Frances1 1fedc3

Warning: over-diagnosis can seriously damage your health!

Whilst in many less developed countries there is a paucity of diagnostic testing and appropriate therapies, we in the West are suffering from the ‘modern epidemic’ of over-diagnosis and over-treatment. Today’s highly sensitive biomarker and imaging tests increasingly identify asymptomatic or very mild conditions that if left untreated would not cause symptoms or reduce longevity. A recent report on mammography screening in the UK suggested that 19% of breast cancers were over-diagnosed, and a US task force concluded that PSA-based prostate cancer screening over-diagnosed up to 50% of tumours. Other over-diagnosed and over-treated conditions include thyroid cancers as well as a range of cardiovascular diseases, chronic kidney disease and ADHD. At best treating such subjects is an imprudent use of health service funds; at worst ‘patients’ suffer both psychological and physical harm from their diagnosis and subsequent treatment. Of course effective screening for cancer and other serious conditions is vital, but how can the problem of over-diagnosis be at least alleviated when tests (and cut-off values) must be sensitive enough to detect pathologies that really require treatment?
When diagnostic tests are evaluated for accuracy the average sensitivity and specificity are reported. But of course individuals vary, and diseases have stages of severity. What is needed is the identification of those patients for whom treatment will do more good than harm. Similarly average results in therapeutic trials may be positive, so negative effects in some patient groups are not evident, but again the potential benefit of a treatment should be weighed against possible harm according to disease severity. And subjects being screened should surely be informed about the risk of over-diagnosis. Yet in a recent random sample of 500 Australians, only 10% of the women who had had mammography, and 18% of the men who had had prostate cancer screening reported that they had been told about the limitations of these tests.
There is also an urgent need to scrutinize the panels of medical professionals setting disease definitions. Diagnostic thresholds are frequently lowered without considering the balance between good and harm of treating the additional patient group who have a lower risk or milder symptoms. And although it may sound cynical, panels with three quarters of the members having multiple ties to pharmaceutical companies – some of which will directly benefit from an increased number of patients with the disease under discussion – surely can’t be unbiased!
Hopefully appropriate action can be taken before the seemingly inexorable trend towards over-diagnosis makes patients of us all!