The Mortality Gap: Autistic People Die Younger, and It Isn't Inevitable

On average, autistic people die younger than non-autistic people. Pooling 12 studies, researchers found that autistic people's all-cause mortality was 2.37 times that of the general population^[1]. Many of the causes behind that gap, from epilepsy and chest infections to accidents, are ones that health and care services exist to treat or prevent.

Each of these numbers stands for lives that ended sooner than they should have: a chest infection recognised too late, seizures never properly reviewed, a person in distress who met services that would not adapt. Nothing in the evidence says those outcomes are inevitable. The UK researchers who have estimated the gap using life tables call it an inequity that better support, inclusion and access to services could reduce^[2].

Around twice the mortality, in country after country

The best single summary is a 2022 meta-analysis in JAMA Pediatrics. Across 12 studies and 154,238 autistic participants, all-cause mortality was 2.37 times that of the general population (95% CI 1.97–2.85), a finding rated moderate confidence^[1]. National studies, including some published since, point the same way:

• Sweden: 2.60% of 27,122 people diagnosed between 1987 and 2009 died during follow-up, against 0.91% of matched controls. That is 2.56 times the odds of death^[4].
• Denmark: among 1,912,904 children born between 1980 and 2010 and followed to 2013, those with an autism diagnosis had twice the hazard of death (adjusted hazard ratio 2.0). Deaths were rare: 68 people, 0.3% of the group^[5].
• Taiwan: 64,685 autistic people had a hazard ratio of 2.28 for death from any cause compared with non-autistic people of the same age and sex (2008–2019)^[6].
• Australia: in New South Wales, 35,929 autistic people aged 5–64 had mortality rates 2.06 times the general population's (2001–2015)^[7].
• United States: autistic adults aged 65 and over in Medicare had 2.87 times the mortality rate of non-autistic older adults^[8]. Among Medicaid beneficiaries, the standardised mortality ratio was 1.44 compared with the Medicaid population as a whole^[3].
• United Kingdom: adults diagnosed autistic had 1.71 times the mortality rate of matched comparison participants if they had no intellectual disability, and 2.83 times if they had one^[2].

These ratios aren't interchangeable and each study uses its own comparison group, so the chart shows consistency, not a league table. The Medicaid ratio is probably smaller partly because its comparison group, people on a means-tested programme for those on low incomes or with disabilities, already lives shorter lives than the wider US population, as the study's own two comparisons show^[3].

About six years of life lost in the UK, and why "16 years" is the wrong number

A figure of 16 years has circulated widely. It traces back to the Swedish study, where the average age at death of those who died was 53.87 years in the autistic group and 70.20 among controls^[4]. That difference is often described as a life-expectancy gap, but it isn't one. It compares the average age of the small share of each group who died during a limited follow-up, in a mostly young cohort, so it can't be read as years of life lost^[2].

A UK team tested the claim with life tables, the method national statisticians use. They followed 17,130 adults diagnosed autistic without an intellectual disability and 6,450 with one, each matched to ten people from the same GP practice. Without an intellectual disability, estimated life expectancy at 18 was 74.6 years for autistic men and 76.8 for autistic women, against 80.7 and 83.3 for their comparison groups. That is about six years lost (6.14 for men, 6.45 for women). With an intellectual disability, the estimated loss was 7.28 years for men and 14.59 years for women^[2].

The authors call these "apparent" estimates that probably overstate the gap for autistic people as a whole. Only around 1 in 450 people in the data had an autism diagnosis, and diagnosed adults were likely to have higher support needs. Even so, they concluded that there is a group of autistic people who experience premature mortality, "which is of significant concern"^[2].

The most recent US estimate, published in September 2026, covers 2,048,046 Medicaid beneficiaries with an autism diagnosis between 2000 and 2020. It puts their life expectancy at birth at 64.9 years. That is 5.6 years shorter than for all Medicaid beneficiaries and 13.8 years shorter than for the US general population^[3]. The second comparison also takes in the poverty and disability many Medicaid beneficiaries share. The authors add that their figures may underestimate autistic life expectancy, because enrolees are likely to have lower incomes and higher support needs than other autistic people^[3].

In later life the gap may narrow. Among US Medicare enrollees aged 65 and over, those who died did so at a median age of 72 if autistic and 75 if not. That is again age at death, not life expectancy, but the researchers suggest that autistic people who reach 65 may live to a more similar age to their peers^[8].

What autistic people are dying of: seizures, infections and injuries

The excess isn't confined to one cause. The meta-analysis found raised deaths from both natural causes (pooled rate ratio 3.80) and unnatural causes such as injury (2.50), with low confidence^[1]. In Sweden, mortality was raised for every cause category examined except infections. Odds ratios were among the highest for diseases of the nervous system (7.49), and epilepsy was the most common cause of death among autistic people with an intellectual disability^[4]. In Australia, the leading cause was "nervous system and sense disorders", which includes epilepsy, for autistic people with an intellectual disability, and "injury and poisoning" for those without^[7]. In Taiwan, mortality was higher for most specific causes except cancer^[6]. See our article on autism and epilepsy for more on seizures.

The newest US data make the pattern sharper. Among Medicaid beneficiaries with an autism diagnosis, deaths from influenza were 10.55 times the expected number. Malnutrition was 7.56 times, and pneumonitis from inhaling food or fluid 6.92. Accidents, drowning and pneumonia were also markedly raised^[3]. Each has well-established prevention or treatment, which is what makes their over-representation so troubling.

US death certificates from 1999 to 2014 point the same way. Injury accounted for 27.9% of 1,367 deaths where autism was recorded on the death certificate. Drowning made up about 40 times as large a share of those deaths as in the general population (proportionate mortality ratio 39.89). That compares shares of deaths, not death rates^[9].

Suicide is also part of this picture. In Sweden, autistic people had 7.55 times the odds of dying by suicide compared with matched controls^[4]. In England, the LeDeR programme reviews deaths of autistic people and people with a learning disability. Among autistic adults without a learning disability who died in 2021–2024, suicide, misadventure or accidental death together made up around a third (33.0%) of deaths, the most common group of causes. More than half (53.7%) of those reviewed had depression, and reviewers found care that was not consistently adapted to autistic adults' needs^[10]. These deaths are not a consequence of being autistic, and they are not inevitable. The findings point to unmet mental health needs and care that did not adapt, and services can change both. Our article on autism and suicide research covers the evidence, and what helps, in depth.

Next most common in LeDeR were circulatory disease (17.2%) and cancer (15.1%), though its small numbers may not represent all autistic adults^[10].

The gap is widest for autistic women and people with an intellectual disability

Every dataset here that reports it finds higher risk for autistic people with an intellectual disability: see the table below, and the Taiwanese and Australian studies^[6][7]. A small Swedish study published in 2026 followed 113 people diagnosed as children in the 1970s and 1980s, most with an intellectual disability. By 2023, 9.7% had died, against 2.7% of a matched comparison group (hazard ratio 3.77)^[11]. But the gap is not confined to this group. In Sweden, the UK and Taiwan, autistic people without an intellectual disability also had higher mortality^[4][2][6].

People with a learning disability, autistic or not, face a gap of their own. In England in 2024, their median age at death in LeDeR was 62.8 years, 19.0 years below the general adult population (not a life-expectancy measure), and 39.0% of their deaths were from avoidable causes, against 21.1% in the general adult population^[10]. See our articles on autism and intellectual disability and the learning disability health gap.

For women, the pattern is more complex. The meta-analysis found a larger relative excess for autistic women and girls than for men and boys, with low confidence^[1]. In the US Medicaid data, autistic women and men had almost the same life expectancy (65.0 and 64.6 years), but women in general live longer, so the deficit was far larger for women. The authors link this to more psychiatric conditions^[3]. Taiwan also found higher risk for autistic women^[6], but Sweden found higher overall odds for males, with the female excess concentrated among those with an intellectual disability^[4]. The UK authors suggest under-diagnosis may explain part of the female gap: if autism and intellectual disability are more often missed in women, those diagnosed may have the highest support needs^[2].

A gap in health and care, not a fixed fact of being autistic

It would be easy to read these figures as something built into autism. The research can't fully separate biology from care: the Swedish researchers said plainly that they could not tell whether the excess reflected "shortcomings in care provision, increased general biological vulnerability, or both"^[4]. But much of the excess sits within reach of care:

• It is linked to co-occurring conditions. In Denmark, 57 of the 68 autistic people who died (83.8%) had a co-occurring mental, behavioural or neurological condition. Among people with those conditions, an autism diagnosis added no further mortality risk, and the Danish authors suggest the risk may be mediated through, or shared with, those conditions^[5]. They are conditions that can be diagnosed, monitored and treated.
• Life expectancy responds to circumstances. US Medicaid life expectancy for autistic people rose from 63.9 to 65.6 years between 2000–2004 and 2015–2019. In 2020, the first COVID-19 year, it fell by 4.1 years, a bigger drop than in the other two populations studied^[3].
• It follows adversity. The UK researchers point to discrimination, victimisation, unmet mental health needs, inadequate care, homelessness and poor access to healthcare. Difficulty communicating with clinicians can delay the diagnosis of conditions as common as constipation or depression^[2].
• Reviews find failures that can be fixed. LeDeR reviews of autistic adults' deaths describe delays in recognising needs and escalating concerns, poorly coordinated care, and care not consistently adapted to autistic adults' needs^[10].

The World Health Organization notes that autistic people often face stigma and discrimination, "including unjust deprivation of health care"^[12]. Our article on healthcare barriers for autistic adults shows what that looks like in practice.

What the evidence asks of us

The researchers largely agree on the response: better knowledge of autism "in all medical specialties" and coordinated medical care^[4]; better service design, social support and enforcement of equality law^[2]; better management of co-occurring physical and mental health conditions^[7]; and targeted action on preventable excess deaths^[3]. Together, the evidence points to five priorities:

• Treat co-occurring conditions as a priority. Epilepsy, respiratory illness, eating and swallowing difficulties and mental health conditions feature prominently in these deaths. They should be identified, reviewed and treated, not put down to autism.
• Make healthcare adapt. The WHO says autistic people's health needs require "a range of integrated services"^[12]. That means reasonable adjustments, communication support and clinicians who understand how autistic people may show pain, illness and distress.
• Focus on the groups the data identifies. Plan and monitor services specifically for autistic women and for autistic people with an intellectual disability.
• Diagnose autistic adults. The UK authors estimate that only around 1 in 10 autistic people had been diagnosed over the period they studied. They warn that undiagnosed people may face their own risks, such as lack of support, problems at work and financial exploitation^[2].
• Count and learn from every death. Mortality data on autistic people depend on diagnosis records. Routine statistics broken down by sex and intellectual disability, and reviews like LeDeR, are how anyone will know whether the gap is closing.

More research is collected in our Autism & Health research library.

Language: this piece uses identity-first language ("autistic people"), the preference of most autistic-led organisations, and uses clinical terms such as "autism spectrum disorder" only when naming a study or diagnostic category.

Why we're publishing this

Health Insurance UK is a commercial health-insurance resource, not a charity and not a campaign. Autistic people, their families and autistic-led organisations have long pushed for early deaths to be taken seriously, and they deserve the figures accurately, with sources and without exaggeration. The honest numbers are serious enough. The case for change is theirs. Our part is to stand alongside it and say plainly what the data shows: autistic people are dying sooner than they should, often from causes that care could reach.

How to read this data

These figures come from observational studies of registers, GP, insurance and death-certificate records. They show associations, not causes. Odds, risk, hazard, standardised and proportionate mortality ratios are not interchangeable, and each study has its own comparison group and period, so figures from different studies should not be read as a trend. Average age at death is not life expectancy, and life-expectancy figures here are period estimates, not predictions for any individual. Every study relies on a recorded diagnosis, and many autistic adults, especially older adults and women, are undiagnosed, so diagnosed groups may include more people with high support needs and inflate the estimates. All of these studies come from high-income countries or regions.

Use this data

Free to cite with attribution to Health Insurance UK. Charities, campaigners and journalists are welcome to reuse these figures and charts. For the underlying figures as a spreadsheet, get in touch.

Important: This article summarises published research for general information. It is not medical advice and cannot be used to diagnose, treat or predict anything about an individual. If you have concerns about your own or someone else's health, please speak to a qualified health professional.

Sources

1. Catalá-López F, Hutton B, Page MJ, et al. Mortality in Persons With Autism Spectrum Disorder or Attention-Deficit/Hyperactivity Disorder: A Systematic Review and Meta-analysis. JAMA Pediatrics, 2022;176(4):e216401. jamanetwork.com/journals/jamapediatrics/fullarticle/2789090
2. O'Nions E, Lewer D, Petersen I, et al. Estimating life expectancy and years of life lost for autistic people in the UK: a matched cohort study. The Lancet Regional Health – Europe, 2024;36:100776 (doi:10.1016/j.lanepe.2023.100776). eprints-gro.gold.ac.uk/34378/…
3. Li G, DiGuiseppi CG, Blanchard A, Russell MT, Ing C. Autism Spectrum Disorder and Life Expectancy Among Medicaid Beneficiaries. JAMA Network Open, 2026. jamanetwork.com/journals/jamanetworkopen/fullarticle/2853885
4. Hirvikoski T, Mittendorfer-Rutz E, Boman M, Larsson H, Lichtenstein P, Bölte S. Premature mortality in autism spectrum disorder. British Journal of Psychiatry, 2016;208:232–238. doi.org/10.1192/bjp.bp.114.160192
5. Schendel DE, Overgaard M, Christensen J, et al. Association of Psychiatric and Neurologic Comorbidity With Mortality Among Persons With Autism Spectrum Disorder in a Danish Population. JAMA Pediatrics, 2016. europepmc.org/article/MED/26752506
6. Wu CS, Tsai HJ, Chien YL, Gau SS. All-Cause Mortality and Specific Causes of Death in Autism: A Nationwide Analysis. Autism in Adulthood, 2025. europepmc.org/article/MED/40151655
7. Hwang YIJ, Srasuebkul P, Foley KR, Arnold S, Trollor JN. Mortality and cause of death of Australians on the autism spectrum. Autism Research, 2019;12:806–815. europepmc.org/article/MED/30802364
8. Krantz M, Dalmacy D, Bishop L, Hyer JM, Hand BN. Mortality rate and age of death among Medicare-enrolled autistic older adults. Research in Autism Spectrum Disorders, 2023;100:102077. europepmc.org/article/MED/36685335
9. Guan J, Li G. Injury Mortality in Individuals With Autism. American Journal of Public Health, 2017. europepmc.org/article/MED/28323463
10. King's College London. Learning from Lives and Deaths – People with a learning disability and autistic people (LeDeR): 2024 report, key infographics. King's College London, 2026. kcl.ac.uk/ioppn/leder/leder2024/leder-2024-key-inforgraphics.pdf
11. Marinopoulou M, Mårland C, Gillberg C, Billstedt E. Mortality in Autism: A Longitudinal Register-Based Study. Journal of Autism and Developmental Disorders, 2026. europepmc.org/article/MED/41989534
12. World Health Organization. Autism (fact sheet). WHO, updated 2025. who.int/news-room/fact-sheets/detail/autism-spectrum-disorders