Overdiagnosed: Making the Case Against Finding Too Much

Overdiagnosed: Making People Sick in the Pursuit of Health was published by Beacon Press in 2011. Its authors are H. Gilbert Welch, Lisa Schwartz and Steven Woloshin. All three are physician-researchers. They spent much of their careers studying the outcomes of medical testing. Cancer screening was a particular focus.

The book is not a memoir. It has no single patient at its center. It is an argument, built case by case across several conditions. Modern medicine drives to find disease earlier and more thoroughly. That drive has a real cost. It finds abnormalities that meet a technical definition of disease. But those abnormalities would never have caused a symptom. Nor would they have shortened a life.

The authors distinguish this carefully from misdiagnosis. A misdiagnosis is wrong. An overdiagnosis is, by their definition, correct. It is a real cancer, correctly identified under a microscope. It simply would never have grown fast enough to matter. In some cases it would never have grown at all. Either way it would not have mattered in that person's remaining lifetime. The problem is not the test's accuracy. The problem is what happens after an accurate test finds something. Biology was never going to act on it.

The book's title captures its overall claim. The pursuit of ever-earlier, ever-more-thorough detection has a cost. It can turn healthy people into patients. All the burden of treatment follows. And no one's life is extended.

The book is organized around a series of case studies. They are drawn from different areas of medicine. Cancer screening is prominent among them. Chapters examine breast cancer screening. Another examines prostate cancer screening, through PSA testing. The book also covers other conditions beyond cancer. Together they build a general argument. It applies across specialties, not only in oncology.

Each case follows a similar structure. What does the screening test find? What fraction of that would never have caused harm if left alone? And what happens to people treated for a condition that was never going to cause symptoms? The prostate cancer chapter is one of the book's most detailed. It draws on PSA screening data available at the time. USPSTF notes that overdiagnosis is common with PSA screening. It estimates 20 to 50 percent of screen-detected prostate cancers are overdiagnosed.

The book also lays out the statistical concepts behind its argument. It does so in accessible terms. Lead-time bias is one. Early detection makes survival time look longer. It does not change when death occurs. Another is the difference between disease incidence and disease mortality. Testing can inflate incidence. Mortality is the number that actually shows whether more testing saves lives.

There is a closing section on what better screening could look like. It would be more selective, and grounded in evidence. The section also covers how a patient might think about a screening decision. Overdiagnosis becomes a real, countable harm rather than a hypothetical one.

The book made a central conceptual contribution. Overdiagnosis is a real, distinct, and measurable harm. It is not a semantic quibble. That claim has held up. It is now part of how federal agencies describe screening tradeoffs. That is discussed below.

The book uses mortality data rather than survival statistics alone. That is the right way to judge whether more screening saves lives. The method is sound. It reflects a real, ongoing debate within evidence-based medicine. That debate is about how screening programs should be evaluated.

The authors write clearly. They build a statistical case against overtesting. They do not dismiss the emotional weight of a cancer diagnosis. The book does not read as cold, or dismissive of patients' fear. That is a real risk for books making this kind of argument.

The book is now more than fifteen years old. Screening guidance has changed materially in several areas it discusses. USPSTF revised its breast cancer screening recommendation in 2024. It now recommends mammography every two years from age 40. Before that, USPSTF advised women in their 40s to decide with their clinician. The older guidance was in place when the book was written. So any specific screening age or interval in the book reflects 2011 guidance. It is not current guidance.

The book's argument carries careful qualifications. Taken in isolation, it can be misread. It can look like an argument against screening in general. That is not what the authors argue. Their case is for more selective, evidence-based screening. It should be calibrated to where evidence shows benefit outweighs harm. It is not a case for abandoning screening. A reader should be careful here. The book critiques certain tests, in certain populations, at certain intervals. Those critiques are not a blanket dismissal of screening as a category.

The book also predates part of the current debate. That debate concerns screening in specific higher-risk populations. Expanded genetic risk assessment is part of it. It can change the benefit-harm calculation for an individual. The book's population-level analysis does not fully capture that.

This book is itself a useful model for evaluating claims about screening. That holds for cancer and for other conditions. Its method is worth naming explicitly.

The book has a framework for a strong claim about a screening test. Such a claim separates disease incidence from disease mortality. It also states clearly what fraction of detected cases would have caused harm if left alone. It states what fraction would not have. A weak claim treats "found early" as the same as "lives saved." It skips that distinction.

A strong claim also names the population the evidence applies to. The benefit-harm balance of a screening test differs by age. It differs by risk factors and other individual circumstances. USPSTF's own grading system exists for that reason. It makes those population-specific judgments explicit and current. A weak claim generalizes across age groups. It does not acknowledge that the same test can differ by group. The balance can be favorable in one age group and unfavorable in another.

Readers can apply the same test to the book's own claims. Check whether a specific figure it cites still matches USPSTF's current recommendation. Guidance in several areas the book covers has been revised since 2011.

USPSTF recommends screening mammography every two years for women aged 40 to 74. That is a Grade B recommendation, published in 2024. It replaced the earlier advice that ages 40-49 should decide individually. That earlier guidance is what the book was responding to.

For prostate cancer, USPSTF's current recommendation reflects the book's concern directly. For men aged 55 to 69, PSA-based screening gets a Grade C. That means the decision should be individualized. It is not offered as a blanket recommendation. The potential benefit is modest. USPSTF cites about 1.3 prostate cancer deaths prevented per 1,000 men screened over 13 years. USPSTF lists this as its final 2018 recommendation and says an update is in progress. Considerable potential harms come with it. Those include false positives and overdiagnosis. They also include treatment complications such as incontinence and erectile dysfunction. For men 70 and older, USPSTF recommends against PSA-based screening entirely. That is a Grade D.

NCI has its own breast cancer screening page, current as of this writing. It states the overdiagnosis concern in nearly the same terms the book uses. Some cancers found during screening may not have caused problems during a woman's lifetime. Those would not have needed treatment. NCI also covers false-positive results. It says those can cause emotional distress. They can also require further procedures such as biopsies.

The convergence is notable. The book argued against unchecked expansion of screening in 2011. Federal guidance now describes the same tradeoff. The language is similar, even where the specific numbers have moved. For the fuller current picture, see cancer overdiagnosis and breast cancer screening. See also false positive cancer screening results. That page covers the anxiety and follow-up burden the book describes.

The book has one measurable legacy. So does the broader body of research its authors contributed to. It is visible in how USPSTF itself now writes its recommendations. Current USPSTF statements routinely include an explicit discussion of "harms of screening" alongside benefits. USPSTF's prostate statement, for example, gives overdiagnosis estimates. It also gives different grades for different age ranges. That is the kind of open weighing of harms and benefits the book argued for. We cannot say the book caused this change. Still, the match between the book's argument and current federal guidance is easy to check for yourself.

This book suits a reader with one question in mind. Why is "more testing" not automatically the same as "better care"? Such a reader should be at ease with statistical reasoning in plain language. The book is particularly useful for a reader facing a screening decision. It helps to understand overdiagnosis before that conversation. It is not a replacement for the conversation.

It does not suit a reader looking for a current screening recommendation. Several of the specific guidelines it references have since changed. It also does not suit a different reader. Some want the book to settle whether they, personally, should be screened. It cannot do that. That decision depends on individual risk factors. Current guidance is built to weigh those. It is a decision to make with a clinician who knows that person's history.

Cancer Explained runs free of charge and free of ads; donations are what keep it that way, if you're able.

• USPSTF — Breast Cancer: Screening
• USPSTF — Prostate Cancer: Screening
• NCI — Screening for Breast Cancer

This page discusses Overdiagnosed for education. It is not medical advice, and nothing here is a judgement of anyone's real medical care.