Tuesday, August 14, 2018

Overdiagnosis in lung cancer screening: don't tell, don't ask?

- Kenny Lin, MD, MPH

Although the U.S. Preventive Services Task Force recommended in 2013 that current and recent smokers 55 to 80 years of age with at least a 30 pack-year history receive annual low-dose CT screening for lung cancer, family physicians have been slow to implement this recommendation in their practices. Concerns about this screening test include the quality of the supporting evidence (which the American Academy of Family Physicians judged to be insufficient) and potential harms, including overdiagnosis and overtreatment of tumors that, left undetected, would never have caused symptoms during a patient's lifetime. An analysis of the National Lung Cancer Screening Trial (NLST) suggested that one in five lung cancers were overdiagnosed. In recognition of the balance of benefits and harms of lung cancer screening, the Centers for Medicare & Medicaid Services requires that eligible patients first have a "counseling and shared decision making visit" with a clinician that utilizes a patient decision aid prior to undergoing a scan.

A previous study of screening for other cancer types found that clinicians mentioned overdiagnosis as a potential harm less than 10 percent of the time. Are lung cancer screening discussions any different? In a study published this week in JAMA Internal Medicine, researchers evaluated shared decision making (SDM) using the validated Observing Patient Involvement in Decision Making (OPTION) scale in a sample of transcribed physician-patient conversations. Relative to the mean total visit length (just over 13 minutes), physicians spent a mean of 59 seconds discussing lung cancer screening. None of the conversations mentioned decision aids, and the mean total OPTION score was 6 out of 100 (where 0 indicates no evidence of SDM and 100 indicates SDM at the highest skill level), reflecting that physicians rarely informed patients about harms of low-dose CT scans or asked patients how they valued these harms.

This lack of attention to harms of lung cancer screening is concerning because the magnitude of overdiagnosis may be considerably higher than previous estimates. Researchers recently analyzed data from the Danish Lung Cancer Screening Trial, in which participants underwent 5 annual low-dose CT screenings (compared to 3 in the NLST) and concluded that two-thirds of lung cancers were likely overdiagnosed. In an accompanying commentary that compared the methods used to estimate overdiagnosis, AFP Deputy Editor Mark Ebell, MD, MS and I stressed the importance of communicating with patients about this "often underappreciated harm of screening":

Patients can make informed choices about low-dose CT only if practitioners fully disclose all the potential harms of screening, including the risk of overdiagnosis. It will be important to researchers to continue to refine estimates of lung cancer overdiagnosis, allowing physicians to provide more accurate information to our patients.

To best serve patients, primary care physicians and pulmonologists must do better than 59-second conversations about lung cancer screening that only mention potential benefits. We need to take the time to tell patients about harms such as overdiagnosis, and ask them how they value these harms relative to the benefits, before ordering the scan.