DMS profs. express concerns for overdiagnosis of cancer

Dartmouth Medical School professors H. Gilbert Welch and William Black explored a relatively new term in the medical dictionary: “overdiagnosis.” Their paper was recently published in the Journal of the National Cancer Institute.

Welch and Black are on faculty at Dartmouth Medical School and The Dartmouth Institute for Health Policy and Clinical Practice.  Welch currently teaches Public Policy 26 – a new course on health policy in its inaugural offering – at Dartmouth College, and he has incorporated his research on early detection and screening into its syllabus.

What exactly is overdiagnosis?  In short, it is the diagnosis of a disease in an early form that will not change a patient’s life expectancy or quality of life. This can lead to unnecessary treatments or procedures.

Welch and Black walked through different types of cancers and identified the negative effects in screening for asymptomatic patients.  For example, in a 10-year randomized trial of two groups of women in Malmö, Sweden, 741 breast cancers were detected in the mammography group – 447 by screening and 264 clinically – as opposed to 591 in the control group.

Fifteen years after this time-period, 35 more cancers were detected in the control group; these “catch-up” cancers still leave a gap of 115 “extra cancers” in the mammography group (a gap of 150 initially, minus 35 more detections).  This may mean that 115 out of 741, or 16%, were “overdiagnosed.”

Over the last 35 years, the rate of new diagnoses has significantly increased for thyroid, prostate, breast, kidney, and skin cancer, more than doubling for thyroid cancer (4.9 per 100,000 to 10.6 per 100,000) and nearly tripling for melanoma.  However, the death rates have remained relatively constant – or even decreased – for each cancer.

Why is the rate of detection increasing?  First, not all physicians have the same subjective interpretation of what exactly entails a threatening form of cancer, due to drawing different threshold lines.  Second, the data varies from randomized trial to trial, which will certainly impact a health care provider’s decision.  Third, the patient demand for testing is increasing, due to increased access to medical knowledge on the Internet or other media.

“There is a spectrum of abnormality,” Welch remarked in an interview with the DUJS, “and lowering the threshold for diagnosis has the potential to harm [the patient].”

The threshold for abnormal is not always clear, such as a PSA (Prostate Specific Antigen) count of greater than 4 ng/mL when determining whether a patient has prostate cancer.

Welch and Black note that “individuals below the threshold can still have cancer,” such as a patient with a PSA of 2.5 ng/mL, but they write that “it may be better to simply ignore small abnormalities.”

They conclude the article with hopes to incorporate “formal coursework on how to evaluate screening tests and how to recognize overdiagnosis” in medical schools.  Rapidly advancing technology brings enthusiasm to the community, and moving forward, we need more long-term studies to determine at what stage of a particular subtype of cancer treatment should commence; such work is essential to the future of comprehensive cancer care.

Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *