Rise in post-pandemic autism diagnoses primarily reflects improved detection in girls and women

A new study reveals that increased autism diagnoses after COVID-19 are driven by heightened detection of females, challenging longstanding stereotypes and emphasising the need for gender-sensitive screening approaches.

Much of the apparent post-pandemic rise in autism diagnoses in one large US health system was concentrated among females, according to a new study that suggests clinicians are finally identifying girls and women who would previously have been missed. Researchers found that the overall rate of newly recorded autism diagnoses rose by 7.1% a year from 2020 to 2024, but that increase was driven by females, whose incidence climbed by 19.0% a year after 2020, while male incidence was stable or declining. (jamanetwork.com)

The study, published in JAMA Network Open on 22 July 2026, analysed electronic health records for 171,134 people in a Midwestern US health system who had at least one clinical encounter every year from 2016 to 2024. It split patients into children, adolescents and adults, and tracked first recorded autism diagnoses before and after the arrival of COVID-19. Females were diagnosed an average of 3.4 years later than males, at 15.7 years compared with 12.3 years. Saint Louis University, where psychiatrist Erick Messias led the research, said the work was among the first to trace sex-specific diagnosis patterns across all three life stages within one system. (jamanetwork.com)

The sharpest change was among teenage girls. In adolescent females, the annual incidence of new diagnoses rose by 22.3% from 2020 to 2024; among girls aged up to nine, it rose by 17.4% after 2020, while boys in that younger group saw a 2.0% yearly decline. Over the full study period, male annual incidence stayed higher but broadly flat, whereas female incidence nearly doubled, from 284 new diagnoses in 2016 to 513 in 2024. (jamanetwork.com)

The authors argue that this should not be read as evidence that autism itself suddenly became more common after the pandemic. In the paper, they wrote that “aggregate growth masked a significant sex disparity”. When the anomalous year 2020 was removed from the joinpoint analysis, the female turning point shifted back to 2018, suggesting the pattern pre-dated the worst pandemic disruption. The researchers said the persistence of the rise through 2024 may reflect more sensitive surveillance, wider acceptance of telehealth assessments and broader diagnostic criteria, rather than a simple post-COVID surge in underlying prevalence. (jamanetwork.com)

That interpretation fits a wider rethink about how autism presents in girls and women. The JAMA paper says females may show higher social motivation and camouflaging strategies that allow them to evade screening tools designed around male presentations, with social demands in secondary school then exposing difficulties that were missed earlier. The researchers also said their findings challenge the long-standing four-to-one diagnostic ratio between males and females. A separate JAMA Network Open study published in October 2024 found that the diagnosed ratio in 2022 was closer to three to one, and had been narrowing in both children and adults over time. (jamanetwork.com)

Part of the administrative picture has also been changing. Diagnoses coded as Asperger disorder, removed as a separate category from DSM-5 in 2013, continued to fall as clinicians shifted to the broader autism spectrum disorder label; the researchers said the old F84.5 code was declining by roughly 23% across groups. The study drew on Saint Louis University’s Advanced HEAlth Data Institute and SSM Health’s Virtual Data Warehouse, which together contain records for more than five million patients across Illinois, Missouri, Oklahoma and Wisconsin. Alongside Messias, the author team included Emily Casanova, Regina Huang, Siri Battula, Rachel Loftin, McKenna Walsh and Ping-I Lin. (jamanetwork.com)

The authors were careful about the limits of the evidence. This was a single health system, not a national sample, and the analytic cohort was 81.5% White. Because inclusion required annual contact with the system, it likely overrepresents people with stable insurance and regular access to care, while underrepresenting those with fragmented care or lower socioeconomic status. The adult cohort was also 69.4% female, which may increase the opportunity to identify autism in women. And because the analysis relied on ICD-10 billing codes rather than fresh clinical assessments, it measures recorded diagnosis, not true prevalence in the community. (jamanetwork.com)

Even with those caveats, the practical message is that the recent rise in autism diagnoses is not evenly spread, and services built around the stereotype of a young boy with autism are likely missing a substantial share of female patients. The paper says “diagnostic capacity must extend beyond early childhood”, especially as later recognition can delay access to support, adjustments and community for adolescents and adults whose needs have long gone unexplained. (jamanetwork.com)