Korea's thyroid cancer epidemic followed income upward. So did the correction, and the deaths did not

For almost every serious disease, being poorer means getting it more often and dying of it more often. Thyroid cancer in South Korea broke the first half of that rule and kept the second.

The pattern, and the control that makes it legible#

Researchers at Seoul National University combined the Korea Community Health Survey, the National Health Insurance database and cause-of-death records from Statistics Korea to track screening, incidence, postoperative complications and mortality by income between 2006 and 2015.

Incidence over that period rose from 46.6 per 100,000 in 2006 to 115.0 in 2012, then fell to 63.5 by 2015. Mortality did not fluctuate.

What turns that from a curiosity into evidence is the negative control. The same team ran the same analysis on lung cancer and stroke, two conditions where nobody expects a detection artefact. For both, higher income went with lower incidence and lower mortality — the ordinary direction. Thyroid cancer inverted it for incidence while keeping it for death.

The correction reached the same people first#

The sharpest finding is about what happened after Korea began to worry publicly about overdiagnosis. Incidence fell, and so did postoperative complications — and both falls were more pronounced among higher-income people than lower-income people.

The same group that was first to be over-examined was first to stop being over-examined. Whatever the mechanism — better information, more access to a clinician willing to counsel against a scan, more capacity to decline one — the benefit of the country's second thoughts arrived unevenly, in the same direction as the harm had.

That pattern recurs downstream. In the Korean multicentre cohort we reported earlier this week, higher income independently predicted choosing active surveillance over immediate surgery for a low-risk microcarcinoma, with an odds ratio of 1.79. Income shapes who is over-diagnosed, how quickly a population stops over-diagnosing them, and who is offered the chance not to be operated on afterwards.

How the scans actually happened#

Korea has never had a national thyroid cancer screening programme. The scans were opportunistic, and a separate study identifies one of the routes.

Analysing the 2014 Korean National Cancer Screening Survey, researchers looked at 2,270 women over 30, of whom 569 — a quarter — had been screened for thyroid cancer. What predicted it was whether they had been screened for breast cancer.

Mammography roughly halves the effect that ultrasound produces, which points at the machine rather than at health-consciousness in general: a woman attending for a breast ultrasound is already lying down with a probe on her upper body. The study's authors conclude that thyroid screening in Korea is performed on an opportunistic basis and call for a national initiative to curb it in asymptomatic people.

What these studies cannot show#

1. Both are observational. The income study is a time-trend analysis of population rates, which can establish that patterns move together and cannot establish that one caused the other in any individual.
2. Screening behaviour in both studies is self-reported through national surveys, and people misremember whether a scan they had was of their thyroid.
3. The breast screening study is a single cross-section from 2014 with 2,270 respondents. It describes an association at one moment, not a trend, and cannot show that the breast appointment produced the thyroid scan.
4. Higher postoperative complications among higher-income patients follow from higher rates of surgery. That is a consequence of being diagnosed more, not evidence that their operations were worse.
5. Higher mortality among lower-income patients is not explained by these data. Later presentation, comorbidity and treatment access are all candidates and none is tested here.