From the Frontline

Does early diagnosis really save lives?

BMJ 2012; 344 doi: http://dx.doi.org/10.1136/bmj.e4252 (Published 19 June 2012)
Cite this as: BMJ 2012;344:e4252
  1. Des Spence, general practitioner, Glasgow
  1. destwo{at}yahoo.co.uk

Language is important. For many years the popular press clamoured for more screening, because detecting a disease early is always better. We couldn’t get enough of health checks. But now we are falling out of love with screening, with reports and evidence of unnecessary interventions, overdiagnosis, and medical harm. Many of us puff out our chests in self satisfied vindication, but the intuitive and near religious belief that early diagnosis is a good thing has not gone away: government and charities now lobby to promote the early signs of a disease and to “educate” primary care doctors. Early diagnosis is everywhere—for example, the National Awareness and Early Diagnosis Initiative and the “three week cough” campaigns in cancer.1 2 Other lobbyists press for early diagnosis of rheumatoid arthritis, mental illness, hypertension, diabetes, dementia, and the rest. Early diagnosis is portrayed as wholesome, clean cut, apple pie medicine: who could disagree with such an agreeable concept?

But early diagnosis is simply screening rebranded. The logic that if you spot the condition early you can improve the outcome is exactly that for screening. But unlike in screening, the concept of early diagnosis generally has no intellectual framework, rigour, research, or systematic reviews. The danger symptoms of cancer, such as cough and bloating, often come from small, confounded case-control studies. These have no measurable positive or negative predictive values for the general population. In the real world these symptoms are so common as to be useless as predictors of early disease. Worse, having “symptoms” tends to suggest incurable late stage diseases. So it is with the research purported to support the current lung cancer and cough campaign,2 which had no impact on the presentation of treatable lung cancer.3 Also, specifically for ovarian cancer and prostate cancer, studies show that even systematic screening for asymptomatic disease doesn’t work.4 5 As for other chronic diseases such as diabetes, we have no evidence that promoting early diagnosis through symptoms affects any hard clinical outcomes.

Big pharma loves early diagnosis: it makes more patients who take more profitable drugs for their whole life. The other certain outcomes of early diagnosis are more testing, more referrals, and more interventions. In turn, these generate more non-progressive disease, “incidentalomas,” and overdiagnosis—impaling anxious patients with yet more fear and non-disease pathology, while the medical fatalists happily ignore all warnings. The thoughtless and emotional language of “early diagnosis” is in reality just bad screening and the worst science.

Notes

Cite this as: BMJ 2012;344:e4252

Footnotes

References

THIS WEEK'S POLL