Researchers discover solutions to gender bias in autism diagnoses

Summary: Study reports of girls and boys show similar rates of concern about ASD and identify several biases that contribute to the inflated sex ratio for autism diagnosis. The findings could help with the early identification of girls on the autism spectrum.

Source: University of Minnesota

Posted in Biological psychiatryA multidisciplinary study conducted by the University of Minnesota showed that equal numbers of girls and boys can be identified as concerned about autism spectrum disorder (ASD) when screened previously, correcting large gender differences in current diagnoses.

“The conventional wisdom is that more boys than girls have ASD,” said the study’s lead author Casey BurrowsPh.D., LP, assistant professor at the University of Minnesota Medical School and psychologist with M Health Fairview.

“Our research shows that girls and boys show similar rates of concern about ASD and identifies some of the biases that contribute to inflated sex relationships. We hope this research will bring relief to women and girls who have struggled socially without knowing why. ”

Using data from the Infant Brain Imaging Studies Networkthe study used a less partial sample that monitored a group of children most likely to develop ASD (eg, newborn siblings of autistic children) aged six to 60 months.

The study found that there are as many girls identified with ASD problems when children are screened early and when they are corrected for sex-based bias in diagnostic tools. This is in stark contrast to the current 4 to 1 sex ratio when following standard clinical referral processes.

“We know that screening processes and diagnostic tools in ASD are often lacking in many girls who are later diagnosed with ASD,” said Dr. Burrows, who is also a member of the Masonic Institute for Brain Development.

“This prevents many girls from receiving early intervention services at a time when they can have the greatest impact in early childhood. Most studies on ASD focus on children after they are diagnosed, lacking information about symptoms in children that were not detected by common screening practices. “

The research team examined whether girls and boys had similar symptoms and found subtle differences in the structure of the core symptoms of ASD. After correcting for these differences, subgroup analysis identified a “high risk” group that had a 1: 1 ratio of males to females.

The study found that there are as many girls identified with ASD problems when children are screened early and when they are corrected for sex-based bias in diagnostic tools. The image is in the public domain

“This approach – impartial assessment, ensuring that our tools are measuring what we think they are measuring – can help address current disparities in autism identification,” according to Jed ElisonPh.D., associate professor at the Institute of Child Development and Medical School and co-author of the paper.

“It is critical to recognize and understand the limitations of traditional diagnostic and screening approaches and to generate creative solutions to identify all children who could benefit from early intervention services.”

Researchers are planning to follow up on this work by examining how children in the high social concern group of primary to secondary school age fare. They are also studying group differences in underlying brain structure and function.

Financing: This study was supported by grants from the National Institutes of Health (R01-HD055741, R01-MH118362-01, R01-MH118362-02S1, U54-HD079124, P50-HD103573 (Project ID 8084), U54-HD086984), Autism Speaks, and the Simons Foundation (140209). Dr. Burrows was supported by an NIH Career Development Award (K12-HD055887).

About this autism research news

Author: Kat Dodge
Source: University of Minnesota
Contact: Kat Dodge – University of Minnesota
Image: The image is in the public domain

See also

This shows a woman's eye

Original research: Closed access.
“A data-based approach in an unbiased sample reveals a gender equivalent ratio of the disorder associated with autism spectrum disorder in early childhood” by Casey Burrows et al. Biological psychiatry


A data-driven approach in an unbiased sample reveals an equivalent gender ratio of impairment associated with autism spectrum disorder in early childhood


Sex differences in the prevalence of neurodevelopmental disorders are particularly evident in autism spectrum disorder (ASD). The heterogeneous presentation of symptoms and the potential for measurement bias hinder early detection of ASD in females and may contribute to discrepant prevalence estimates. We examined the trajectories of social communication (SC) and restricted and repetitive behaviors (RRB) in a sample of siblings of autistic children, adapting measurement biases based on age and gender. We hypothesized that exploiting a potential high familial probability sample, deriving data-based behavioral constructs, and accounting for measurement bias would reveal less discrepant sex ratios than those typically observed in ASD.


We conducted direct assessments of ASD symptoms at 6-9, 12-15, 24, and 36-60 months of age (total Nremarks= 1254) with infant siblings of autistic children (N = 377) and a lower familial ASD-probability comparison group (N = 168; Nremarks= 527). We established the invariance of the measurement between age and sex for separate models of SC and RRB. We then conducted a modeling of the latent class growth mixture with the longitudinal data and evaluated the sex differences in trajectory membership.


We identified two latent classes in SC and RRB models with same-sex ratios in the high-risk cluster for both SC and RRB. Gender differences were also observed in the high concern SC cluster, indicating that girls classified as “high social concern” show milder symptoms than boys in this group.


This novel approach to characterize the progression of ASD symptoms highlights the utility of assessing and adjusting gender-related measurement biases and identifying patterns of sex-specific symptom onset.

Leave a Comment