Cut the out-of-pocket cost of an MRI and almost everyone wants one — including for the headaches that do not need it

Korean health policy has spent a decade extending insurance coverage to services patients previously paid for themselves, magnetic resonance imaging among them. The intended effect is that people who need a scan are not deterred by its price. A study designed to test the unintended effect asked a narrower question: what happens to demand for the scans that were never indicated?

How the question was put#

Published in the Journal of Preventive Medicine and Public Health, the study surveyed a proportional quota-based sample of 1,229 people in Korea using two hypothetical scenarios: uncomplicated headache, and non-specific low back pain. Both are standard examples of situations in which clinical guidelines advise against imaging when there are no warning features. Participants were asked whether they would be willing to have an MRI, before and after a reduction in what they would pay out of pocket.

Half of the participants first saw a version of the vignette that included a physician recommending the scan, so that the effect of price could be separated from the effect of a doctor's suggestion.

The equality result, read properly#

Before the price fell, income predicted willingness: the poorest quintile was the least likely to say they would have the scan. After it fell, that gradient largely disappeared. The slope index of inequality for low back pain fell from 8.15 to 5.37 percent, and for headache from 6.99 to 0.83 percent. The relative index moved from 1.20 to 1.06 and from 1.20 to 1.01 respectively.

Why low-value imaging is not merely a waste of money#

MRI delivers no ionising radiation, so the cost of a low-value scan is not the cost of a low-value CT. What it shares with every other broad look at a person who is well is the finding nobody was seeking. In a German population cohort, disclosing incidental findings from whole-body MRI raised the rate of biopsies that returned nothing malignant by more than it raised the rate returning cancer.

A scan of a normal lumbar spine in a middle-aged adult reliably shows disc degeneration, bulges and facet arthropathy, because those are near-universal with age. Naming them on a report in a patient with non-specific back pain converts an ordinary symptom into a structural diagnosis, and the literature on what follows — further imaging, injections, referral — is the reason guidelines say not to look.

What a vignette study can and cannot support#

These are stated intentions, not scans that happened. People asked hypothetically what they would do systematically differ from people in a consulting room, and the study measures potential utilisation rather than utilisation. It also cannot capture the other half of the transaction: a scan requires a doctor to order it, and physician behaviour under changed reimbursement is not what was surveyed.

What it does establish is a price elasticity for a service where price was the main brake. Whatever restraint existed in the lowest income quintile was financial, and removing the finance removed the restraint. Policies that expand coverage without a parallel mechanism for distinguishing indicated from unindicated use are, on this evidence, expanding both.