Autism spectrum disorder (ASD) is a neurodevelopmental condition presenting in the early developmental period, more often in males than in females (Talantseva et al., 2023; Loomes et al., 2017). According to the latest edition of the Diagnostic Statistical Manual of Mental Disorders (DSM-5) by the American Psychiatric Association (2013), ASD is characterized by restricted, repetitive patterns of behavior, interests, or activities, and deficits in social communication and interaction across multiple contexts (American Psychiatric Association 2013). Difficulties in nonverbal communication behaviors such as gestures can predict the verbal/linguistic level of children with ASD (Kilili-Lesta et al., 2022). The International Classification of Diseases (ICD-11) by the World Health Organization (WHO) describes a full range of intellectual functioning and language abilities for persons with ASD (WHO, 2019). The severity level according to the DSM-5 is based on the intensity of needed support: Level 1 requires mild support, level 2 moderate support, and level 3 intense support (Seretopoulos et al., 2019; Mehling & Tassé, 2016). ASD diagnoses are conferred by specialists, including child psychiatrists, neurologists, and developmental pediatricians, through clinical observations and adherence to specific diagnostic criteria (Seretopoulos et al., 2019; Centers for Disease Control and Prevention, 2023; American Psychological Association, 2021).

There is variability in methodology between the studies assessing ASD prevalence. Some utilize screening tools followed by full assessments for those at risk for ASD. In contrast, others identify the persons already diagnosed with ASD within a population through parent/school report, surveys, or examining administrative/medical/school records/databases. The prevalence of ASD has been increasing in the past decades (Solmi et al., 2022), and systematic reviews showed that ASD presented globally with a 0.6–1.28% cumulative mean/median in children (Talantseva et al., 2023; Elsabbagh et al., 2012; Tsai, 2014; Adak & Halder, 2017; Salari et al., 2022; Zeidan et al., 2022). In 2014 in the United States (US), 1.7% of children aged 8 years old held the diagnosis based on the Center for Disease Control data from 11 sites/states (Baio et al., 2018), which increased to 2.3% in 2018 (Maenner et al., 2021) and to 2.8% in 2020 (Maenner, 2023). These studies measured prevalence by reviewing medical and special educational records for children receiving special education services in public schools. In Europe, ASD prevalence ranged from 0.6 to 3.1% for children 6 to 11 years of age (Talantseva et al., 2023; Narzisi et al., 2018; Chiarotti & Venerosi, 2020; Fuentes et al., 2020). In Greece, the prevalence was calculated at 1.2% for children 10–11 years of age, using administrative data (Thomaidis et al., 2020). A large-scale study in England used data from the English national public school student database to determine a 1.8% prevalence for ASD among 7,047,238 students aged 2 to 21 years (Roman-Urrestarazu et al., 2021).

Systematic reviews on ASD prevalence revealed a mean/median male-to-female (m:f) gender ratio for ASD of 4.2–4.3 (Talantseva et al., 2023; Elsabbagh et al., 2012). It ranged from 1.1 to 16.0 globally (Tsai, 2014; Adak & Halder, 2017), 1.33 to 16.0 for Europe, and 2.2 to 6.7 for the US (Elsabbagh et al., 2012). A systematic review examining the ASD gender ratio reported it at 4.1 globally, which dropped to 3.2 for high-quality studies (Loomes et al., 2017). The large-scale English, total student population study, reported a 4.3 gender ratio (Roman-Urrestarazu et al., 2021). Children with ASD primarily could additionally get diagnosed with intellectual disability (ID) (Seretopoulos et al., 2019; Zeidan et al., 2022), attention-deficit/hyperactivity disorder (ADHD), learning disability, sensory processing disorders, or epilepsy, among other conditions (Bougeard et al., 2021; Casanova et al., 2020; Doshi-Velez et al., 2014), which might also change with age (Supekar et al., 2017). Approximately 30–40% of children with ASD were estimated to be nonverbal/ minimally verbal (ASD-NV/MV) (Maltman et al., 2020), defined as having a limited communicative verbal output, using fewer than 20 words during a 10- to 20-min language sample (Koegel et al., 2020; Kasari et al., 2013), or identified at the preverbal/first words developmental language phase/level (Tager-Flusberg & Kasari, 2013).

In the Republic of Cyprus, the setting for the current study, most of the population speaks the Cypriot Greek dialect, and screening tools for ASD translated into Greek could be utilized since it is one of the official languages. Some examples include the Social Communication Questionnaire (Karaminis & Stavrakaki, 2022) and the Autism Spectrum Screening Questionnaire (Alexopoulou et al., 2018). However, even after screening, there are a limited number of trained professionals like child psychiatrists, child neurologists, and developmental pediatricians who could provide an official ASD diagnosis to a child. The common tools utilized include the Autism Diagnostic Observation Schedule-Generic (Lord et al., 1999) or the Autism Diagnostic Inventory-Revised (Lord et al., 1994). For children already diagnosed, many are included in the special education database and school records. This database, prepared by the Ministry of Education, Sport, and Youth (MOESY), is shared confidentially and separately to each public school at the beginning of the school year for review and correction, based on school records. Unfortunately, no official reports were published by the MOESY with statistics of diagnoses in the past years. The ASD diagnoses are recorded in the children’s educational record for those who provided the required medical reports/certificates to refer them for a full evaluation (van Kessel et al., 2020), through the District Committee for Special Education and Training (DCSET). The Cyprus Department of Special Education Services in the MOESY keeps the special education database of all children with established special needs, receiving free services provided by the state, such as speech-language therapy, special education, and support by an aide, per school year, as defined by the 113(I)/1999 to 2020 Special Education Laws and the relevant Regulations (Cyprus MOESY, n.d.). The database shared with each public school includes information on the name, date of birth, school, grade, main diagnosis, comorbidities, and support services received. Although the special education database is currently the most comprehensive resource to record existing ASD cases in Cyprus, the information contained for each child is limited and lacks important characteristics that can be obtained from the school records for each child and school statistics. In private schools, the diagnosis was reported to the school by the parents.

Children with ASD in the Cyprus public general education system could be placed either in the general classroom (with/without an aide), or in a special unit, which is a designated classroom integrated into a mainstream school, manned with a special education teacher and necessary school aides. Children with more severe symptoms of ASD could be placed in special schools, and they had access to additional free therapy services, like occupational therapy, music therapy, physiotherapy, special gymnastics, and hydrotherapy, as needed. With these provisions provided by the MOESY in Cyprus, it was thus expected that children with mild symptoms be placed in mainstream classrooms, children with moderate symptoms in special units, and children with more severe symptoms in special schools, based on individual merit decisions made by the DCSET. Children with ASD with limited communication skills can be provided with augmentative/alternative communication (AAC) devices as needed, only for those referred to the DCSET for an evaluation by a specialist (Cyprus MOESY, n.d.).

To date, no published studies reported the prevalence, gender ratio, and characteristics of ASD in the Republic of Cyprus. Also, there was a lack of public policy for reporting ASD cases and no central database to help track prevalence. Recently, on April 10, 2024, the National Strategy for Autism (NSA) was announced in a press release (Republic of Cyprus & Deputy Ministry of Social Welfare, 2024), aiming to monitor incidence and prevalence between the years 2024 and 2028. No actions have been implemented to date. After completing a thorough literature search in PubMed, SCOPUS, and Google Scholar, this is the first investigation aimed at gathering data on existing childhood ASD prevalence and characteristics in Cyprus. Our primary aim was to establish ASD prevalence, gender ratio, and key characteristics among children in Cyprus. To achieve this, we employed school reports leveraging the most comprehensive and reliable database at our disposal, the special education database with supplementation from school records and statistics.

Methods

Participants

The current study utilized a cross-sectional study design, following the Strengthening the Reporting of Observational Studies in Epidemiology (von Elm et al., 2014). The referent population consisted of children attending kindergarten/pre-primary and elementary/primary level schools in the five districts under the control of the Republic of Cyprus (Nicosia, Limassol, Larnaca, Paphos, and Ammochostos). The population frame included all children between the ages of 5 and 12 years, enrolled in public/private schools where Greek was the primary language of instruction during the 2022–2023 school year.

One-level convenience cluster sampling was employed, with each school representing a cluster including all its enrolled children. The school sample comprised all participating schools and the student sample comprised all children from each school, within the age range provided. According to data from the MOESY for the 2022–2023 school year, an estimated 64,990 children aged 5–12 years were enrolled in 764 schools across districts, including both the public and private sectors (Cyprus MOESY & Department of Primary Education, n.d.).

A power analysis was utilized for the study. The estimated sample size required for the 1.9% (middle-point) prevalence found for Europe (0.6–3.13%) across studies (Talantseva et al., 2023; Narzisi et al., 2018; Chiarotti & Venerosi, 2020) was 7163 children with a ± 0.5% margin of error in estimating the prevalence with 95% confidence interval (CI, 0.5–1.5%) and considering a 60% potential school non-response. A high non-response rate was expected because the data was collected close to the end of the school year when schools were busy with end-of-the-year procedures. In this study, the 9990 children that participated in the sample were greater than the required estimate. To determine whether schools that had at least one child with ASD enrolled were more likely to participate in the study due to their sensitivity to the subject, a chi-squared test was used to check for selection bias. The total number of schools which reported one/more children with ASD enrolled was compared to the number of schools which reported none. No selection bias was found (p=0.41); thus, schools with children with ASD enrolled were equally likely to participate in the study as were those without.

Procedure

Data collection occurred between February 20 and June 23, 2023, using a new online anonymous questionnaire. Schools were invited to participate in the study through email, social forums, and social media platforms. Schools were given instructions to allow a person entitled to access to school records, statistics, and the special education database by the MOESY, namely a principal, special educator, or speech-language pathologist (SLP), to prepare all the anonymous information for the questionnaire, and subsequently participate in the study on behalf of their school. The MOESY permitted access to these sources anonymously and researchers provided specific instructions to utilize them to ensure data validity indirectly. The schools with access to the special education database had to report additional anonymous information for all the students identified with an ASD diagnosis if any.

Measures

The school questionnaire included 50 closed-ended and multiple-choice questions in Greek. It was developed by our research team, based on a literature review and knowledge of the local educational system (Cyprus MOESY & Department of Primary Education, n.d.; US Department of Health and Human Services, n.d.; Zablotsky & Black, 2020; Kogan et al., 2009). Questions were developed to collect necessary information anonymously to fulfill the study objectives. To ensure face validity, the questionnaire was piloted by 15 special educators/SLPs to confirm the clarity and application of all questions, and the pilot sample was not included in the study. The questions inquired about summation data from each school rather than individual data for each student to ensure anonymity and efficiency in answering the questions. Therefore, each school represented a row of data in the database, with each column representing each variable studied.

The questionnaire had four sections: (1) school information (e.g., school system, level, type), (2) overall children information (e.g., number enrolled, gender, age groups), (3) ASD children information (e.g., number of children diagnosed with ASD, comorbidities, grade), and (4) linguistic/communication status information (e.g., bi/multilingualism, ASD-NV/MV, AAC). In the last section, schools had to report how many children with ASD had exposure to other languages besides Greek/Cypriot Greek and provide the different languages found within their school statistics. Specifically, the SLP of each school was asked to report how many of the children with ASD in their school could be considered NV/MV based on the definition provided within the questionnaire (nonverbal, using fewer than 20 words, not combining words yet) according to their best judgment, after collecting a short 10-min language sample. The validity of their responses was not independently confirmed due to a lack of parent report tools in evaluating this subgroup of children, which has severe difficulty completing linguistic tests due to their poor pragmatic, comprehension, verbal, and nonverbal skills.

Data Analyses

The Cronbach’s alpha coefficient was found to be 0.49 for Section 1 (school information) due to the low number of questions regarding unrelated demographic school characteristics, 0.91 for Section 2 (overall student information), 0.96 for Section 3 (ASD student information), and 0.59 for Section 4 (linguistic/communication information), which had few questions. Due to the limited/unrelated questions for Section 1 and Section 4, the McDonald’s omega coefficient was measured at 0.64 for Section 1, 0.98 for Section 2 and Section 3, and 0.74 for Section 4.

Statistical Analysis

Categorical characteristics were shown as absolute (n) and relative (%) frequencies. Skewed distributions of continuous measures (e.g., overall children, children with ASD per school) were displayed via Wilk normality test and QQ plots. Prevalence was calculated as the proportion of children with ASD among the overall children, presented as a percentage with 95% CIs. Age was classified into four age groups: 5 to 6 years, 7 to 8 years, 9 to 10 years, and 11 to 12 years. All data from each participating school site were used. The data was weighted to ensure equal representation of both school systems (general/special education system) and reviewed for accuracy, and missing values were noted.

ASD diagnosis was treated as a binary dependent variable (yes/no), while school and student characteristics were categorized as independent variables. Prevalence and proportion CIs were calculated at 95% using the Wilson and Fisher’s exact score method. The Pearson chi-square tests were used to compare proportions and to examine associations between categorical variables. Odds ratios (ORs) were calculated to compare the two subgroups (non-ASD, ASD). Logistic regression analysis was conducted on all variables to investigate the potential associations of demographic factors (gender, age group) and school-related factors (system, level, type, area, and district) on ASD diagnosis. The factors significantly associated with an ASD status were included simultaneously as independent variables in a multiple logistic regression model, to adjust for potential moderators/confounders (e.g., gender, age group, district). p-values less than 0.05 for two-sided statistical tests and 95% CIs not including 1.0 for prevalence ratios were considered statistically significant. The R-Studio platform (RStudio Team, 2020) was used for data analysis.

Results

School Sample Information

A total of 117 schools completed the online questionnaire, with a sample size of 10,033 children aged 5 to 12 years. Most schools (97.4%) were general education/mainstream schools (15.1% of total mainstream schools), while few (2.6%) were special schools (30.0% of total special schools). We applied weights to ensure equal representation of the school systems, resulting in a final weighted sample of 9990 children (15.4% of the total population aged 5 to 12 years). Supplementary Tables 1 and 2 provide detailed characteristics and representation of the participating schools.

More than half of schools (51.3%) were kindergarten/pre-primary schools, representing only 21.1% of the children sample, because of their smaller enrollment size, as they were not allowed to report the numbers of children younger than age 5 (pre-school). About 46.2% were elementary/primary schools representing the majority (78.5%) of children. Regarding the school type, most (90.6%) were public schools, representing 17.4% of all public schools, while a few (9.4%) were private, representing only 7.1% of private schools. Among participating schools, 46.2% were urban, while 53.8% were rural. Only 10.3% of the schools had a special unit (Supplementary Table 2). A median of 54 (23–123 IQR) children were enrolled per school, comprising 30 (14–62 IQR) males and 28 (11–60 IQR) females, with one (0.0–2.0 IQR) child with ASD. The median ASD prevalence per school was 0.7% (0.0–2.7 IQR) overall and in the general education system, regardless of whether the school had a special unit or not, and 42.9% (21.4–48.7 IQR) in the special education school system (Supplementary Table 3).

Child Sample Information

The distribution of the overall children across the variables in the sample is displayed in Table 1.

Table 1 Weighted ASD distribution and prevalence by school/child characteristics
Full size table

The majority (99.6%) of children attended general/mainstream education schools, while a few (0.4%) attended special schools. In mainstream schools, most (78.5%) attended elementary/primary schools, while 21.1% attended kindergarten/pre-primary schools. About 9830 children (98.4%) were enrolled in public and only 160 (1.6%) in private schools. The regional distribution of children in the sample per school district was as follows: 38.7% in Nicosia, 21.7% in Limassol, 14.9% in Larnaca, 16.2% in Paphos, and 8.5% in Ammochostos. We found that 51.7% attended urban and 48.3% rural schools. Of the participants, 51.7% were male, while 48.3% were female, resulting in a 1.1 gender ratio. Age group was reported only for a sub-sample of 5479 (54.8%) children, and the distribution into four age groups was as follows: 51.2% aged 5–6 years, 17.5% 7–8 years, 19.3% 9–10 years, and 12.0% 11–12 years.

ASD Group Information

A total of 178 children were reported as already diagnosed with ASD by a specialist, resulting in an overall point childhood ASD prevalence of 1.8% (1.5–2.1%) in the 2022–2023 school year. In mainstream schools, this was slightly lower at 1.6% (1.4–1.9%), while in special schools, it was significantly higher (p<0.001) at 41.9% (28.4–56.7%). Regarding their school level, 31.5% of all ASD children attended pre-primary schools, with a 2.7% (2.1–3.4%) prevalence, and 58.4% attended primary schools with a significantly lower prevalence at 1.3% (1.1–1.6%, p<0.001). The majority (97.2%) of children with ASD attended public schools with a 1.8% (1.5–2.0%) prevalence, only a few (2.8%) attended private schools, and all pre-primary schools, with a 3.1% (1.3–7.1%) prevalence.

Regarding the ASD distribution per district, 34.3% of the ASD group attended school in Nicosia, 24.7% in Limassol, 21.3% in Larnaca, 15.2% in Paphos, and 4.5% in Ammochostos. The majority (62.4%) of children attended urban schools with a 2.2% (1.8–2.6%) prevalence, while the remaining (34.2%) rural schools, with a lower 1.4% (1.1–1.8%, p=0.005) prevalence, which became not significant after controlling for gender, school level, and system. The age group distribution for the 90 children ASD group sub-sample was as follows: 50.0% at 5 to 6 years; 21.1% at 7 to 8 years; 15.6% at 9 to 10 years; and 13.3% at 11 to 12 years. The overall ASD prevalence in the age group sub-sample was 1.6% (1.3–2.0%).

Among the ASD group, 80.3% were male, and only 19.7% were female, resulting in a statistically significant 4.1 (p<0.001) gender ratio, which was greater for children attending mainstream schools at 4.3 (p<0.05), but lower at 2.6 (p=0.50) for those at special schools. In the special education system, which primarily accommodates more severe cases of ASD, the difference in prevalence between males and females was not statistically significant.

The characteristics of children with ASD are displayed in Tables 2 and 3.

Table 2 Distribution of school and child characteristics for the ASD subgroup in the weighted sample
Full size table
Table 3 Linguistic characteristics and communication level of children with ASD
Full size table

The grade distribution for the ASD group was as follows: 8.4% in kindergarten/pre-school, 27.5% in pre-primary, 11.8% in first, 12.4% in second, 10.7% in third, 9.6% in fourth, 8.4% in fifth, and another 8.4% in sixth grade. Most (42.1%) of the ASD group was placed in special units, 27.5% in general classrooms with the help of an aide, 20.2% in general classrooms independent, and 10.2% in special teams within special schools. Most children received special education (93.8%) and speech-language therapy (90.4%) services. Only 20.8% of the children in general classrooms (47.7%) received extra support from their teachers. All of those attending special schools received additional services like occupational therapy, music therapy, and special gymnastics services, in addition to special education and speech-language therapy, and 83.3% of them received physical therapy.

The main comorbidities reported in addition to the ASD main diagnosis were ADHD for 37.6% and ID for 10.7% of children (Table 3). About a third (33.7%) of children diagnosed were exposed to more than one language (p<0.001). Approximately half (55.6%) were regarded as ASD-NV/MV, including all (100.0%) of those in special education schools and half (50.6%) of those in general education schools (p<0.001). From the ASD-NV/MV subgroup, only 45.5% were using an AAC system to support their communication needs. The most popular AAC systems utilized by schools were the “Picture Exchange Communication System” (Bondy & Frost, 1998) and high-tech devices with communication and voice-generating software.

Factors Associated with ASD Diagnosis

Table 4 displays logistic regression results with ORs and CIs, indicating a 74.4% (62.8–82.3%) lower likelihood of ASD in females compared to males (p<0.001). Children in special schools had a 44-fold (24–82) greater probability of ASD than children in mainstream schools (p<0.001). Children in primary schools had a 50.8% (31.6–64.5%) lower probability of ASD compared to those in pre-primary schools (p<0.001). Additionally, children in rural schools had a 36.0% (13.1–52.9%) lower probability of ASD than children in urban schools (p=0.004). In the Larnaca district, the 63.5% (8.6–143.0%) higher probability of ASD compared to Nicosia (p=0.018) was no longer significant after controlling for gender, school system, and level.

Table 4 Probability of ASD diagnosis by school/child characteristics
Full size table

Table 5 displays the results of the multivariable logistic regression model. After considering all other factors, females and children in primary schools had a significantly lower probability (p<0.001), unlike those in special schools, who had a significantly greater probability of ASD (p<0.001).

Table 5 Multiple logistic regression analysis for ASD prevalence by gender, school system, level, area, and district
Full size table

Discussion

This was the first study reporting the prevalence, gender ratio, and characteristics of children with ASD in Cyprus. Our findings revealed a 1.8% (1.5–2.1%) overall point prevalence for ASD in children aged 5 to 12 years using school report during the 2022–2023 academic year, with a 4.1 gender ratio. This corresponded to about 975–1365 children diagnosed. Males, the children enrolled in pre-primary schools, and those in special schools had a significantly greater probability of ASD (p<0.001). The recently announced NSA reported data from government bodies like the Health Insurance Organization (HIO) and the MOESY, with discrepancies across databases. Specifically, the HIO reported about 1190 children aged under 18 years with the diagnosis, based on assessment records for service provision, a number within the prevalence range found in the prevalence study. In contrast, the MOESY reported 858 children with ASD attending public pre-primary, primary, and special schools (Republic of Cyprus & Deputy Ministry of Social Welfare, 2024), slightly lower than the study estimate. This difference could be due to the inclusion of private schools within the prevalence estimate of the current study.

For the general education system, the prevalence of ASD was 1.6% and as expected, for special education, much greater at 41.9%, due to the nature of special schools. Both the 1.8% weighted and the 1.6% general education ASD prevalence were greater than the global prevalence range (0.6–1.28%) reported by recent systematic reviews using various methodologies (Talantseva et al., 2023; Elsabbagh et al., 2012; Tsai, 2014; Adak & Halder, 2017; Salari et al., 2022; Zeidan et al., 2022), as well as that reported in Greece (Thomaidis et al., 2020). However, they were within the prevalence range reported in Europe (0.6–3.1%) (Chiarotti & Venerosi, 2020) and in the US (1.7–1.9%) (Bougeard et al., 2021). These findings agree with the 1.8% prevalence found in the large-scale English total public school population study (Roman-Urrestarazu et al., 2021), greater than the 0.4–1.6% prevalence range reported by a study in Europe (Bougeard et al., 2021), but significantly lower than the average 2.8% (2.7–2.8) latest ASD prevalence for 2020 in the US (Maenner, 2023), based on school/medical records.

Gender was found to be an effect modifier for the ASD diagnosis, as ASD prevalence was significantly greater (p<0.001) in males than in females. The 4.1 gender ratio was comparable to ratios reported globally (4.2–4.3) (Talantseva et al., 2023; Elsabbagh et al., 2012), in the US (3.8) (Maenner, 2023), in Europe (1.1–16.0 range) (Elsabbagh et al., 2012; Roman-Urrestarazu et al., 2021), and in Greece (4.1) (Thomaidis et al., 2020). Unlike general education schools, the presentation of ASD across genders was not significantly different in special schools. This confirms the findings of a study showing a greater gender ratio in the absence of ID (Rivet & Matson, 2011), and that females with ASD can have lower intellectual ability (Seretopoulos et al., 2019). It might be that females with ASD with normal intellectual ability are harder to identify than males. In Cyprus, children with ASD and comorbid ID can be placed in special schools; therefore, females may be over-represented in special rather than general education schools. This could explain the lower m:f gender ratio found in special schools, suggesting better detection of ASD in females with greater severity, and lower detection of level 1 females with ASD in mainstream schools.

After controlling all factors associated with the ASD diagnosis, the higher ASD prevalence in urban areas and the Larnaca district was no longer significant. Only gender, school system, and school level were significantly associated with the odds of ASD prevalence, which were greater for males (p<0.001), children in special schools (p<0.001), and, unlike the results by Adak and Halder (2017), those enrolled in pre-primary schools (p<0.001). This could reflect a difference in educational policy across studies and might be confounded by school grade, as children in this study could be enrolled in the pre-primary at age 5, and in first grade at age 6, splitting the age group between two grade levels. Additionally, children diagnosed with ASD in Cyprus are allowed to delay their enrollment in primary school in first grade for up to 2 years, meaning they could belong to the 7 to 8 age group, but still attend at the pre-primary level, to be given more time to develop their social, communication, and academic skills. No significant difference was found for ASD prevalence across age groups, contrasting a recent systematic review by Talantseva et al. (2023) where it was significantly greater in the 6- to 12-year-old subgroup than in the other subgroups. The difference in the age cut-offs could explain this since in the current study age 6 was considered in the younger age group of 5 to 6 years. Unfortunately, the age of identification or the duration of the diagnostic procedure was not examined in the current study, which could affect the age group results.

Approximately 20.2% of the ASD group receiving services were independent in the classroom and 27.5% needed the support of an aide. If considered mild support, 47.7% of children with ASD in Cyprus received services at support level 1. A total of 42.1% of children enrolled in special units could be considered receiving moderate support at level 2. A total of 10.2% of children with ASD were enrolled in special teams in special schools and received substantial support at level 3. The vast majority (90.4–93.8%) received speech-language therapy and special education respectively at school, indicating that even cases at level 1 needed support. No information was collected on the duration/frequency of services. Although children with ASD often present with sensory disorders or have poor fine motor skills, this was not reflected in the occupational therapy provision data in this study, because it was only provided within special education schools. The prevalence of comorbidities reported was within the range reported by Bougeard et al. (2021), but the percentages were reduced for ADHD, ID, and epilepsy (Maenner, 2023; Bougeard et al., 2021; Doshi-Velez et al., 2014).

About one-third (33.7%, p<0.001) of the ASD group were reported as bi/multilingual, which could affect their linguistic skills in the Greek language and Cypriot Greek dialect. This was almost double that in the English total school population study, with only 18.5% bilinguals, who were less likely to be diagnosed with ASD (Roman-Urrestarazu et al., 2021). In contrast, the proportion of bi-/multilingualism reported in the current large-scale school report study was about half of the proportion (71.4%) of a convenience sample of 56 children with ASD in a case-control study in Cyprus, where most (46.4%) of the bi-/multilingual subgroup was exposed to the English language (Kilili-Lesta et al., in press). In Cyprus, a previously British colony, English is frequently taught as a second language in schools from the elementary school level and in private practice (Fotiou, 2022), and it could be one of the languages children with ASD were exposed to besides Greek/Cypriot Greek. A greater overall report rate (55.6%) for children with ASD-NV/MV status was found in the current study, which was higher than the 30–40% reported in the literature (Maltman et al., 2020). The definition of the ASD-NV/MV status was given within the school questionnaire, based on the provided literature (Tager-Flusberg & Kasari, 2013), and the schools responded to it subjectively based on their judgement after acquiring a brief language sample and their knowledge of each ASD case. No uniform language sample analysis method was enforced. After the current study, a case-control study in Cyprus examining the linguistic differences within a sample of 56 children with ASD found that by utilizing a new, valid, and reliable parent report questionnaire to measure linguistic level and status between the participants, only 39.3% were found to be classified as ASD-NV/MV (Kilili-Lesta et al., in press). Therefore, the differences in ASD-NV/MV status across studies could be explained by the limited availability of assessment tools in Greek, to objectively determine which children belong to the ASD-NV/MV group in Cyprus. Interestingly, in the current study, only 45.5% of children reported as ASD-NV/MV were using AAC systems to support their communication. This could be explained by the various steps needed in the procedure of assessment and implementation of AAC (Voniati & Christopoulou, 2017) and the limited training specialists usually receive about AAC (Theodorou & Pampoulou, 2022; Wallis et al., 2017).

Limitations and Future Research

Some limitations should be considered in interpreting the findings. First, the ASD diagnosis was not confirmed for reported cases. The schools utilized their special education database, school statistics, and records to respond, but no independent verification of accuracy was realized. Second, the special education database only included children receiving services; hence, it missed children with the diagnosis attending schools without special services. Third, pre-primary schools were limited to reporting children already 5 years old or older. Moreover, the sample had a low representation of private schools. Additionally, the linguistic level of the children could not be verified due to the limited availability of tools measuring the developmental language phase/level of children with ASD. The study design allowed only prevalence and odds ratios to be calculated, lacking the ability to measure incidence or assess causality. Lastly, information on the age of identification and socioeconomic status was not collected for this study, unlike other studies.

Future studies could study ASD prevalence in Cyprus using future centralized data of confirmed ASD cases after public policy for ASD monitoring is enforced by the state, as well as encourage more participation by private schools. Furthermore, the age range in future studies could be limited to children aged 6 to 12 years, to better align with age groups of other studies. Moreover, researchers in future studies could utilize a cohort study design to calculate both childhood prevalence and incidence, along with risk ratios. Information on the socioeconomic status of the children, age of diagnosis, or special education/therapy services frequency/duration could also be useful information for collection in future studies. Additionally, objective tools in the Greek language need to be utilized to measure the linguistic level of children with ASD in Cyprus. Moreover, professional training for SLPs and special educators on available AAC means is necessary to support the communication needs of the children classified as ASD-NV/MV. Currently, public policy must be established to monitor ASD point prevalence and incidence in Cyprus on a state level.

This study marks the first to assess the prevalence and characteristics of ASD in the Republic of Cyprus. Our findings revealed a 1.8% ASD childhood prevalence in 2023, with a 4.1 gender ratio. This underscored the significance of early detection and tailored interventions, especially among males, and those in pre-primary and special education settings. Finally, there is a compelling need for targeted public health initiatives emphasizing early screening, and support services, particularly for males and children in pre-primary/special education, to enhance overall community, well-being, and inclusion.