Evidence-Based Treatment

About Placebos and Randomized Trials

Entering “bipolar disorder” on a large search engine today yielded a list of 19.3 million sites (can you believe that?!).How do you know who to listen to?

One of my mentors once said, “Jim, there are a lot of treatments out there. Why don’t you start with the ones that have some evidence that they work? “Sounds smart, no? But what is “evidence”, and where do you find it?

With all the health care treatments being offered these days, especially on the net, your main defense against ineffective treatments is to understand randomized controlled trials. Here’s what you need to know about this standard research technique.

First, you need to know that “placebos” work! If you take 200 people with depression, and give 100 of them Prozac, and the other 100 a sugar pill, what happens? In the Prozac group, about 50 or 60 will not be depressed (using a special test to detect depression) in about 1 month. In the sugar pill “placebo” group, how many will be better? In most of these kinds of research studies, around 30 of the 100 who took placebo will have recovered. If we had another 100 on a wait list during this time, only about 10 of them would have recovered.

This means that getting a pill, with the hope and expectation that it will help, really “works” to treat depression. Not for everybody, but definitely for some. The same thing is true for panic disorder treatments: placebos generally help about a third of patients.

You can see what this implies. If someone offers you a “treatment” for panic disorder, they should be able to show you that their “treatment” is better than a placebo! After all, placebos work! So if someone says “look here, these people got better taking this herb” — well fine, we’d expect some people to get better taking that herb. The question is, would people taking the herb get better more often that people taking a sugar pill? If not, why don’t you just take the sugar pill — it will surely be cheaper, and maybe safer too.

All right then: what treatments out there for panic disorder have been shown to work better than a placebo? Well, you’d have to see the results of some sort of test, where the treatment was directly compared to a placebo, as with our 200 depressed people above. Such a test is called a “randomized controlled trial”. Randomized means the 200 people were split into two groups without any kind of bias, just some sort of system for putting them in one group or the other “at random” (usually a computer generated random number list is used for this).

“Controlled” means that there was a placebo group. And “trial” means that the proposed treatment and the placebo were given over a period of time, then compared using some sort of measurement. In the case of panic disorder, the most common measure is: how many in each group are “panic-free”?

Now you’re ready to look at the “evidence” for panic disorder treatments, because fortunately there are several randomized controlled trials to examine. If this “evidence” business bothers you because it sounds too much like Western Medicine, or “science” — and you want something different, something more Alternative, read this.