Alzheimer's Tests and Insurance

Word of a test of spinal fluid for Alzheimer's Disease has started discussions of whether one should get a test for an incurable disesase. I wrote a paper with colleagues from Boston University, University of Michigcan and Duke that looked at the potential impact of a genetic marker for AD on adverse selection in private long term care insurance. Basically, we confirmed a story that was consistent with adverse selection whereby persons with the genetic marker were more likely to actually use nursing home care (the most expensive form of long term care) and that their knowledge of their genotype was associated with them being more likely to purchase insurance.

The paper has a discussion of the issues surrounding the use of genetic markers as underwriting/risk adjustment tools. The logic of the Genetic Information Non-Discrimination Act of 2008 (GINA 2008) is to ban the use of genetic markers for such purposes. Interestingly, GINA 2008 doesn't ban the use of genetic markers for use in writing long term care insurance policies, while it does ban the use for virtually all other types of insurance. It is unclear to me if this was purposeful, or just an oversight.

Our paper concludes with a discussion of the issues around the potential use of genetic markers to underwrite/risk adjust long term care insurance policies and premiums (which is not currently done). While the reflexive instinct of many is to say this should be banned, if the public policy goal is to expand the rate of coverage by long term care insurance policies, it is not clear that is the correct intuition. It really boils down to whether the goal is actuarial fairness, or a more general notion of moral fairness (you can' control your genes).

The spinal fluid test for AD is not a genetic marker, but the presence of a test that could predict the onset of a debilitating disease such as AD will likely make it more difficult for long term care insurance coverage to expand unless the test is directly used, because you would expect insurers to assume applicants have inside information, to raise prices for all comers, further driving out better risks (at least with respect to a test like this).

The best policy to insure agaisnt long term care is to spread risk broadly, which means to get as many persons into an insurance market as possible. With insurance that is rare, the adverse selection risk is great, and anything that provides information to only one side of this type of market is likley to hinder uptake of this type of insurance, which only occurs when a buyer and seller of an insurance policy can agree on a price/premium.