A Dec. 17 Wall St Journal article outlines how data mining is being used to identify persons with selected illnesses. While the objective in this case is to find patients for clinical trials, the same approaches are used to target folks for ads, and potentially could influence heath, long term care, or life insurance. For anyone close to the U.S., it would be hard to avoid the political football being played with health insurance over the last few years. But there is a fundamental question about the nature of insurance in a data mining world: is it no longer ethical?
First you need to really consider what insurance is. If you have 1000 homes in a town, one burns down each year, the average replacement cost is $100,000 then a premium of $100/year is the minimum for an insurer to break even. But most insurance companies are “for profit”, and do not want to break even. Actually, they do best when they don’t pay out benefits at all. And here we see the essential tension between the insurer and the insured. Is it any surprise that when a new health care law takes effect in the U.S. that health insurance companies are dropping their higher risk patients? This is what for-profit means! When you can reduce your risks and raise your revenues you deliver profit.
But the WSJ article points out a disturbing trend — data mining to isolate affected individuals. The article points out that the records used; credit card records, online shopping, search preferences, TV viewing habits, vacation styles, cars driven are part of the mix. “We are now at a point where, where based on your credit-card history, and whether you drive an American automobile and several other factors, we can get a very, very close bead on whether or not you have the disease state we’re looking at.” Roger Smith, Sr. VP of Acurian. (WSJ)
If Insurance companies can sort out higher risks, then the amortization of risk over large populations can be eliminated. In effect the insurance premium becomes a pre-payment of future expenses (plus overhead and profit.) Increasingly insurance companies are following this path. Flood insurance (historically supported by US Government reinsurance) is rising in price dramatically, and home-owners in wildfire risk areas are being dropped, potentially if their neighbors have not cleared their properties. It is easy to argue that folks who build glass houses should have higher ‘stone’ insurance rates, but as this moves from a large “pool” to a “pool-of-one” the concept of insurance disappears.
With the emergence of full genome mapping, individual life style tracking, big data and data mining, the profit opportunity will drive insurance companies to sort customers as accurately as possible. The result may be the elimination of insurance, at least in certain areas, as a way to manage risk.
An informative counter-point is the U.S. medicare system, this spreads premiums over all employed persons and insured persons, and does not discriminate by preexisting conditions or the more advanced technology that is emerging. Of course it is a not-for-profit government program which is yet another political football.
Has your insurance been affected by these processes? … Don’t comment if so, this blog is mined!