Dyslexia is widely recognized as a learning disorder and neurodiverse condition that affects the accuracy and fluency of reading in ways that are atypical relative to an individual's age, education, and/or intellectual ability (1, 2). A child's phonological processing ability is considered one of the strongest predictors of literacy acquisition (3, 4). However, definitions of dyslexia vary across disciplines, regions, and contexts. For example, some frameworks view dyslexia as a distinct diagnostic disorder, while others treat it as a cluster of reading-related difficulties or symptoms. Although there is ongoing debate and variation in how dyslexia is defined across disciplines and contexts, we adopt the definition of dyslexia as a neurodevelopmental condition primarily characterized by difficulties with accurate and/or fluent word recognition, along with poor spelling and decoding abilities. This aligns with definitions provided by the Rose Report (5) and Snowling and Melby-Lervåg (6). Dyslexia may co-occur with other learning or emotional difficulties. It is one of the most prevalent neurodiverse conditions and commonly occurs alongside conditions such as dyspraxia, dyscalculia, Attention Deficit Hyperactivity Disorder (ADHD), and Autism Spectrum Disorder (ASD) (7).

Although global dyslexia prevalence among children and young people (CYP) is estimated to be 7.1% (8), figures can vary widely depending on criteria for diagnosis and language (9, 10). The long-term effects of dyslexia for CYP compared to those without the condition include social, emotional, and behavioral issues, school exclusion, increased likelihood of attending a youth offending institute, and overall poor educational outcomes (11–13).

Schools are an ideal setting to universally screen for children with dyslexia and mitigate against harmful educational outcomes. Yet children with dyslexia are often diagnosed late due to a lack of awareness among teachers and parents regarding the signs and symptoms of the condition. Adopting a universal approach to screening in schools (screening tools can be used before school entry) (14, 15) can help overcome barriers that exist in identifying children with dyslexia that may relate to finance and other socio-demographic factors.

The precision of a school-based screening tool for dyslexia is important and should ideally result in a high percentage of true positives (children who are correctly identified as at risk, known as its sensitivity) and a high percentage of true negatives (children who are correctly identified as not at risk, known as its specificity (16)). However, no single screening tool or set of screening tools can achieve 100% sensitivity and specificity, thus, rendering false positives and false negatives inherent to any tool (3) and is a trade-off and balance between sensitivity and specificity (16, 17). Additionally, screening methods for dyslexia in school settings can be broadly categorized into subjective (e.g., teacher or parent questionnaires, self-report tools) and objective methods (e.g., direct testing of phonological awareness, rapid automatized naming, or decoding skills). Each category carries its own strengths and limitations in terms of reliability, feasibility, and alignment with diagnostic criteria.

Currently, many schools use different approaches to screen for CYP with dyslexia. This review was guided by answering the following research questions: (1) What dyslexia screening tools are most commonly used in school-based settings for children and young people aged 4–16 years? (2) How frequently are combinations of tools employed, and in what ways? (3) Which tools demonstrate the highest reported sensitivity and specificity? and (4) What gaps remain in the evidence regarding socio-demographic equity of school-based screening?

Figure 1 presents the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) diagram and details the search and selection process applied during the screening of articles.

We identified 16 eligible studies of school-based screening for dyslexia in this review. Table 1 presents a summary of these studies including screening tools used and study outcomes (their informants, sensitivity, specificity, the language of administration and prevalence of CYP screening positive for dyslexia by gender and demographics) to improve transparency and facilitate comparison across different contexts. Studies were conducted across 11 countries and six World Health Organization regions (19): European (N = 5), Western Pacific (N = 4), the Americas (N = 2), Eastern Mediterranean (N = 2), Southeast Asia (N = 2), and African (N = 1). Screening tools were available in different languages including English, Mandarin, and Italian.

Participants ranged in age from 4 to 16 years, and the study sample size ranged from 95 to 9,964 participants. In general, early primary school studies tended to emphasize phonological awareness and letter knowledge, whereas studies in older primary and secondary school students focused more on fluency, comprehension, and spelling. This suggests that the applicability of a tool may be age- or stage-dependent.

Seven studies recruited over 1,000 participants: China (N = 3), Italy (N = 2), Norway (N = 1) and Thailand (N = 1) and of these four studies reported similar screening positive prevalence for dyslexia; 3.1–4.9%. Of the eight studies which included a gender distribution for dyslexia screening prevalence, all of them reported higher figures among boys compared to girls.

Of the 16 studies we identified in our review, seven studies reported the socio-economic status of CYP screened. Of these seven, only three studies reported on the socioeconomic status of participants and its association with screening positive for dyslexia (20–22). Socio-economic status was assessed using parental occupation and education, and monthly family income, alone or in combination. Cai et al. (20) found that high levels of parental education was associated with less likelihood of screening positive for dyslexia while no significant association was found between family income and screening positive for dyslexia. Similarly, Lerthattasilp et al. (22) found higher monthly incomes were associated with less likelihood of screening positive for dyslexia and an association between parents with low levels of education screening positive for dyslexia. However, an earlier study by Choudhary et al. (21) showed that the socioeconomic status between CYP with dyslexia and those without (control group) did not differ. Socio-demographic factors such as parental education and family income were inconsistently reported, limiting conclusions regarding equity and generalizability.

No studies reported on the ethnicity of CYP who screened positive for dyslexia and only one study conducted in the USA (23) reported on ethnicity of the study population's CYP reflecting limited socio-demographic reporting. Other studies reported ethnicity in terms of nationality such as China or made no mention of ethnic status.

Dyslexia screening tools were used in combination and concurrently in 12 studies; the mean number of tools used was 3.7 (standard deviation = 2.7), with four studies using only one tool and two studies using eight tools to screen, reflecting variability in methodological approaches. Three studies adopted a staged screening for dyslexia whereby further screening tests were administered if a certain cut-off was achieved at the first stage. Different combinations of dyslexia screening tools were used as well as ‘other tools' which assessed the impact of dyslexia in terms of function as well as other concurrent problems. For example, the strengths and difficulties questionnaire (20) which is an emotional and behavioral screening tool and the Wechsler intelligence scale for children (24). Supplementary Table 2 shows the other screening tools that were identified in this review.

The most used dyslexia-specific screening tool was the Developmental Dyslexia and Dysorthographia 2nd edition (DDE-2). This albeit different subtests of DDE-2 were used four times in two separate studies (25, 26). Among other tools, the most used was the Raven Progressive Matrices which was used in four studies (22, 25–27).

Five studies did not report the prevalence of screening positive for dyslexia highlighting gaps in reporting and limiting cross-study comparisons. Among the remaining 11 studies the reported figure ranged from 3.1% to 33%. The highest figure of 33% was observed in a study from Pakistan (28) where the authors used the Learning Disability Checklist, a non-specific dyslexia test, which may explain the high figure and the likelihood of false positive. High rates of prevalence may be attributed in part to the lack of early assessment and screening strategies at school level (28, 29). In general, reported difficulties in using these tools included time constraints, need for trained personnel, and variability in teacher familiarity with the measures.

Additionally, 11 studies did not report the sensitivity or specificity of the screening tools used. In the five studies which did report these metrics, sensitivity was consistently higher than specificity in all combination of tools administered, for example, Chan et al. (30) reported the sensitivity of the Hong Kong Behavior Checklist of Specific Learning Difficulties as 86% while the specificity was 33%. Nergård-Nilssen and Friborg (32) reported the highest sensitivity of their screening protocol (90.6%) which used the Dyslexia Marker Test for Children. Fletcher et al. (23)'s combined seven tools to achieve the second highest sensitivity metric of 89% screening children approximately aged 5 years. It included the following: (1) Comprehensive Test of Phonological Processes, (2) Rapid Automatized Naming, (3) Alphabetic knowledge, (4) Peabody Picture Vocabulary Test-Revised, (5) Word reading, (6) Woodcock Johnson–Revised (WJR), and (7) Primary Reading Inventory Scale development.

The screening tool with the highest reported specificity of 81% was in a study conducted by Ching et al. (31) where students were screened using the Hong Kong Specific Learning Difficulties Behavior Checklist for Junior and Secondary. The Dyslexia Marker Test for Children in a study by Nergård-Nilssen and Friborg (32) scored the second highest of 70%.

We identified 17 different dyslexia school-based screening tools that were used in the review (Table 2). Students were the main informant (N = 11) followed by teachers (N = 5) and then one tool used by parents (Dyslexia Assessment Questionnaire (21)).

In undertaking dyslexia screening in schools, other tools were used to assess the impact of dyslexia and or concurrent problems (Supplementary Table 2), showing how some protocols integrated cognitive and emotional measures alongside dyslexia-specific screening. We identified 23 other tools that were used, and which assessed features such as intelligence through the Wechsler Intelligence Scale for Children (WISC), emotional and behavioral screening using the Strengths and Difficulties Questionnaire, or symptoms of psychosis and mood through the Kiddie Schedule for Affective Disorders and Schizophrenia. Of these other tools, all bar one where the parent was the informant, required completion by the young person.

Our narrative systematic review reveals significant global differences in school-based dyslexia screening for children and young people (CYP) aged 4–16 years. The studies we included covered various countries, populations, and languages, making direct comparisons challenging. Additionally, few studies provided data on the sensitivity and specificity of the screening methods used. As a result, we were unable to determine whether school-based dyslexia screening should rely on a single tool, or a combination of tools used either together or in sequence. Our findings reinforce the need for clarity in how dyslexia is defined and operationalized in screening contexts. Tools used without consistent criteria or theoretical underpinning may contribute to under- or over-identification, especially in diverse populations.

The most commonly used dyslexia screening tool was the Developmental Dyslexia and Dysorthographia tool (second edition), which consists of eight subsets. In studies, different subsets of this tool were used for screening. It was developed based on norms from a sample of 1,200 children aged 7–14 years (33). The second most frequently used tool was the Raven Progressive Matrices, which was always combined with another screening tool. While standardized across various populations, this tool is designed to assess abstract reasoning and problem-solving abilities through pattern recognition (34), rather than specifically measuring dyslexia.

We identified several screening tools, such as the Strengths and Difficulties Questionnaire, that were not specifically designed for dyslexia screening but instead assessed the impact of dyslexia. For example, these tools measured the effects of dyslexia on an individual's psychosocial functioning, as well as emotional and behavioral difficulties. This type of supplementary screening could play an important role in school-based assessments, helping to identify children and young people with dyslexia who face the greatest challenges and guiding how resources are prioritized for them.

Research to date for dyslexia screening recommends the use of a tool with a sensitivity score of 90% (35, 36); high sensitivity scores are likely to identify nearly all children with dyslexia and that additional tests are then used to confirm an accurate diagnosis; removing the number of children identified wrongly (false positives). In our review, we found one study by Nergård-Nilssen and Friborg (32) that used a Dyslexia Marker Test that met this recommendation reporting a sensitivity of 91%. This tool consisted of six subtests: letter knowledge test, phoneme isolation, phoneme deletion, RAN, working memory, decoding and spelling and covered three areas: ability, attainment and diagnostic. The only other study conducted by Fletcher et al. (23) that nearly met this recommendation, reported a sensitivity score of 89% and used five screening tools.

The prevalence of children screening positive for dyslexia varied widely across studies, ranging from 3% to 33% of the sample population. The highest prevalence was reported in a Pakistani study (33%; 27), which used a Learning Disability Checklist. We also found that only 50% of the studies reported on the gender of students, 19% on their socio-economic status, and none on their ethnicity among those who screened positive for dyslexia. When considering educational stage, screening practices also diverged: tools for early years (ages 4–6) prioritized early phonological skills, rapid naming, and alphabet knowledge, whereas upper primary and secondary school tools often assessed fluency, comprehension, and spelling. Framing tools by school stage as well as age may help schools select the most developmentally appropriate instruments.

The wide variation in the dyslexia screening tools used in school-based research may reflect a lack of consensus on the best approach. A recent systematic review found that psychologists in English-speaking countries show little alignment in their methods for assessing dyslexia (37). Furthermore, the definition of dyslexia itself has evolved over time, with arbitrary cut-offs and the absence of clear diagnostic criteria (5, 6). This has led to differences in operational definitions and screening tools, each emphasizing different skills and characteristics (38). As a result, the lack of a unified approach to screening limits the ability to implement universal methods and complicates decisions within schools and the broader education system regarding which tool to use for dyslexia screening (38).

Various school-based screening tools are used to identify children with dyslexia, but there has been no consensus over time regarding the best approach. The wide range of dyslexia screening tools identified in this review raises questions about whether such diversity is necessary and or reflects variation in definition. Increased collaboration between researchers and front-line educators could help address these differing approaches and establish an evidence-based screening method. In the meantime, an initial screening with a highly sensitive tool, followed by more specific tests to assess detailed deficits and their impact, may provide a beneficial approach.