A team of researchers at Georgetown University and six other medical centers has developed a simple blood test  they say can predict, with 90 percent accuracy, whether an individual will develop Alzheimer’s Disease within 2-3 years.  If it works as advertised, such a test could have a profound impact on the long-term care insurance market.

The study results, published in the journal Nature Medicine, are very preliminary. If the test works as hoped, it has enormous potential benefits, including the ability to advance efforts to find a treatment for the disease. If it makes early intervention possible, this test could dramatically reduce long-term care costs.

But it also has a powerful unintended consequence: It could destroy private long-term care insurance or even any future voluntary government insurance program. A widely available test to predict Alzheimer’s would make any form of voluntary long-term care insurance impossible. “It would be a huge game changer,” one insurance actuary told me.

Here’s why: If you take the test and find you are almost certain to develop cognitive impairment, you are far more likely to buy long-term care insurance.

If insurers have access to the test results, they will either deny you coverage or charge you much higher premiums than someone whose results are negative.

But even if insurers are not allowed to see the results (and federal law already prohibits health insurers from demanding results of certain medical tests), a cheap and accurate Alzheimer’s test still could sink the voluntary insurance model.  Insurers would simply assume that buyers know they are likely to contract the disease and they’d raise premiums sharply to account for that. The result: the classic insurance death spiral.

Already more than half of all LTC insurance claims are for cognitive impairment. And surveys show that those who think they are going to contract Alzheimer’s (even without an accurate test) are far more likely to buy than those who don’t.

There are huge uncertainties here. First, the Georgetown research was very preliminary with a small sample. Much more needs to be done before the test can be commercialized. Second, 2-3 years advance warning is probably not enough to make much difference for insurance buyers. People usually purchase in their 50s or early 60s and most won’t show symptoms of Alzheimer’s for 20 years.

However, Howard J. Federoff, executive vice president of health sciences at Georgetown University Medical Center and an author of the study, says his team wants to know if their test can identify biological changes even sooner. In a recent interview with the radio program Science Friday, he said his group had already begun designing that follow-up study.

The test identifies changes in blood cells that predict development of cognitive impairment. Other researchers are looking at ways to identify whether an individual has a genetic predisposition to cognitive disease. If any of those tests pan out, the impact on LTC insurance would be even more profound since genetic markers could be identified at a very young age.

For now, this new research promises a potentially powerful diagnostic tool.

In the best case, it would help others develop an effective way to treat—or even prevent—Alzheimer’s. Until now, drug therapies aimed at treating this disease have been deeply disappointing. But Federoff wonders whether that’s because these drugs are being given too late. He asks whether patients would benefit by receiving the meds before they start showing symptoms. His test could make that possible.

But there is a potentially profound downside. What happens if people learn they are fated to suffer cognitive impairment and, as a consequence of that knowledge, lose access to one of the few resources available to pay for what they know will be very costly care? As research into these sorts of predictive tests continues, it is a question we need to keep asking.