How Your Health Care Works, Part III "statistics"

When I first began to study statistics in undergraduate my professor said, "there are three kinds of lies, there are lies, damn lies, and statistics". It wasn't until many hours of study and conversations that I fully understood what he meant. As a practicing doctor I was often frustrated with the media's reporting of studies and stats. "A recent study showed......!" The frustration came from the general population's acceptance that if it came from a study it must be true. Pharmaceutical companies, politicians, and media agencies use such generalizations to sensationalize or drive home a point or opinion that they support. This is actually dangerous, you can see some of the most blatant misuse and ultimate effects by looking at the Vioxx debacle. How many people died because of a misuse of statistics? This is one that you know about, how about the misused stats that we are depending on today to make decisions, that we will find out later are killing us?
When explaining this to people I have tried to use everyday stats that we make general assumptions about. 1% fat milk! sounds great right, must be 99% less fat? Actually milk is only 3% fat to begin with, so it is 66% less fat. Overall it is unnerving to consider that we accept stats from anyone with an agenda, they can be spun, adjusted and presented in a way that can ultimately influence us to make choices that are not necessarily in our best interest.
Here are four pieces of advice to help with stats, you don't need to be a statistician or even college educated.
1. Consider the source, what does this person or organization have to gain by presenting this statistic in this fashion?
2. WAIT! any truly significant statistic that changes what we do or the way we think will surface again and again, the rest will ultimately be dis proven and you will never hear from them again.
3. Ask as many critical questions as you can, assume it is bullshit to start. I know, sounds rough but, if you change what you are doing and it ultimately turns out like say, vioxx, you will be dead and it won't matter. If you assume it is crap until "proven" otherwise you have lost nothing.
4. Whenever possible talk to an objective professional who understands statistics. Admittedly there is no one who is truly objective, but at least someone who has your best interest in mind and at heart.
In medical care when beginning a new treatment I suggest you ask the "number needed to treat" (that is how many people are treated with this before you get one good result?) If the NNT is 500 but 1 in 5 people get side effect, well, whether or not you risk it in this situation depends on how serious the condition is. The decision is ultimately up to you, but be informed. I further recommend that you ask what the top five side effects are and how many people end up with those? These questions can ultimately be extrapolated to procedures, surgeries etc. Finally, ask how this treatment, procedure etc. compares with others, if it is no better than say, a cheaper, more well known treatment or procedure why are we even talking about it.
In closing, statistics are complex, obtuse and unfortunately adjustable. Remember a statistic is not "THE TRUTH" be a cynic, it will serve you well!!!