Various updates about life from my perspective.

[Read more →]

(via Metafilter)

The BBC is running a series of articles on statistics and on how to spot flaws in media (and other) reports of statistical findings.

I’ve read lesson 1, but not the others, but it looks hopeful. I hope they apply these to their site’s editorial standards.

I had cause recently to go (re-)read Directive 15 (because I had been speaking authoritatively about it and hadn’t reviewed it in quite some time - a couple of decades, probably). The last time (before this time) I thought I had reviewed it (or potentially related regulations with the IRS or some other governmental organization - I’ve had a lot of trouble trying to find such a citation, by the way), I learned that I couldn’t self-identify as White, because at that time only folks who were 7/8 White or more were allowed to self-identify as White. So I could only identify as Asian/Pacific Islander.

On further research this morning it sounds like I might have been reading an older regulation (if I had to guess, it was issued by a governmental or private agency in Virginia, but I can’t narrow it down further than that - I just don’t know) based on the older “One-drop rule“. It also sounds like in practice, this kind of rule was really only applied to folks who could be categorized as African Americans, but when I read it I’m pretty sure I thought it applied to me as well.

For reference of anyone who might find this via IBARW, I am half Chinese American and half Caucasian American by blood, ancestry, parentage. I can tell you where my families are from, but it also works if I say I’m 1/2 Taishanese, 1/4 Welsh, 1/8 English and 1/8 Scot. Cool, huh? I’m proud of my inherited ancestries.

Anyhow…

[Read more →]

This is largely stimulated by a blog post by Junkfood Science from today.

I’ve talked about the dangers of interpreting relative risk before. Even if the actual change in risk is 0.3 (between 99.5% and 99.8%), the relative risk can be 150%, scare the crap out of you and get you to take a foolish risk.

Below is a chart out of one of the links JFS references, which is from a Letter sent to the Archives of Internal Medicine in 2002. The authors analyzed six (at the time) current studies on statin therapies and risk reduction of total mortality among different sets of populations. (click the thumbnail for the full-sized graphic - legible but still not very large)

(From Letter to the Archives of Medicine, “Exaggerated benefit of statin treatment in the elderly?“, by Ravnskov, U, et al., July 20, 2002.)

If you apply these numbers to actual total time per individual potentially saved between statin and non-statin lifetime use, you’re actually looking at (and this is especially stupid) about 15.6 days extra life after treatment for 30 years. 30 years of taking expensive medicines with potentially very harmful side-effects for a little more than 2 weeks extra life, statistically, which means basically no guarantee for any specific individual. If you look at the chart above, you’ll actually see that some of the studies show better mortality rates for folks not on statins.

But still you get (ethical or not) medical professionals willing to testify that people, even seniors, should take (expensive, risky) statins for the rest of their lives for the sake of (no promises now!) a couple extra weeks of life. I guess it’s up to you, but I’d rather not take the statins, or pay for them, honestly.

The other point I wanted to make about fat/fit/health/medicine is that we have now achieved the status, in U.S. medicine at least, of turning medicine into a technical support interaction. In this kind of interaction, because it is statistically sensible to do so, (wherein 95% - 99% of all complainants can have their health concerns addressed, as a technical issue, by a set number of very well understood - if not characterized - procedures), everyone gets treated as if:

[Read more →]