Map of life expectancy at birth from Global Education Project.

Wednesday, December 22, 2004

Pills, damn pills, and statistics

It's been a rough few months for the pharmaceutical industry, what with news about antidepressants increasing the risk of suicidal thoughts and behavior in children, so-called "hormone replacement therapy" and various nasty consequences, most recently news about Cox 2 inhibitors and cardiovascular disease, and now NSAIDs, plus some less publicized problems.

It's also been a rough patch for the people who were taking these drugs, obviously, and now a lot of folks out there are confused and don't know what the heck to do.

At the heart of these problems we have a tangled intersection of epistemology (the theory of knowledge - how we come to believe what we think we know [redundancy intended]), difficult value judgments, politics and yes, industry greed. The latter is to be expected, of course, but the epistemological thicket and the values fog make it difficult to properly contain.

Which thread to pull first in this skein? I pick the field called biostatistics. Medical practice and applied biomedical research depend very much on the language of probability, which is called risk when we care about what happens. But the so-called positivist epistemology which underlies the scientific approach to knowledge has a very hard time with probability. It is not clear in what sense statements about probabilities are verifiable. The blow dried guy comes on TV and tells us that there is a 50% chance of rain tomorrow. Tomorrow comes, and it rains or it doesn't rain. Guess what? He was right!

I tell you that you have a .5% (that's .005 out of 1) chance of having a heart attack in the next year, but if you start taking this pill your chance will be .75%. But that's not actually true any more if we learn more about you. For example, if we test your Low Density Lipoprotein ("bad" cholesterol) and find that it's below a certain number, your risk is less to begin with and is increased even less if you take the pill. Was I wrong when I gave you the figures of .5% and .75%? No, those are the odds when we don't know about the cholesterol, so I was right. And no matter what happens next year, whether you have a heart attack or not, I was still right. And what if I give this news to, say, 100 people? In order for me to be "right," does 1/2 of a person need to have a heart attack? No, I could still be right if nobody has a heart attack, or even if 10 people do, although I would then say that the outcome which ensued was an "unlikely" one. Well, it isn't unlikely any more, it happened! The probability of events which have actually occured is 1, 100%, no matter how unlikely they were before they occurrred.

What if I tell you that the pill increases your risk of attack by 50%. Does that sound scary? It probably isn't worth it. What if I tell you it increases your risk by only .25%, in other words that 2.5 people out of 1,000 will suffer a heart attack because of taking this pill. Maybe you'll take a chance now. But both of those statements are true -- in some sense, but neither of them may be true for you! In the end, you're either going to have a heart attack or you aren't, right?

9 comments:

buy cialis said...

I don't trust pills due the amount of chemicals that pharmaceuticals use to produce them, I prefer homeopathy instead!

buy generic viagra said...

no, pill from certain trustful pharmacies are a big deal burt only if you have a prescription.

Buy Clomid said...

Grass on your doctor and laboratory operators around all medications and herbal products you are irresistible if you are treated with buy Clomid online. Together with allergy symptoms such as lump of lips, tongue, or face or hives in some patients may look the following conditions: ovarian enlargement presenting as abdominal or pelvic suffering, flushing, nausea, vomiting, or diarrhea, blurred understanding, headache, abnormal uterine bleeding.

Buy Valium said...

Valium, influences on receptors in the intelligence and gamma aminobutyric acid is discharged. Neurotransmitters are the couriers between neurons, and they are kept in neurons in the imagination and in the nervous system. As gamma aminobutyric acid is discharged into the perspicacity by means of consuming [url=http://www.buydiazepamsite.com]buy generic valium online[/url], the congress turns more relaxed, peace and/or sleepy. Gamma aminobutyric acid may as graciously slow or block the issue of additional impertinence signals in the brain.

Buy Ambien said...

But they neutral can’t get to sleep no matter how much they stand in want to. For the first-class effects please swatting and follow the manufacturer’s marker accurately before delightful Ambien. This is called insomnia and is the calculate why people keep tossing and turning, twitch positions at night, assume from a book or use the old fashioned method of buy ambien no prescription! It is a common luggage for doctors today to pull down cases of patients complaining anent not able to sleep or get ample sleep.

www.soria-3d.com said...

Quite helpful piece of writing, thank you for this article.

Hotels miami beach said...

Nice Information. Great great i was looking all day for this post on google and finally i got it.thanks so much for posting such an amazing post: ): ): )

BroddyAdams said...

Thanks untuk infonya. Semoga ekonomi Indonesia makin baik setelah munculnya para pemikir-pemikir ekonomi baru di tanah air. Mudah-mudahan anda adalah salah satunya.

CRM Software

Anonymous said...



Acheter Viagra en ligne
Viagra Pharmacie