Are we 'overdiagnosing' harmless problems?

Many patients are being “overdiagnosed” and given unnecessary medical treatment for problems that will never harm them, says new research. The study has featured prominently in the press, with the Daily Mail reporting “a plague of overdiagnosis”, and The Independent saying that “an over-reliance on healthcare threatens to bankrupt the world”.

The article, published in the British Medical Journal, argues that overdiagnosis poses a significant threat to human health by labelling healthy people as sick and wasting resources on unnecessary care. Overdiagnosis is when somebody is diagnosed and treated for a condition that is unlikely to cause them harm.

The authors of the report say there is growing evidence that “too many people are being overdosed, overtreated and overdiagnosed”. For example, they say, research has found that almost a third of people diagnosed with asthma may not have the condition, while up to one in three breast cancers detected by screening may actually be harmless.

What is overdiagnosis?

The authors state that overdiagnosis is when a person without symptoms is diagnosed with a disease that will not cause them to experience symptoms or early death. More broadly defined, overdiagnosis refers to the related problems of increased dependence on medical care and overtreatment, which leads to healthy people with mild problems or at low risk being “reclassified as sick”.

As a result of overdiagnosis people potentially face the harms of unnecessary tests and side effects from treatment, and resources that could be directed at other patients are wasted.

The authors of this new article say that many different factors are driving overdiagnosis, but a key contributor is technological advance. Tests and diagnosis methods are now so numerous and sensitive that even the tiniest of harmless abnormalities can be detected, they argue.

How does overdiagnosis happen?

The authors say that people can be overdiagnosed and overtreated in a number of ways:

Screening programmes can detect disease that may be in a form that will never cause symptoms or early death (sometimes called pseudodisease). Contrary to popular notions that cancers are universally harmful and ultimately fatal, the authors point out that some cancers can regress, fail to progress or grow so slowly that the individual concerned dies from other causes. They say there is now strong evidence from randomised trials that a proportion of cancers detected through screening may fall into this category.
Tests for specific disease and disorders have become increasingly sensitive, enabling less severe forms of disease to be detected. A substantial proportion of the abnormalities that are detected will never progress, they say.
Diagnostic scanning of the abdomen, pelvis, chest, head and neck can reveal ‘incidental’ findings in up to 40% of people being tested for other reasons. Most of these incidental abnormalities are benign but cause anxiety and lead to further investigations, they say.
Overdiagnosis also occurs because of the changing diagnostic criteria for many diseases, so that people at lower risk and with milder problems are being defined as sick. For example, say the researchers, most elderly people are now classified as having at least one chronic condition, while many women treated for osteoporosis (weak bones) may be at very low risk of a fracture. The authors argue that diagnostic criteria are often set by panels of health professionals “with financial ties to companies that benefit directly from any expansion of the patient pool”.

What are examples of overdiagnosis?

The authors say there is evidence that the problem of overdiagnosis may exist across many conditions (including those for which underdiagnosis may also be a problem) and cite research on overdiagnosis in several different areas. The authors provide their opinions on various examples of what they see as overdiagnosis:

Breast cancer – a systematic review has suggested that up to a third of breast cancers detected by screening may be overdiagnosed, which means they would not actually cause harm or early death if left untreated.
Thyroid cancer – the likelihood of tests detecting a thyroid abnormality is high but the risk that it will ever cause harm is low. Many of the newly diagnosed thyroid cancers are the smaller and less aggressive forms that don’t require treatment, which itself carries risks.
Gestational diabetes (diabetes that develops during pregnancy) *–* an expanded definition of this condition now means that nearly one in five pregnant women are classified as having it, while the evidence of benefit for being diagnosed is weak.
Chronic kidney disease – an expanded definition of this condition means that one in 10 people in the US are now classified as having the disease. One study estimates that up to a third of people over 65 meet the new criteria, yet each year fewer than 1 in 1,000 of this group will develop end-stage kidney disease.
Asthma – the authors concede that while asthma can be underdiagnosed and undertreated, one large study suggests that almost a third of those diagnosed may not have the condition and that two-thirds of this group do not require medication.
Pulmonary embolism (a blockage in the artery leading to the lungs, caused by a blood clot) – while pulmonary embolism is potentially fatal, the authors say that newer and more sensitive diagnostic tests are leading to the detection of smaller clots that may not require treatment.
Attention deficit hyperactivity disorder (ADHD) – a widened definition of this condition has led to concerns about overdiagnosis, with one study showing that boys born at the end of the school year have a 30% higher chance of diagnosis and 40% higher chance of requiring medication than those born in the following month.
Osteoporosis – expanded definitions of this condition mean many women at low risk of fracture may be given treatment that can result in adverse effects.
Prostate cancer – research shows that the risk of a cancer detected by prostate specific antigen (PSA, a marker for prostate cancer found in the blood) being overdiagnosed may be more than 60%.
Lung cancer – the authors cite research suggesting that around 25% of cases of lung cancer detected by screening may be overdiagnosed.
High blood pressure – the authors cite research suggesting there is a possibility of “substantial overdiagnosis” of high blood pressure.
High cholesterol – the authors cite research that estimates that up to 80% of people being treated have near-normal cholesterol levels.

To clarify, the above statements reflect the opinions of the study’s authors, not Behind the Headlines or NHS Choices.

What are the causes of overdiagnosis?

The authors say that overdiagnosis is being driven by several factors, including:

technological advances that can detect ever-smaller harmless abnormalities
commercial and professional vested interests involved in the expansion of disease definitions and the writing of new guidelines for diagnosis and treatment
legal incentives that ‘punish’ underdiagnosis but not overdiagnosis
health system incentives favouring more tests and treatments
the cultural belief in early detection of disease and medical intervention

What solutions do the researchers recommend?

The authors argue that action is needed to tackle the problem of overdiagnosis. They say that medical professionals should aim to differentiate between benign abnormalities and those that will go on to cause harm, while the public and professionals should be given “more honest” information about the risks of overdiagnosis, particularly related to screening.

The authors say new protocols are being developed to bring a more cautious approach to the treatment of incidental abnormalities. They say consideration should be given to raising the thresholds of what is defined as abnormal – in breast cancer screening, for example. At a policy level, reforming the process of defining disease is urgently required, they argue, to rule out financial or professional conflicts of interest.

The authors point out that concern about overdiagnosis does not preclude awareness that many people with genuine illness miss out on healthcare. They argue that resources wasted on unnecessary care can be much better spent treating and preventing genuine illness. “The challenge is to work out which is which.”

Conclusion

This is a powerfully argued and controversial article that claims that many people are being overdiagnosed and overtreated for mild problems that may never cause them harm. It is timed to coincide with the announcement of an international conference on the topic, which will take place next year, partly hosted by the BMJ and Bond University, where some of the authors are based. It should be noted that the article is not a systematic review of the evidence on screening or on overdiagnosis, but a strong opinion piece that cites research in support of the authors’ argument.

Nevertheless, the paper is a useful contribution to the complex debate about how far healthy people should be screened or tested, and how far conditions that may or may not cause harm in the future should be treated. It’s a difficult subject that provokes strongly opposing views among doctors and researchers. For example, a recent paper published in The Lancet argued that everyone over 50 should be given statins because they have been found to reduce the risks of heart attacks even in healthy people.

The article raises a number of concerns about cancer screening in particular, and it should be noted that in the UK the Department of Health announced last October that a full review of the potential risks and benefits of the NHS Breast Cancer Screening Programme is to take place. In terms of cancer screening, much will depend in the future on how far researchers are able to discriminate between ‘harmless’ slow growing cancers that do not need to be treated and those that are more aggressive.

The article should be seen in the context of previous advances in technology and treatments that have brought established benefits in terms of detecting certain conditions early. For example, high blood pressure is symptomless, but it is a recognised risk factor for cardiovascular disease, and there is good research to show that treatment to reduce high blood pressure saves lives.

Overall, the concept of overdiagnosis is one that needs careful consideration, particularly how its potential harms relate to the potential harms of failing to diagnose a disease. Is it better to risk side effects in patients than to risk missing a serious health problem? The issue is perhaps too big to cover in a single article, although this latest opinion paper raises some extremely interesting and thought-provoking points on the issue. Perhaps the key thing now is to look at overdiagnosis on a condition-by-condition basis, for example in the way that alleged overdiagnosis in breast cancer screening is being examined in the UK.

Although there are clearly strong feelings on both sides of the overdiagnosis debate, examinations of the issue should be as comprehensive, objective and evidence-based as possible and focus on specific topics. Overdiagnosis of one condition could be problematic, while it might produce few negative results for other conditions. Next year will see an international conference on the issue of overdiagnosis, which should stimulate both debate and research on this important issue.