Abstract
INTRODUCTION: Autism spectrum disorder (ASD) shows marked clinical heterogeneity and a pronounced sex disparity in diagnosis, with male-to-female ratios of up to 4.6:1 in the Spanish educational context, suggesting potential systematic under-identification in females. This systematic review critically examines gender bias in ASD diagnosis, focusing on females, and assesses whether current diagnostic criteria adequately capture Level 1 presentations. METHOD: A literature search was conducted in December 2025 across American Psychological Association's database (APA) PsycInfo, Dialnet, PsicoDoc and Education Resources Information Center (ERIC), using Population, Exposure, Comparator, Outcome (PECO) and Patient, Intervention, Comparison, Outcome (PICO) frameworks. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) process, 26 studies (2015-2025) were included, encompassing observational, psychometric, technological, quasi-experimental and qualitative designs. Overall methodological quality was good to very good (mean CRF-QS = 15.7), and risk of bias was predominantly low. RESULTS: Findings indicate consistent diagnostic inequities: delayed diagnosis in girls, reduced sensitivity of standard instruments to less externalizing phenotypes, the role of social camouflaging, and moderating effects of comorbidities and contextual factors. CONCLUSION: Diagnostic bias appears multifactorial and calls for gender-sensitive, multi-method and intersectional approaches to ensure equitable access to diagnosis and support.