Ron Bailey Unzipped

By Razib Khan | December 14, 2010 1:07 am

Over at Reason Ron Baily has an excellent piece up, I’ll Show You My Genome. Will You Show Me Yours? He reviews his results from two genotyping chips, and has placed his results online. I doubt readers of this weblog will learn anything that new, though the article might prove illuminating to friends & family. But, some cautions:

– Even papers which make it through “peer review” may not be reporting robust results. So Baily reports the breastfeeding & IQ gene-interaction findings…but the statistics in this might be flawed. Reason should also just tack the Genomes Unzipped odds ratio post to that article. Bailey knows how to interpret the results (with a grain of salt), but I don’t think how important this is comes through in the piece.

– Second, he repeats the argument that genomic screening might make insurance unfeasible:

In 2008 Karen Pollitz, director of Georgetown University’s Health Policy Institute, wrote in the journal Managed Care, “Much of life’s uncertainty about health will become much more known to us, and since insurance is all about protecting people from the unknown, that will be a profound change.” So profound that Pollitz thinks it could ultimately make the insurance industry obsolete….

Unlikely. There’s still a fair amount of randomness in health outcomes. This is probably a function of the stochasticity of development and environment. I think people need to stop repeating this argument (which I found plausible until I thought about it more deeply).

Oh, and a final minor note, Bailey notes that contrary to family rumor he doesn’t have Native American ancestry. I don’t know what Pathway does, but 23andMe really can’t go back more than five generations apparently.

(Acknowledgement, Daniel MacArthur, Ph.D.)

CATEGORIZED UNDER: Genetics, Genomics

Comments (3)

  1. The concern about insurance and genetic testing does have merit to it, absent significant insurance regulation, although it isn’t unqiue to genetics and has a larger privacy element.

    For example, suppose that you are selling car insurance to new drivers. Traffic violation and accident history can’t be used to price policies because nobody in that group has any traffic violation or accident history. But, there are lots of statistically significant assocations that an insurer could use. Some of them are credit histories and grade point averages. Gender has a huge demonstrated impact that is used. Insurers would use race if they could, and instead use zip codes as a proxy for it. DNA markers for traits like novelty seeking (and you could compile a list of five or six markers that would be even more robust) would make the rating more fine grained and would probably have a strong enough relationship relative to the kinds of underwriting factors already used to put some people in a high risk auto insurance pool before they even put their keys in the ignition.

    Progressive Auto insurance is hopscotching over genetic tests to measure the phenotype more directly by offering premium reductions to people who agree to put a black box in their car (to measure things like accelleration patterns out of stop signs, speed above speed limits, and sudden stops) and are shown by the data mining to be safe drivers.

    You can impose “community rating” that limits the ability of health insurers to do this kind of thing, and the Health Care Reform law (popularly called Obamacare) does just this in the United States, and most non-American health care systems have similar limitations on linking premium cost and actual risk (or decouple the two entirely with a tax funded universal health care system).

    The hard question is not whether DNA tests could dramatically increase health insurance rates for some people and reduce them for others (surely they could because they accurately predict a significant number of high cost pre-existing conditions that have major premium and insurability impacts). Instead, the hard question is whether we should care. It is a matter of degree. The better you can predict someone’s future health care costs, the more health insurance looks like a system of prepaying medical costs rather than a system of societally sharing unpredictable health care costs. The ultimate in predictable insurance costs is permanent life insurance — you will die, and if the policy will pay out in a certain face value without further premium payments the cost of the benefit is certain, just the timing is uncertain and once predictable factors are considered that uncertainty isn’t as big as you might think.

    Predictability (from any sense) forces us to parse cost sharing from risk sharing. It makes the transfer payment element of insurance that is concealed in high uncertainty situations visible. Maybe that isn’t a big deal. Maybe we should try to have more predictability and openly subsidize the basic needs of those who predictably aren’t able to meet their needs.

    The main difference between rating based on genes, and rating based on say, triglyceride levels, or blood sugar counts, or weight, or history of smoking, or gender, is that you know earlier. Rating insurance based on genes is very similar to rating insurance based on family history of mental illness, cancer, diabetes, heart disease, etc.

  2. Qohelet

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar