Last updated 8/24/2024
If the treatment you’re considering is cheap and harmless, and you think it will help you, you might want to start there, unless your symptoms are so bad that you need a treatment most likely to help.
If the treatment you’re considering is expensive or not likely to be harmless, then it should have evidence that it works better than a placebo.
What is “Evidence”?
Research evidence means there’s more reason to believe a treatment will help you than “someone says so”. Suppose someone says “Rub this cream on your face, it will help you not drink alcohol, it’s only $100 a jar.” And now they show you comments (“testimonials”) from people who have used the cream and are not drinking alcohol anymore. Well, I guess that’s better than having only the word of the person who’s selling the cream!
So that’s the first level of evidence: testimonials. The next level is similar: testimonials from a lot of people, not just a few. The person selling the cream says “hundreds of people have used the cream and stopped drinking.” No, that doesn’t count. That’s just “someone says so” dressed up as evidence. Actually, so are testimonials, unless you have good reason to believe those were real people who weren’t somehow benefiting from saying nice things about the cream.
If a treatment is expensive or potentially harmful, you need a reason to trust that it “really works”. You can trust a doctor’s recommendations, because presumably they have good reasons themselves to believe a treatment really works (but they’re not profiting from it; that’s what makes them different than your basic internet sales job. Okay, if they are profiting, like surgeons who only make the big money by doing surgeries, yes, you have to be careful there too).
If you really want to decide for yourself whether to accept a treatment being offered, and it’s either expensive or potentially harmful, here’s what I think you need to trust: evidence, from testing.
Evidence comes from testing
You can test an idea by predicting an outcome, like this: “if the idea is correct, X will happen. If the idea is not correct, X will not happen.”
But we want to know more than the result. A scientist can show how they did the test. Like a cooking recipe: another person could follow the same steps and expect to get the same result. Now we don’t have to trust the scientist who did the test. We can believe the idea because we see how it was tested.
But what if we can’t understand a complicated recipe? We can we rely one more thing: results and methods that have been “peer-reviewed”. The scientist showed them to people who can understand the recipe, and they concluded that the research was valid. So the article was published in a “peer-reviewed journal”. All the major statements on this website are supported by peer-reviewed articles.
What is the best test for a treatment you might consider?
The best evidence: placebo tested
Here’s the key idea: when you believe a treatment will help you, that belief makes the treatment likely to help you. Canned green beans: if you believe they will treat your depression or lower your anxiety, there’s a good chance that will happen — about a 25-30% chance, on average, if this is your first treatment.
But if your depression (or anxiety or substance use or traumatic recall) is really bad, if you are really suffering, you want a treatment that has an even better chance of helping you. You want a treatment that has been tested against belief alone, one that has been tested against… a placebo.
The typical placebo is a “sugar pill”. Or it might be a believable “sham” treatment, like rubbing a cream on your knee for arthritis — a cream that has no active treatment in it, for comparison with some other cream, or even with knee surgery.
A good placebo, if it is believable, is almost always better than no treatment at all. Because our bodies can heal themselves, sometimes a little, sometimes a lot. Believing we are going to get better somehow can begin that healing. It’s amazing. It’s sort of miraculous. This doesn’t work all the time; and for some serious problems, it may not help much. And yet it can, even for illnesses like cancer (rarely).
This treatment “really works”
To say a treatment “really works” means it works even better than a placebo. This requires a formal research study in which participants are told they might get the treatment being tested or they might get a placebo treatment, and they won’t know which group they’re in. They are randomly assigned to one group or the other. So these are called “randomized clinical trials”, or RCTs.
RCTs provide the best form of evidence that a treatment works. But they need to be large, and they are complicated to run. So sometimes a treatment you might consider may not have this level of evidence. Sometimes, there’s just “clinical experience.”
Why clinical experience is tricky
On one hand… a clinician accumulates experience. If they’re careful with their observations and conclusions, they can know things that randomized clinical trials will never show. They will have tested their ideas by making predictions to themselves and then looking to see if they outcome matches their expectations. They won’t ignore the times when their prediction was not correct. For example, if a patient does not return for further care, they’ll assume the treatment didn’t work.
Or if the patient is uncertain about improvement, the wise clinician will take this as evidence the treatment did not work as intended, rather than twisting this result into a more positive outcome.
On the other hand… Consider a clinician who really believes that a treatment can help. For example, take me and lamotrigine. If I’ve recently seen some patients get much better on lamotrigine, this will strengthen my belief that lamotrigine is a good treatment for people like them. The next time I see such a patient, I will describe lamotrigine in even more positive ways (even just body language). Seeing and hearing my belief in lamotrigine increases the chances that lamotrigine will help my new patient.
Now we have a self-reinforcing spiral: the more I believe in lamotrigine, the more strongly I will describe its potential benefits. This makes my patient more likely to believe they will get better on lamotrigine – which increases the chances that they will indeed get better! That strengthens my belief, and so on it goes.
By the way, lamotrigine does have some randomized trials showing that it works better than a placebo. But you can see how my belief in its benefits could amplify itself.
Any provider who sees their treatment produce a good result — even if it was a placebo result — is likely to get excited about having helped someone. That excitement can help the treatment-provider convince other people that this particular treatment “really works”. Good outcomes are easier to see than bad ones, because sometimes people who didn’t benefit just go away, whereas people who got better say “thank you” (and buy some more treatment). This makes the treatment-provider even more convinced, and so on.
This is why you must be careful…
Don’t trust enthusiasm
…unless the product is cheap and harmless. Otherwise, watch out for people who believe their treatment can help you (especially if they are making a lot of money providing that treatment),
Instead, expect evidence that the treatment is better than a placebo (because placebos are cheap and harmless). That means a randomized trial — or at least results that have been published in peer-reviewed journals. If no such evidence is available, or if some studies find the treatment better than a placebo, but others don’t; then look for an experienced clinical opinion. Or maybe several, if you can get them.