Winding back the harms of too much medicine.

Distinguishing the sick from the healthy has always been a fundamental challenge for medicine. A chief concern has been to guard against missing disease, with the focus on problems of underdiagnosis and undertreatment. Yet with the modern technological expansion of healthcare in rich developed nations, sceptical voices have long warned of the flipside—too much medicine. Mounting evidence about the threat to human health from overdiagnosis, and the harms and waste from unnecessary tests and treatments, 5 now demand that we meet one of this century’s key challenges: how to wind back medical excess, safely and fairly. In 2002 the BMJ published a theme issue called “Too Much Medicine?” with articles on the medicalisation of birth, sex, and death, among other aspects of ordinary life. Its opening editorial wondered whether doctors could become pioneers of de-medicalisation, handing back power to patients, resisting disease mongering, and demanding fairer global distribution of effective treatments. A decade later, as data on overuse and overdiagnosis mount, the BMJ announces a “Too Much Medicine” campaign—this time without the question mark (www.bmj.com/too-much-medicine). Through the campaign the journal plans to work with others to highlight and contribute to the growing evidence base on overdiagnosis and overtreatment. In October last year, for example, a major inquiry cited evidence based estimates that as many as one in five women given a diagnosis of breast cancer as a result of screening would not have been harmed by that cancer. In December, the chair of the DSM-IV (Diagnostic and Statistical Manual of Mental Disorders, fourth edition) task force warned of the potential for “massive overdiagnosis and harmful overmedication” with increasing expansion of the definitions of mental disorders—for example, turning the physical symptoms of cancer or heart disease into a mental disorder “somatic symptom disorder”—in the forthcoming fifth edition of the manual. Data suggest that overdiagnosis exists to some extent across a range of common conditions, including prostate and thyroid cancers, asthma, chronic kidney disease, and attention deficit hyperactivity disorder. Indeed, overdiagnosis may be the norm rather than the exception. This matters because once people are labelled with a diagnosis, a cascade of medical, social, and economic consequences follows—some ofwhich are permanent. The medical label and the ensuing treatment take an emotional and financial toll on the person, while also costing the health system. Importantly, overdiagnosis and underdiagnosis coexist in many healthcare settings, both rich and poor. Concern about overdiagnosis of chronic kidney disease in older people exists alongside evidence that some groups disproportionately experience avoidable harm from serious kidney disease. Because of this and other uncertainties, it will not be easy to communicate effectively about overdiagnosis with professionals and the public. The concept is unfamiliar and counterintuitive to many people. The BMJ’s campaign must be seen as part of a larger effort to combat over-medicalisation, including the “Choosing Wisely” campaign, run by a coalition of US medical specialty societies to combat the overuse and misuse of tests; the recent Avoiding Avoidable Care conference, run by the Lown Cardiovascular Research Foundation (http://avoidablecare.org/); and the second Selling Sickness conference, held last week inWashington DC, which brought together academic critics of the drug industry, health reformers, consumer advocates, and journalists (http:// sellingsickness.com/). Although each has specific concerns and recommendations, all share the same goal: implementing practices that help people who are sick and do not harm those who are well. In addition, the BMJ is a partner in an international scientific conference on preventing overdiagnosis, to be held in September this year in Hanover, New Hampshire, USA. It is hosted by the Dartmouth Institute for Health Policy and Clinical Practice, and