|Turned up to eleven: Fair and Balanced|
Thursday, June 13, 2002
I wanted to get back to science on this one, in particular the intersection of scientific and medical advancement in genomics with our current and potential future medical insurance system. At the outset, let me say this; I don't particularly want nationalized health care(single payer), although I am not completely opposed to the idea. HMO's, which were instituted to save an ailing, bulging health care system, have fundamentally undermined it instead, by putting profits ahead of health. I am not sure if I have discussed this before, but it is a somewhat off topic anyway. Anyone who cares to argue this point, please note I am not referring to for-profit medical device or pharma, which needs to have the profit motive to spur innovation (although, it should be noted that hundreds of thousands of University Professors, research M.D.'s, grad students and post-docs do a great deal of innovative medical research with a much reduced, but not eliminated, profit motive). For profit insurance companies do not have the patient's health as their primary responsibility (to do so would be a breach of fiduciary responsibility). In particular, publicly traded for-profit insurance companies must focus on the bottom line, to the utter exclusion of all else. As a patient population, we cannot allow this to go on forever. Undoubtedly, HMO's served to lighten the burden on state-controlled health plans, and introduced competition into the marketplace, but eventually, this runs (has run?) its course. In any event, this is not what I want to discuss in this post.
The interesting thing to think about, at the juncture of genomics and insurance, is how genomic technology (and proteomics, but in a different way), changes the game. In order to think about this, think about how health insurance is apportioned, and how a profit is made by an insurance company. Now, in auto, home, and life insurance, the rate you pay is tied intimately to personal factors (driving record, age, marital status, education for auto; location, value of home; health, age, risk factors such as alcohol and tobacco for life). For many (most? I haven't looked at the stats), however, health insurance is provided as a part of a benefits package at work, with no regard to risk related to a given employee (you probably have to pass a physical under some circumstances, although I and my family never have). The risk is distributed, over a very large potential patient population. Based on actuarial tables (essentially the aggregated risks based on population statistics), the company predicts how much it will have to pay, and then charges a bit more than that (what the market will bear, of course!). Fundamentally, this is based on poor information. With a little gaze into our crystal ball, we can predict what will happen to this system. (WARNING: Unwarranted, baseless guesswork and speculation ahead! Proceed with caution...)
So we come back to the fundamental question, What can genomics tell us? Well, in a limited sense, it can tell us exactly what the genetic makeup of any given individual is, down to the specific base pair in question (not right now, but within 3-5 yrs). Knowing that, it is much more likely that we can make good, accurate predictions about chronic disease manifestations, in particular the biggest killers in the developed world (the developing world can't get past malaria and TB), Cancer and Heart Disease. Some time in the future I will come back to Paul Ewald's remarkable, intriguing, but as yet unproved hypothesis that microbes are responsible for the lion's share of chronic disease as well as conventional infectious disease (it is known that some cancers, such as Burkitt's lymphoma, can be caused by viral infections, in this case Epstein-Barr virus). In any event, a number of genes are already known to be tied to high cancer risk (c-myc, p53, and rb are a good start, for the cell biologists out there), and others to higher heart disease risk (hypercholesterolemia leaps to my mind, but there are, I am sure, others), so these are perhaps a good model. Within a few years, epidemiological and lab model system experiments will allow for identification of most, if not all, of the risk factors. Guessing a bit, I would suspect that there is a normal distribution of risk in the population (a Bell Curve, if you will). Because we will be able to specifically and conclusively identify these risk factors, we, that is to say the insurance companies, will be able to assign a pretty good relative risk value to each client/patient.
So, you might ask, why is this a problem? Well, I may be wrong, but I think this puts the patients (and, to be fair, the doctor's) in a bit of a tricky spot. After all, when all I have to go on is some loose guidelines, it is hard to be too concerned about whether healthy Jane, who is a Vegan and runs 5 miles a day, is paying too much for health care, while fat slob Pete, who eats 3 steaks a day and jogs from the bed to the couch, is paying too little. When I know for a fact that Jane has a 10-fold greater cancer risk because of genetic factors, and Pete has perfect genes that mean he'll live to be 100 no matter what he eats, that changes the calculus. It no longer is an economically viable model to charge everyone the same rate over a large (say, company/industry/statewide) population, when you have solid knowledge about who is a high risk. Of course, it could be that drug companies will invent a cheap pill for every condition, but even then, wouldn't you rather insure only the people who didn't have to take it?
I suspect (no, I know) I am not the first person to think about this, and there may be very good answers to my concerns. If there are, I would love to hear them. If they are of the "Insurance companies would never do that, they are good public citizens!" variety, leave 'em somewhere else, 'cuz they aren't wanted here. Constructive ideas/criticism/speculation is always welcome, however!