6%, and 5.5%, respectively). Similar results were obtained by analyzing the educational level of fathers: ADHD was more common in children with non-educated or primary education-level fathers, while it was relatively rare in those with high school or university-graduate fathers (9.0%, 7.7%, and 5.6%, respectively). Nevertheless, the differences were not significant (p > .05).
The most frequent type in our sample was the ADHD-combined type (ADHD-C). The prevalence of combined type was 4.7% (N = 68; 95% CI = [3.6, 5.7]), the prevalence of inattentive type was 2.4% (N = 35, 95% CI = [1.6, 3.1]), and that of hyperactive-impulsive type was 0.8% (N = 12, 95% CI = [0.3, 1.2]). There were no significant group differences among the three types concerning child and family characteristics (all p > .05).
Although K-SADS-PL evaluates both current and lifetime disorders, we assessed particularly current comorbid diagnoses in this study. The overall prevalence rate of psychiatric comorbidity was 60.0% (N = 69; 95% CI = [51.0, 68.9]). Moreover, of these 69 children, 18 (15.7%) had two and four (3.5%) had three comorbid disorders. There were no significant differences in comorbidities between boys and girls. Learning disorders and ODD were found in comorbidity most frequently. The frequency of learning disorders was higher in combined and inattentive types as compared with hyperactive-impulsive type (χ2 = 10.69, p = .005). ODD was more common in combined type than the other types (χ2 = 15.45, p < .001). All comorbid disorders according to types are listed in Table 2.
Comorbid Disorders According to Types.
Note. ADHD-C = ADHD-combined type; ADHD-I = ADHD-inattentive type; ADHD-H = hyperactive-impulsive type.
22 of children had more than one comorbid disorder.
Twenty-six children were not diagnosed with ADHD, although their parents or teachers indicated ADHD symptoms. The clinician concluded that the child’s symptoms were better explained by another psychiatric disorder (such as anxiety disorders, learning disorders, and mental retardation) in 15 cases. None of the remaining 11 children had any psychiatric disorders.
Discussion
The aim of the present study was to explore the prevalence of ADHD and comorbid disorders in a school sample of children in Denizli, Turkey. The results revealed that the prevalence of ADHD is 8% among children aged 6 to 14 years. Our results also indicated that 60% of children with ADHD (N = 66) had one or more and 19.2% (N = 22) at least two comorbid diagnoses.
The prevalence of 8% found in this study is higher than ADHD worldwide pooled prevalence (5.29%) and that indicated by the DSM-IV (3%-5%; APA, 1994; Polanczyk et al., 2007). However, our finding is comparable with results from several previous studies worldwide (Polanczyk et al., 2007; Polanczyk & Rohde, 2007; Scahill & Schwab-Stone, 2000; Willcutt, 2012).
Previous epidemiological studies in Turkish children have demonstrated prevalence rates of ADHD within the 8.1%-13.3% range (Ercan et al., 2013; Ersan et al., 2004; Gul et al., 2010). The prevalence of 8% found in this study is quite similar to that found in two of the three previous studies from Turkey. Ersan et al. (2004) created a questionnaire based on the DSM-IV to survey 1,425 children aged 6 to 15 years, and according to parent or teacher reports, 8.1% of children were found to have ADHD. Gul et al. (2010) found that the prevalence of ADHD was 8.6% among children between 6 and 12 years by parent and teacher scales. In both studies, researchers applied only rating scales without a diagnostic assessment for ADHD. Indeed, these rates are in line with those we found by clinical assessment (8%). A recent 4-year longitudinal study in Turkey has estimated a higher prevalence of ADHD. The prevalence rates in the four waves were 13.3%, 12.5%, 12.2%, and 12.9%, respectively, among children 8 to 10 years of age (Ercan et al., 2013). A main difference between the design of this study and ours is that Ercan et al. (2013) assessed a younger population.
Data from epidemiological studies using DSM criteria in child and adolescent population propose that the ADHD prevalence rates do not vary significantly across cultures (Faraone et al., 2003; Willcutt, 2012). As mentioned in the introduction, methodological differences of ADHD prevalence studies such as demographic characteristics of the population, information source, and diagnostic criteria used to define the disorder have a significant impact on the highly variable rates (Faraone et al., 2003; Gathje et al., 2008; Polanczyk et al., 2007; Polanczyk & Jensen, 2008; Polanczyk et al., 2014; Rohde et al., 2005). Parent and teacher rating scales had been administered to determine ADHD prevalence in most studies. Although rating scales are reported as reliable in assessing ADHD, use of structured or semi-structured interviews is recommended for an accurate diagnosis of ADHD (Pelham, Fabiano, & Massetti, 2005). Both rating scales and a semi-structured diagnostic interview were performed in this study, and our results revealed that approximately 1:5 of children whose parents and/or teachers described ADHD symptoms did not meet the DSM-IV diagnostic criteria for ADHD.
The results are consistent with those of previous studies demonstrating a gender difference (Bradshaw & Kamal, 2017; Faraone et al., 2003; Polanczyk et al., 2007; Polanczyk & Jensen, 2008; Willcutt, 2012). The difference between male and female prevalence of ADHD is a well-known fact, and community-based studies have reported a male/female ratio ranging from 1:1 to 3:1 (Skounti et al., 2007). In accordance with previous results, the male to female prevalence ratio was 1.67:1 in our study.
We found some differences in family characteristics between children with and without ADHD. Diagnosis of ADHD was more common in children whose parents had a lower educational degree, but this finding was not significant. Although there was no difference in the number of household members between ADHD and non-ADHD groups, living in an extended family was associated with ADHD. There was an extremely high rate of ADHD (46.4%) among children from an extended family. In spite of the fact that ADHD is a highly heritable disorder, psychosocial adversities may play a role in triggering and maintaining of disorder (Spencer, Biederman, & Mick, 2007). Kim et al. (2009) indicated that a primary caregiver except a biological parent and changes in caregiver were associated with ADHD in a Korean community sample. A study from the United Kingdom also revealed that children with extended early grandparent care exhibited higher levels of hyperactivity and problems in peer relations by age 4 (Fergusson, Maughan, & Golding, 2008). In traditional extended families, where the obligation of childcare is shared among several adults, the authority of parents and family rules may be disrupted. As the study was conducted in only an urban area and small number of pupils in the extended family group, our results might be regarded as spurious. Further studies from different cultures are warranted to determine whether family structure has a role in ADHD.
The prevalence rates of ADHD types have varied between studies. In a recent meta-analytic review, ADHD-I was found to be the most common type based on symptom criteria alone and full DSM-IV criteria. However, the rate of ADHD-C has been detected higher than the rate of ADHD-inattentive type (ADHD-I) when the best estimate procedure was applied (Willcutt, 2012). In our study, we found the rate of the combined type was higher than the other types. The hyperactive-impulsive type was the least common type, which differs from other Turkish samples (Ersan et al., 2004; Gul et al., 2010). One explanation is that previous studies had used only parent and teacher reports without a clinical assessment. Hyperactive-impulsive symptoms of children may be distinguished more easily by parents and teachers. Conversely, identification of inattentive symptoms—especially by parents—may be difficult, and therefore, the problems of these children might remain undetected with simple screening. There were no differences among three ADHD types in terms of sociodemographic features. This finding might be a consequence of the small sample size of subtypes.
In accordance with previous findings (Gillberg et al., 2004), comorbid psychiatric disorders were highly prevalent in our sample. The majority (60%) of children with ADHD had at least one additional impairing disorder. The most frequent comorbid disorders were identified to be learning disorders and ODD, and their distribution according to types was consistent with the literature. Learning disorders are known to frequently accompany ADHD, and the risk is higher especially in the inattentive type (Baeyens, Roeyers, & Walle, 2006; Gillberg et al., 2004). ODD, too, is one of those disorders that are found in comorbidity with ADHD most frequently, and many studies demonstrate that it occurs more often in combined and hyperactive-impulsive types (Adewuya & Famuyiwa, 2007; Byun et al., 2006; Elia, Ambrosini, & Berrettini, 2008; Huh et al., 2011; Takahashi et al., 2007). As comorbid disorders are known to be related with both persistence into adulthood and poorer outcome (Biederman, Petty, Clarke, Lomedico, & Faraone, 2011; Cherkasova, Sulla, Dalena, Pondé, & Hechtman, 2013; Lara et al., 2009; Spencer, 2006; Spencer et al., 2007), early identification and more comprehensive intervention for children with comorbidities are worthwhile.
Limitations and Strengths
Our findings should be understood in the context of some methodological limitations. First, the sample we examined is representative of an urban area of a single city, and the results may not be generalized to rural areas and other parts of Turkey. Second, only public schools participated in the study. Consequently, children from the highest socioeconomic group may be underrepresented. Third, diagnostic interviews were not conducted with screen-negative children. Also, children’s medication status was not examined in the screening phase of study. Children who were under ADHD treatment may have been overlooked due to a positive medication response. So, the prevalence of ADHD may have been underestimated. Finally, as the K-SADS-PL does not cover several disorders such as learning disorders, mental retardation, developmental coordination disorder, and pervasive developmental disorders, the DSM-IV diagnoses of these disorders were made based on clinical assessment and intelligence tests, if required.
There is a dearth of epidemiological studies concerned with childhood ADHD in Turkey. In the majority of previous studies conducted both in Turkey and worldwide, only parent and/or teacher reports were considered for ADHD diagnosis. In addition, some research did not assess age of onset, duration of symptoms, and impairment criteria. In our study, the initial assessment via rating scales was administered to both parents and teachers. To avoid false-negative results, all children with ADHD symptoms as described by either parents or teachers were invited to diagnostic assessment phase. To determine the ADHD diagnosis, the K-SADS-PL was administered to both children and parents, and when required, a telephone survey was conducted by the main teacher to investigate pervasiveness of symptoms across settings. The age of onset, duration of symptoms, and impairment criteria of DSM-IV were also taken into account. Cases better explained by other conditions were excluded through a detailed clinical assessment. Despite the limitations mentioned above, these methodological differences provide relative strength to our work.
Conclusion
Besides its limitations, our study has provided some support to the existing data on ADHD epidemiology. Our results suggest that ADHD is a prevalent disorder, and psychiatric comorbidities of ADHD are common in Turkish school-age children. Careful assessment and management of ADHD and its comorbidities could improve the functioning of children with ADHD. Considering the high prevalence rates and long-term consequences of ADHD, it seems important to develop appropriate health policies for these children.
Footnotes
Acknowledgements
Melike C. B. Sengul, MD, provided thoughtful comments, which improved the manuscript. The authors would like to thank the teachers, parents, and the children who participated in this study for their cooperation.
Authors’ Contributions
Study concept and design: Adil Zorlu, Gulsen Unlu, Mehmet Zencir. Acquisition of data: Adil Zorlu, Gulsen Unlu, Ahmet Buber, Yetis Isildar. Analysis and interpretation of data: Adil Zorlu, Gulsen Unlu, Burcu Cakaloz, Mehmet Zencir. Drafting of the manuscript: Gulsen Unlu. Critical revision: Adil Zorlu, Gulsen Unlu, Burcu Cakaloz, Mehmet Zencir, Ahmet Buber, Yetis Isildar.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Pamukkale University Scientific Research Projects Coordination Unit (Grant 2011TPF001; primary investigator: Gulsen Unlu, MD).
