Monday, August 13, 2007

What do all these studies really mean?


…Increased chance of this side effect…more incidence of that…every day some study comes out pushing or debunking a drug or procedure (angioplasties are no longer the thing, did you know?).

…Like people rushing from one side of a boat to the other, patients and the “worried well” caper to and fro like the puzzled puppets that we have become..

…Now there is a new book by environmental scientists Erik Rifkin and Edward Bouwer trying to cut through the BS and help people decide if these authoritative-sounding edicts from on high are worth a flip.

…The book is called Illusion of Certainty: Health Benefits and Risks.

…The uncertainty, says Rifkin, president of an environmental firm in Baltimore, has been lost in the health area—the numbers have taken on a life of their own.

…The two are used to looking at environmental data and saw that the "don’t know" factor was not being communicated.

…Hang in with HA now…this gets tricky. One measure is called absolute risk reduction. This looks at the difference between two groups, such as a group that got a drug and a group that didn’t. If one person died among the 100 who took the drug, the death rate would be 1%. If two died among those who didn’t take it, the death rate for that group would be 2%. Absolute risk reduction because of the drug—1%.

..BUT, and this is the kicker, some drug companies and yes, journalists, go by relative risk reduction, rather than absolute. This compares raw numbers of people who died. Because twice as many people (2 versus 1) died without the drug, the drug is said to cut the risk of dying 50%.

…That sounds good until you realize it was only a 1% risk anyway.

…Absolute risk is a better measure for patients.

…And, of course, patients should realize a study does not predict an individual result.

…HA remembers the statistic that one in three women will be raped in her lifetime? Three women in a row may look from side to side and wonder who is getting it. Chances are—none!

…Some other women someplace will make up the statistic.

…Studies don’t predict individuals. HA wagers hat people do a better study of the odds on roulette before going to Vegas than the odds on their skipping or getting a test or therapy.

No comments: