I spent some time yesterday listening to a podcast: “The Skeptic’s Guide To The Universe“, where they talked about their favourite shows for 2011, their Best and Worst for the year, among others. One of Dr. Steven Novella’s picks for the anti-science (my words, not his) person of the year was Dr.Oz of Oprah fame (Dr.Oz has his own show now).
When Dr.Oz started out with his program, he was entirely fact-based, research-proven and mainstream in most everything he said. But with time, I guess he ran out of facts, and started delving into unproven “alternative” medicines.
I have no issue with anything alternative (neither does Steven Novella), if someone takes the time and energy so study cause and effect in a scientific way.
The Scientific Method
The basic steps involved in the scientific method are:
- Ask a question.
- Research possible answers to that question.
- Construct a hypothesis on the answer to your question.
- Test the hypothesis by designing an experiment.
- Analyse the results of the experiment and draw a conclusion.
- Communicate your results.
It seems to me that we’ve done pretty good by following this process … we have modern antibiotics, have sent a man to the moon, drive around in cars, have safe drinking water, the PC, and several million other advances.
The Problems with Distorting The Scientific Method
The problems we get into are when we allow this method to go off the rails. By having someone with a vested interest in the outcome have some control of the experiment (financial control?). By taking only the results you like and discarding all of those which contradict your hypotheses. By having a bias in the way the experiment is set up so you will only see the results you expect. By not eliminating, as best you can, any other factor which could affect the outcome of the experiment – so you know you’re testing one hypothesis (more sophisticated experiments can test several hypothesis at once).
The best studies (which experiments are sometimes called) are double blind, which means that neither the researchers nor the subjects in the study know what variable they are controlling until after conclusions about the experiment have been drawn. An example would be a melatonin study where neither party knows who’s been given melatonin and who’s been given a placebo. Both groups are put into identical situations, and after the results (whatever you’re studying) are in and have been analysed then the groups are revealed.
There are hundreds of ways to intentionally and unintentionally screw up the results of the experiment, and you end up with a false hypothesis supported by a faulty experiment.
So What Does All This Have To Do With A Sleep Disorder?
So here’s my point.
There are hundreds of “cures” and “therapies” out there for tackling a sleep issue. Many of them have been handed down from generation to generation, and have been taken as gospel truth with a lot of people.
So was bloodletting.
If you’re ever going to solve your sleep problem, then it would make the most sense to research the most scientifically proven methods of handling that problem. If there’s a study paid for by a pharmaceutical or supplement company (it can be very tricky to find out), treat it with suspicion. If the study isn’t blind or (preferably) double-blind, find out why.
There has been plenty of research into sleep, and we do have proven methods for solving sleep disorders that have been studied, hypothesized, experimented and analysed. It’s not only medicines and supplements that can be evaluated this way – any other therapy can as well.
Accupunture? Lavender? Herbal Tea? Benzodiazepines?
What’s A Good Study Look Like
This is mostly common sense, but it requires some thinking on your part. Learn how to ask questions and think sceptically.
Modern scientific research wants to let reality speak for itself. Look for studies in which the hypothesis could be easily disproved. People have a tendency to see the results that they want to see, so if an experiment is designed to disprove a hypothesis, that’s a good thing.
Every step of the scientific process should be available for peer review. This means that other scientists can look at the experiment and the results and examine it for any problems. Some scientists don’t work in a peer reviewed process – and I would take any conclusions they reached very suspect. Basically, the peer review procedure should take care of any issues regarding the design of the study.
Make sure that there’s lots of data points. If a study comes up with conclusions from it’s peer-reviewed research, but it’s only with 5 people, then treat it as interesting, but wait for a larger study to confirm.
What and Where Are Some Of The Well Researched Sleep Studies?
There’s a few places that you’ll find them. The journal, Sleep, is a product of the Associated Professional Sleep Societies, while The Journal of Sleep Research is from the European Sleep Research Society. Myself, I just use the search function on either site. Here’s an example of what you’ll find in the journals:
- Intensive Sleep Retraining For Chronic Insomnia
- Melatonin, Natures Most Versative Biological Signal?
- The Treatment of Central Sleep Apnea Syndromes in Adults: Practice Parameters with an Evidence-Based Literature Review and Meta-Analyses (this is neat because it’s a study of previous studies)
- Unexpected survival advantage in elderly people with moderate sleep apnoe
You’ll note, that if you were to type in “lavender” into the Sleep site’s search engine you’d come up with a few studies – so it can be used for almost anything sleep related.
These are good tools to have and a good habit to get in to. When someone says to you “the cure for insomnia is ___ (fill in the blank)” you can have a look to see if it’s ever been studied seriously.
And you should use the scientific process whenever you evaluate your own sleep. I myself use a Zeo Sleep Management system to measure my actual sleep. Then I can try to change one (sometimes two or more, but they’re harder to track) variable at a time and see how it affects my sleep.
The fear here of course is that the study is not blind or double blind, and you are likely to have your own biases on how you’d like the results to be. And sleep results can be confounding to figure out. Check yourself to make sure that you’re not seeing patterns that aren’t really there. Better yet, have someone else look at the data and see if they can find a pattern. (If you do try the Zeo route, they have some good cause and effect tools on their user website that can help you with this analysis, and also some coaching to help you form your own hypothesis about your sleep).
The final word on sleep research is to remain sceptical. There are lots of “cures” out there, so you need a process to separate the good from the bad.
