Brave New World of Genetic Testing

In a provocative piece in the NY Times last week, science writer Gina Kolata suggests that the long term care insurance industry may be in a “death spiral.” The culprit, she argues, is genetic testing, which got a boost last month when the FDA approved testing by the company 23andMe. Previously best known for providing genealogical information to those who send in a saliva sample and a $99 fee, the personal genomics company is now authorized to provide information about genetic risk factors as well—for only an additional $26.

One of the ten conditions about which companies may offer information is Alzheimer’s disease. And perhaps the best established genetic risk factor for late-onset Alzheimer’s disease (sometimes called LOAD) is apoE. A gene that codes for a protein involved in cholesterol metabolism, apoE comes in three varieties, prosaically named apoE2, apoE3, and apoE4. Everyone has two copies of the apoE gene, so there are six possible genotypes, of which the most common are E2/E2, E3/E3, E4/E4, E2/E3, and E3/E4. The majority of people (63 percent, in one study of the distribution of the allele in 9 different populations) are E3/E3.

But while there is no way to definitively predict who will develop Alzheimer’s disease, fully 40 percent of people who develop LOAD are among the 25-30 percent of people who carry the e4 variant of the gene. And roughly half the residents of nursing homes have Alzheimer’s disease. As a result, according to spokesmen from the industry, a growing number of people are seeking testing from 23andMe to see if they have the e4 gene. If they have it, they buy long term care insurance. If they don’t, they take their chances.

If this trend continues—and as of 2017, 2 million people have obtained the direct to consumer genetic analyses---the pool of people buying long term care insurance could be heavily weighted towards those who actually will develop the disease. The health insurance industry, however, as consumers are perhaps finally coming to understand in the ongoing Obamacare wars, depends on pooled risk; it only works if the people who buy insurance include some people who will get sick and others who won’t. 

The reason the direct-to-consumer marketing of genetic information potentially spells doom for the long term care industry is that Americans are protected by the Genetic Information Nondiscrimination Privacy Act, which prevents insurers from requiring gene tests or using the results of genetic testing in coverage decisions. That means that ordinary people, without any physician input, can find out if they are at high risk, they can make decisions about buying coverage based on that assessment—and the insurance companies are powerless to intervene. If most of the people who are destined to get Alzheimer’s disease end up with insurance, the insurance company will end up paying out a lot more than they originally anticipated—leading to enormous increases in the rates or bankruptcy of the industry.

Now there are plenty of other reasons that the long term care insurance industry may collapse, and a good number of pre-existing reasons why it’s a poorly designed program. The amount of money it actually provides people is rarely enough to cover their actual costs, whether of home or institutional care. There are many barriers in the way of people using their benefits—for example, many policies require three months of disability before they kick in, which may be three months too long, especially if they are only going to be in a nursing home for three months. And Medicaid is currently available as a back up to pay for institutional care, provided people have “spent down” their personal savings, so the value of long term care insurance derives from its ability to shelter assets.

Analyzing the value of long term care insurance is a conversation for another blog post. But the point I want to make today is that before you rush out and get tested for Apo E, you should be aware of the limited predictive value of the test. While the likelihood of getting LOAD if you are one of the 2.6 percent of the population who have two copies of E4 is 91 percent, the likelihood of getting LOAD if you are one of the 22 percent of the population with one copy of the E4 allele falls to 47 percent. And if you are one of the 76 percent of the population with no E4 alleles, you still have a 20 percent of getting Alzheimer’s. Because the poor “negative predictive value”—because even with a negative test, you have a substantial risk of getting the disorder—physicians and organizations such as the Alzheimer’s Association have for years recommended against routine Apo E screening.

Interestingly, Apo E determination has been available to Europeans as a direct to consumer test for years. A study to find out what the short- and long-term psychological consequences were for patients with a positive test. Though the study only looked at people who requested testing, it found that there were no significant adverse consequences of getting bad news. So fear that you will become anxious or depressed is probably not grounds for resisting the impulse to be tested for Apo E4. But you should remember that it’s just a risk factor—and one over which you have no control—and plenty of people who test negative will still develop Alzheimer’s disease.