personal genomics and nocebo

[gusl]

One argument against doing something like 23andMe on yourself is that there is very little potential placebo, but significant potential nocebo and the stress that comes with knowing that you have slightly higher odds of a deadly condition. This is because medical information is framed in negative terms: diseases, rather than e.g. healthy lifespan.

Humans, even the most educated, are bad at understanding small differences in probabilities. I wonder how the health gains that come from the increased knowledge compare to the losses due to increased stress and hypochondria. Is one better off not getting exams at all? Or leaving it all to the doctors?

See also: cyberchondria

Comments

[tedesson.livejournal.com]

I think diagrams like these are helpful for visualizing these sorts of questions:
http://blog.oscarbonilla.com/2009/05/visualizing-bayes-theorem/

This wired article has a graph plotting rates of different diseases over time:
http://www.wired.com/wiredscience/2008/06/revisualizing-y/
Amazing the improvements through 1973. The improvement in health care technology factor over time, is always left out of static presentations. Would you think differently about your risks, if you knew the probability of a cure being discovered in the next 5 years was 50%?

In a slightly different form, Wolfram|Alpha has charts:
http://www.wolframalpha.com/input/?i=heart+disease+risk+50yo+male
of the way different measurements correlate with risks. They don't have these for most diseases. And, it's hard to say how much control you have over things like lowering your blood pressure from a normal level.

For myself, I've decided to reduce my frequency of breast cancer screenings, to the NIH recommended rate, because it's not clear to me that the screenings reduce my risk of disease and death. This is primarily because the screenings for a woman my age have a high false positive rate, and a low false negative rate. If I reduce the frequency of screening by half, I have impacted the false positive frequency more than the liklihood of missing a real disease.

I've thought about doing 23andme, but until there is more actionable steps one can take with the information (beyond what I already know), I'm going to pass.

And on the placebo effect, do you do things to encourage it? A 30% efficacy is significant.

[cwarner.livejournal.com]

!!!! OMG Venn diagrams are such a great idea! I am one of those people with a plethora of irrational fears. I think constructing Venn diagrams would be a great way to get a more intuitive sense of actual risks of things which would in turn maybe make me less terrified.

Kind of related to the nocebo response, knowing that an elevated heart rate increases my risk of having a reaction from my allergy shots frightens me, elevating my heart rate and causing...well so far I haven't had a dangerous reaction...but enough of a reaction that the doctors have had to lower my dose...

It probably depends to some extent on the person whether knowing one's risks for diseases would be helpful, harmful, or neither. I'm guessing that it would be harmful for people like me.

[gustavolacerda.livejournal.com]

I find the animated plot on Wired completely baffling.

[bram.livejournal.com]

You know there's a gene that affects one's response to placebos? (If you're told you have this gene, does that mean your placebo effect goes up even more?)

If you dig far enough into your genome you can find things to think about positively--genes giving a statistical benefit of a few points of IQ, fast-twitch muscles, etc... As well as things to tweak--you can learn how quickly you metabolize caffeine and alcohol, medications, vitamins, etc.

[gustavolacerda.livejournal.com]

<3

(I'm not surprised they've identified a gene associated with placebo, since they already identified a chemical that blocks placebo effects, naloxone.)

[cwarner.livejournal.com]

How...

[cwarner.livejournal.com]

Wow!