Most of the causes of diagnostic failure we have so far reviewed are fairly obvious. In this post however, we will look at a less evident cause which is overdiagnosis. In his paper titled Overdiagnosis: an unrecognised and growing worldwide problem in healthcare, John Brodersen defined this as the diagnosis of deviations, abnormalities, risk factors, and pathologies which would never cause symptoms, and will never result in morbidity or mortality. In a similar definition in their paper titled How to prevent overdiagnosis, Arnaud Chiolero and colleagues referred to overdiagnosis as ‘the diagnosis of an abnormality that is not associated with a substantial health hazard and that patients have no benefit to be aware of’.
Thanya Pathirana and colleagues, in their paper titled Mapping the drivers of overdiagnosis to potential solutions, traced the root causes of overdiagnosis to five ‘domains‘. The first domain they identified was the ‘culture of too much medicine‘, a phenomenon that is driven by the belief that more tests and screening are good, and that is amplified by ‘fears of uncertainty, ageing, death, and disease’.
The second domain consists of health system drivers which, by ‘expanding disease definitions‘, end up categorising more people as having a disease. Chiolero and colleagues gave the examples of diabetes and hypertension as diseases where the boundaries between risk and disease have become blurred, and where the threshold for diagnosis has progressively been lowered; these factors, they asserted, have resulted in what they called an ‘epidemic of diagnoses‘. Bodersen referred to this concept as disease mongering, explaining that this arises ‘because it is cheaper and faster to invent new “diseases” than new pharmaceutical drugs’. Pathirana and colleagues added that the financial incentives that physicians receive for investigating and treating more patients contributes to the problem.
The third domain discussed by Pathirana and colleagues encompasses technology drivers such as the increasingly sensitive tests in use today. The authors argue that by detecting minor incidental abnormalities of uncertain clinical significance, these technologies inevitably lead to overdiagnosis. According to Chiolero and colleagues, these technologies also result in the treatment of asymptomatic individuals with the attendant risks this entails.
In the fourth domain are patient and public drivers, and central to this are the ignorance of the public on the ‘limits to, and harms of, medicine’, and their hopeful expectation that doctors should ‘do something‘. Chiolero and colleagues also added the public’s increasing ‘need for reassurance‘ as an important factor in this context.
The final domain of overdiagnosis that Pathirana and colleagues discussed is the professional, and this revolves around how the fear of missing diagnoses encourages doctors to practise defensive medicine and request excessive tests.
Overdiagnosis is unfortunately a threat to patient safety and it engenders a variety of perils. For example, Brodersen pointed out that the treatment of overdiagnosed conditions exposes patients to the physical risks of the intervention without influencing its prognosis. Brodersen also emphasised the psychological harm of overdiagnosis that results from false positive screening tests for incidental lesions. Choliero and colleagues stressed this point further when they said most lesions detected incidentally during screening are indolent or slow growing, and they ‘would have never become symptomatic before the death of the screened individual’.
Beyond the physical and the psychological hazards of overdiagnosis are its financial implications. This was addresed by Russell Harris and colleagues in their paper titled The harms of screening: a proposed taxonomy and application to lung cancer screening. In this article, they explored the medical, societal, and work-related economic losses, as well as the mundane hassles and inconveniences that are associated with overdiagnosis.
In their approach to mitigating the problem of overdiagnosis, Chiolero and colleagues recommended the following preventative measures:
- Raising awareness of the problem to healthcare professionals and patients
- Avoiding unnecessary diagnostic tests
- Avoiding screening without demonstrated benefits
- Establishing risk factor thresholds for screening and diagnostic tests
- Targeting screening to at-risk population groups
- Framing risks in absolute rather than in relative terms before screening or testing
- Adopting shared decision-making with patients before screening or testing
- Using terminologies that do not overemphasise the pathology of low-risk lesions
In their own recommendations, Pathirana and colleagues focused on measures which may help to limit the impact of professional and technology drivers of overdiagnosis. In this regard, they advocated for ‘more rigorous evaluation of the effects of both new and existing diagnostic technology on health outcomes’, and for ‘stronger regulation of the advertising of new tests and treatments to the public and health professionals’. They also encouraged ‘paying greater attention to managing and reducing conflicts of interest with industry’, and for professionals to be rewarded for the quality of the care they provide, and not for its quantity. Just as they advocated for the inclusion of overdiagnosis in medical education curricula, Pathirana and colleagues also urged medical practitioners to always consider less aggressive treatment options such as watchful waiting and active surveillance.
In the next post, we will look at the remedies of diagnostic failure.