Getting Risk Right

Understanding the Science of Elusive Health Risks

[Amazon Link]
(paid link)

Another pick obtained for me via the Interlibrary Loan folks at the University Near Here. Thanks, folks!

I've been interested for a while about the general topic of risk, and how it might best be regulated in a free society. I'm still in the weeds, but maybe if I keep reading…

The author, Geoffrey Kabat, is an epidemiologist at the Albert Einstein College of Medicine; he's deep in the field of quantifying health risks. This book outlines the pitfalls and biases that researchers investigating causes of disease can fall prey to. And the realities of funding and publication pressures on researchers can, and does, incentivize a lot of bad, misleading, and unreproducible results. (For more on that topic, see the recent Wired article about John Arnold.) The discussion here is technical and valuable.

Kabat then turns to four recent case studies, two bad, two good. The bad: efforts to show that cell phone usage causes brain cancer, and the notion of "endocrine disruptors" in the environment. Both areas were sensationalized, and wasted a lot of funding that could have been spent on more productive areas.

The good: research into dangerous "Chinese herbal supplements" containing aristolochic acid; and how varieties of human papillomavirus (HPV) can result in various cancers, mostly cervical. These are heroic stories, as the scientists involved i-dotted and t-crossed their studies in order to tease out valid results, exposing disease processes that can take decades to actually kill you.

Kabat's prose is unfortunately wooden and to-be-sure mannered, and he's mind-numbingly diligent about dropping names and professional affiliations of the researchers he discusses. Fine, but that gets in the way of the narrative. (He's no Daniel Kahneman, sorry.) But, even given that, the book is a good discussion on how to do research right, and also wrong.

He (also unfortunately) doesn't get close to my main concern: once you've identified and quantified a risk, what's the proper policy response, especially given that people have widely varying tolerance for risky behaviors?

Also, a pet peeve: he refers to "socioeconomic inequality" (p. 179) when he almost certainly means "poverty".


Last Modified 2024-01-26 6:52 AM EDT