Debates around whether autism is overdiagnosed have dominated recent media headlines. These have continued despite the fact that the government has just published an independent review that confirms that there is no evidence of overdiagnosis at scale.
To support claims of overdiagnosis, we would need a range of data. Right now, that data doesn't exist for autism.
In this article we will explain:
- why more people are talking about this topic
- what overdiagnosis is
- what the evidence says, and
- why claims of overdiagnosis are incorrect, unhelpful and harmful
Why people are talking about overdiagnosis
More people are being diagnosed as autistic. This is likely due to a combination of factors. For example, greater awareness of autism, better recognition of groups historically missed by services, and increased need for support.
At the same time, we've seen more referrals for autism assessment, more people joining long waiting lists. But there’s no investment to match this; services are unable to keep up, and too many people left without vital support.
In response, last year the government commissioned the Independent Review into Mental Health Conditions, ADHD, and Autism. This review has now been published and concludes that there is no evidence of overdiagnosis at scale.
One of the big questions for the review was how common autism is. This is known as its prevalence. This is different from incidence, which refers to the number of new cases over a set time. Both prevalence and incidence form part of the data required to make judgements on under or overdiagnosis.
The media has widely reported on overdiagnosis. It has oversimplified a complex issue. It suggests that if autism diagnoses are rapidly increasing and services are struggling, autism must be being overdiagnosed. However, that does not reflect the evidence and the definition of overdiagnosis.
Defining overdiagnosis
To understand whether overdiagnosis claims are valid, we need to define it.
Overdiagnosis refers to the correct diagnosis of a medical condition that would have been unlikely to cause symptoms. In these cases, the diagnosis causes avoidable harm. For example, a very small or slow growing cancer which probably wouldn’t have caused health problems in an individual. However, screening for it and detecting it led to unnecessary tests, treatment or anxiety.
Additionally, it’s important to understand that underdiagnosis can harm autistic people. Without a diagnosis, an autistic person might not understand why they find certain things difficult. They might blame themselves, adding to their distress. Even without additional support, many autistic people find a diagnosis gives them permission to be kinder to themselves.
Overdiagnosis is different from misdiagnosis, where the diagnosis itself is wrong. An example of misdiagnosis could be a late-diagnosed autistic person with autistic burnout might have been previously misdiagnosed with depression.
To better understand overdiagnosis and autism, we reviewed the research literature since 2008. We found that overdiagnosis on an individual basis tends to refer to an autism diagnosis which essentially does more harm than good. However, multiple other issues can lead to researchers discussing autism more broadly as being overdiagnosed. This includes incorrect use of diagnostic measures and methods, a lack of clinical expertise in cases where a diagnosis is wrongly given, or where one condition is misdiagnosed as another.ii Some outlets also describe overdiagnosis as diagnosis rates overtaking population prevalence estimates.
Understanding prevalence is vital for decision makers delivering services or allocating resources. It’s important to understand if there are examples across the country where people are being wrongly diagnosed or where a diagnosis doesn’t help. But it'll take a very long time to do properly. And it's vital that work is carried out using all relevant data and evidence, including an understanding of what is underpinning people’s support needs and whether these are being met.
The evidence to help us understand if autism is overdiagnosed
It’s important to be cautious about strong claims about autism under or overdiagnosis. Understanding whether autism is overdiagnosed requires multiple different kinds of evidence, including:
- Representative population prevalence data (how many people are autistic on a population level)
- Incidence data (how many new cases have been measured in a given period)
- Data on diagnostic accuracy and reliability
- Data on how valid the current diagnostic thresholds, measures and tools are
- Experience and outcome measures, such as whether a diagnosis and support is helping people and how.
- Inequalities data covering age, sex/gender, ethnicity, learning disability, socioeconomic circumstances and geography.
Drawing on this range of data would allow you to more confidently provide an estimate on the number of autistic people in a given population (prevalence), how many new cases have been measured in a given period (incidence), and whether you agree with current assessment process and how these are carried out by specialists. You would also have robust data which can say with certainty whether a diagnosis and the support your system provides is helping people and how.
There are scenarios where flagging overdiagnosis could be helpful. For example. If the diagnosis isn’t helping people, if the support doesn’t work, and if there is solid and long-term evidence of clinical errors, then citing overdiagnosis could show where else to prioritise resources. For example, it may point to a need to invest in more staff, change training modules, improve commissioning and regulation processes, or boosting the support packages offered.
Currently, we don’t have all this data available. The evidence is not there. The autism chapter of the Adult Psychiatric Morbidity Survey (APMS) was withdrawn in May this year due to data errors and methodological challenges.ii And the Independent Review’s report concluded that a multitude of things contribute to more people seeking help, including rising increased distress and increased awareness.
To understand over or underdiagnosis, we need better integrated data. It would cover autistic people’s access to services across education, employment, justice, welfare, NHS, and social care systems. This would need to sit alongside experiences of autistic people and their families or carers. The data would need to capture the dynamic and changing nature of autistic people’s support needs, the help they receive, and their experiences of that help.
The current system for supporting autistic people is failing. Too many people are many going without any support at all, including people stuck on long waiting lists.iv To ensure everyone can benefit from the help they need after a diagnosis, there needs to be a formal support system to improve people’s lives. These improvements would be reflected in data about people’s experiences and outcomes.
What existing evidence says about overdiagnosis
It is not possible to conclude there is an overdiagnosis of autism based on current available evidence. Making this claim without evidence is both unhelpful and harmful. Until there is better data to support overdiagnosis, and the data is used to improve outcomes and service planning, then these claims lack evidence and continue to stigmatise autistic people.
The Independent Review is clear that we cannot claim there is currently overdiagnosis at scale. They recommend new data collection to understand population prevalence, but recognise this will take time. They also point to potential risks of misdiagnosis and overdiagnosis in future, but only if our systems continues to fail autistic people.
What the review and wider evidence does show clearly is that there are growing instances of autistic people’s needs going unmet, long waits for assessment, a lack of support, too many examples of care being rationed, and experiences and outcomes not improving.
Looking to the future
We hope that the conclusions in the prevalence review draw a line under the constant debate and questioning on the topic of overdiagnosis. It is harming our community and focusing attentions away from what matters, which is meeting rising levels of distress and unmet needs. Our immediate focus now must be on effective support before and after diagnosis for those who need it. We need to see government addressing the unacceptable long waiting times for a diagnosis. And the government should be adequately resourcing the services that currently do provide support, advice and advocacy.