Autism – Male vs. Female

For decades, autism spectrum disorder (ASD) has been portrayed as a condition far more common in males than females. The classic statistic—that boys are diagnosed with autism four times as often as girls—has shaped research, clinical practice, and public understanding. But is this 4:1 ratio a fundamental biological truth, or a shadow cast by our own diagnostic systems and societal blind spots?

A groundbreaking new study from Sweden, leveraging decades of nationwide health registers, provides the most comprehensive answer yet. By meticulously separating the influences of age, time, and generation, the research paints a dynamic picture: the so-called “male bias” in autism is not static. It is changing, and those changes tell a powerful story about diagnosis, awareness, and the girls and women who have been overlooked.

Autism is a neurodevelopmental condition characterized by differences in social communication and patterns of repetitive or restricted behavior. Its prevalence has risen significantly since the early 2000s, a change largely attributed to broader diagnostic criteria, increased awareness, and improved services.

Yet alongside this rise, a persistent mystery has lingered: Why are so many more boys diagnosed? Several theories have been proposed:

• The “Female Protective Effect”: Girls might need a higher “genetic load” of risk factors to show behaviors that cross the diagnostic threshold.
• The “Extreme Male Brain” theory: Autism may represent an amplification of typical male cognitive styles (systemizing over empathizing).
• “Camouflaging”: Girls often learn to mask autistic traits by mimicking social peers, making their struggles less obvious.
• Diagnostic Overshadowing: Co-occurring conditions like anxiety or ADHD in girls can overshadow core autistic features, leading to missed or delayed diagnoses.
• A simple sex bias in research and diagnostic tools, which were historically developed and validated on boys.

The critical question has been: Is the high male-to-female ratio (MFR) real, or is it an artifact of our failure to recognize autism in females? Previous studies hinted at the latter, showing the MFR appears lower in adults and younger children, suggesting a “catch-up” in female diagnoses over time. But no large study had untangled the complex web of age at diagnosis, calendar period (year of diagnosis), and birth cohort (generation) to see which factor was driving the change.

Sweden offered the perfect laboratory for this investigation. The country has a universal, publicly funded healthcare system and meticulously linked, population-based national registers dating back decades. These include:

1. The Medical Birth Register: Records of all births since 1973.
2. The National Patient Register: Tracks all hospital and outpatient diagnoses from 1987 onward, using standardized international codes (ICD-9/10).
3. The Multi-Generational Register: Links family information across generations.

The researchers identified every child born in Sweden between 1985 and 2020 whose parents were both born in Sweden (over 2 million individuals). They then followed this cohort from birth, tracking who received an ASD diagnosis from age 2 until the end of 2022.

This design allowed them to apply a powerful epidemiological tool: the Age-Period-Cohort (APC) model. This statistical approach is like a time-traveling detective, separating three overlapping influences:

• Age Effect: How diagnosis rates change as a person gets older (e.g., most diagnoses happen in early childhood).
• Period Effect: How diagnosis rates change in a specific calendar year for all ages (e.g., a sudden jump in 2013 after the DSM-5 criteria were published, affecting everyone diagnosed that year).
• Cohort Effect: How diagnosis rates differ between generations born in different years (e.g., are kids born in 2010 more or less likely to be diagnosed than kids born in 1990?).

By modeling these three forces simultaneously, the study could determine whether the shifting MFR was due to:

• Girls being diagnosed later in life (Age effect)?
• A specific year’s diagnostic changes affecting everyone (Period effect)?
• Or a fundamental shift in how autism presents or is recognized in younger generations (Cohort effect)?

The analysis confirmed that all three forces are at play, but with dramatically different impacts on the male-to-female ratio.

1. The Age Effect: Diagnosis is a Childhood Event (Mostly)

As expected, ASD diagnoses cluster in early and middle childhood. The incidence rate rises steeply around ages 3-5 and then declines. This pattern was consistent across the study period. However, the crucial finding here was that the MFR changes dramatically with age. For the youngest children (under 6), the MFR was around 2.5:1. But for adults diagnosed after age 30, the ratio ballooned to over 6:1. This stark difference is a smoking gun for late diagnosis in females. It suggests that many autistic girls are slipping through the net in childhood and only receiving a diagnosis when their struggles become untenable in adolescence or adulthood—often when seeking help for anxiety, depression, or other co-occurring conditions.

2. The Period Effect: The DSM-5 Shockwave

The study captured a massive period effect around 2013-2015. This aligns precisely with the publication of the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), which replaced the old subcategories (like Asperger’s) with a single “spectrum” model. This change immediately broadened the criteria, leading to a surge in diagnoses across the board for all ages in those years. Importantly, this surge increased the MFR temporarily. Why? The researchers hypothesize that the new criteria, while broader, may have still been more readily applied to males, or that the increased clinical focus on autism following the DSM-5 release initially prioritized identifying the “classic” (and often male) presentation.

3. The Cohort Effect: A Generation of Change (The Most Important Finding)

This is where the study delivers its most profound insight. When looking at different birth cohorts—groups of children born in different decades—a clear and powerful trend emerges: the MFR is falling dramatically in younger generations.

• For the cohort born in the late 1980s, the MFR was approximately 4.5:1.
• For the cohort born in the late 1990s, it dropped to about 3.5:1.
• For the most recent cohort born in the late 2010s, the estimated MFR had fallen to around 2.5:1.

This isn’t about girls being diagnosed later; it’s about girls being recognized earlier in more recent generations. This is a cohort effect. It points to a generational shift in awareness, clinical practice, and possibly even the expression of autism itself. Clinicians, teachers, and parents are now more attuned to how autism can manifest in girls—through intense, “camouflaged” social effort, special interests that are socially acceptable (e.g., horses, pop stars), or internalizing symptoms like anxiety. Diagnostic tools and understanding are slowly improving.

This study dismantles the notion of a fixed, biological 4:1 male-to-female ratio. Instead, it reveals a ratio in flux, shaped by society.

1. The “Female Protective Effect” May Be Overstated: If girls require a higher genetic burden, we would expect the MFR to be stable across cohorts and ages. The dramatic fall in the MFR for younger, recently diagnosed girls suggests that much of the historical disparity was diagnostic, not biological. Many autistic girls were always there; we just didn’t see them.
2. Camouflaging Has a Cost: The extreme age effect (MFR of 6:1 in adults) is a direct measure of the toll of camouflaging. Girls and women spend years exhausting themselves trying to pass as neurotypical, leading to high rates of burnout, anxiety, and depression before they finally connect with an autism diagnosis that provides understanding and support.
3. Diagnostic Criteria and Clinician Training Matter: The period effect shows that changes in manuals like the DSM have immediate, population-wide impacts. The subsequent cohort effect shows that with time, education, and advocacy, those same criteria can be applied more equitably. This is a call to action for continuous refinement of diagnostic tools and mandatory training on gender differences in autism presentation.
4. We Need Lifelong Support: The data confirms that autism is a lifespan condition. Services and diagnostic pathways cannot stop at age 18. We must build robust, gender-sensitive pathways for autistic girls and women to receive assessment and support in adolescence and adulthood.

The study is not without limits. It only includes individuals with Swedish-born parents, potentially missing patterns in immigrant populations. It also captures only diagnosed ASD, so undiagnosed individuals are not represented. However, its strength—using a near-complete national dataset—far outweighs these constraints.

The findings are a clarion call. The historical focus on the male presentation of autism has created a diagnostic blind spot with real human consequences. The falling MFR in young cohorts is proof that progress is possible. It is a testament to the advocacy of autistic women and girls who have tirelessly described their experiences, and to clinicians who are finally listening.

The goal now is not to achieve a 1:1 ratio, but to ensure that any autistic person, regardless of gender, can access a timely, accurate diagnosis and the support they need. The Swedish data shows we are moving in that direction. The challenge is to accelerate that change globally, so that the next generation of autistic girls doesn’t have to wait until adulthood to understand themselves.

https://www.bmj.com/content/bmj/392/bmj-2025-084164.full.pdf