Insomnia Demonstrates the Future of Evidence-Based Medicine: Treatment Scores

The most important thing a medical review article should do is figure out the "net benefit" for the patients studied in the past. Yet, medical review articles mostly don't even try to do this. Why not? Because it is tremendously difficult. It has been said to be impossible!

Take insomnia for example. You would think it would be easy to figure out the science for insomnia treatments. It's an extremely common medical problem. Using the Treatment Score Analyzer™ software at I created a diagnosis for insomnia.

The problems begin immediately. How do you define insomnia? It's not nearly as straightforward as you think it would be. Are we talking about primary insomnia? Or secondary insomnia caused by some other medical issue such as chronic pain? What time scale are we looking at? 2 weeks? 3 months? The reality is that you have to combine studies over different time periods when you review the literature.

What patients are we looking at? Kids? Teens, Adults? I am focusing on adults only, but with all the inclusion criteria and exclusion criteria, or lack thereof, describing the patient population can be much more complicated.

Insomnia can mean several things. Insomnia can mean difficulty falling asleep, difficulty staying asleep, or waking up too early. Insomnia can mean poor quality sleep, so that you wake up feeling like you haven't slept well.

For this review, I decided to quantify treatments that were studied with the Insomnia Severity Index as the outcome measure, because it is a questionnaire that tries to take many of the components of insomnia into account. After a ridiculous amount of work, because the tools we need for science-based medicine are still in their infancy, I came up with a short treatment list and Treatment Scores.  

Treatment Scores represent the net treatment benefit for the patients studied in the medical literature (on a 100 point scale).

Why this list of treatments? I chose Ambien (zolpidem) because it is one of the most common prescription medications for insomnia. I actually wanted to review 5 mg, not 10 mg, but I could not find enough appropriate studies.

I chose to review "indiplon" as a treatment because I was able to find a high-quality randomized controlled trial that used the Insomnia Severity Index as an outcome measure. Indiplon is interesting, because if you read Wikipedia ( it sounds like the company obtained FDA approval, but the FDA apparently asked for more studies, and the drug company, Neurocrine, may have simply given up at that point. It's not entirely clear, but that will be another interesting essay for another day. Why did they give up?

I chose to study "tart cherry juice," because a friend asked me a question something like this: "I read about tart cherry juice for insomnia in a magazine, what do you think?" When I looked up tart cherry juice in the medical literature it turns out that there is biochemistry supporting it. It apparently increases melatonin levels and may also have anti-inflammatory properties, both of which might be mechanisms to help people sleep better. At Treatment Scores we don't care if a treatment is Western medical, Eastern medical, natural, herbal, or any other type of alternative medicine, we only care about the data.

Wouldn't it be nice if you could look up a diagnosis, get a list of all the possible treatments for that diagnosis, and see what the Treatment Scores are? That's what we are working towards with the tools at and are blogging about here at

In fact, what we want to get to is Treatment Scores and Treatment Grades. We want to figure out if a treatment is a "grade A treatment" according to the medical literature, which is the highest grade, or a grade C treatment, which is average, or a grade E treatment, which is the lowest positive grade. In other words we are using grades A, B, C, D, E for positive net benefits found in the medical literature. For treatments that are harmful according to the medical literature, we give them a grade F. See the disclaimers at the end of this document. We have Treatment Scores and Treatment Grades in our graphic below:

Ambien (zolpidem) is interesting, because unlike some of the previous prescription medications for insomnia that were in the benzodiazepine (Valium) class of drugs, Ambien (zolpidem) allegedly does not produce tolerance, sedative effects, or signs of withdrawal when the drug is stopped (at least according to some sources). Ambien (zolpidem) is in a completely different chemical class, the "imidazopyridine family."

How was the Treatment Score calculated for Ambiem (zolpidem)? I used the Treatment Score Calculator™. It looks like this:

The Treatment Score Calculator™ is a tool that allows you to put the main statistic, or main outcome measure on the left, and the positive side benefits and negative side effects on the right. The Treatment Score Calculator™ allows us to organize the statistics and then adjust the main statistic up or down to create the Treatment Score. Note that I don't have every single side benefit and side effect scored on the right side of the calculator. Eventually, I just ran out of time, and had to stop where I was (but will complete this for a future blog post). However, I have read that the side benefits of "quality of sleep," "time to fall asleep," and "total sleep time" are all positive. The negative side effects are relatively rare as can be seen by the low numbers. What are the mathematical formulas that need to be used to translate the things on the right into values that can be added or subtracted to the main outcome measure on the left? These are all things that we are working on.

Where do the numbers in the Treatment Score Calculator™ come from? They come from medical studies. We capture the statistics in what we call STAR™ Blocks. STAR™ stands for "statistic and a reference." Below is an example of a STAR™ Block. The statistic on the right in the "rose-colored" cell is what is most important, but also important are all the other reference information: the title, authors, journal name and so on. By doing evidence-based medicine this way, with 4 tools, the Diagnosis Tool™, Treatment Organizer™, Treatment Score Calculator™, and STAR™ Blocks, we make the entire process TRANSPARENT. We have made the process visual so that patients, doctors, nurses, and everyone can see what is being done.

This type of quantified evidence-based medicine is in its infancy. We are taking baby steps. There are many issues yet to be overcome, and I want to list some of my frustrations.

Frustration: There are over 130 treatments for insomnia! There is a list at Wikipedia here: ( Some of the treatments don't require a prescription; some of them do require a prescription. We need to know the Treatment Scores for all 130 treatments. But imagine the work? Just to do the three treatment scores above has taken me days of time and effort. Imagine having to do it for 130 treatments! However, we are getting there. With "big data," "natural search engine technology," and other tools that we are building out, we will make it doable. It has been difficult and time-consuming to come up with the Treatment Scores and Treatment Grades for these three treatments for insomnia. In fact, if I didn't have the tools at, it would have been nearly impossible. A lot of time was spent searching the medical literature and a lot of time was spent reading the medical literature. We need to make this process of quantifying treatment effects faster and easier. The information is right there in the medical literature. Medical articles have the main treatment outcome. Medical articles usually have the frequency of the main side effects. However, all these statistics need to be combined to come up with the net treatment benefit for the patients studied. Doing this should be the focus of every single medical review article. All medical review articles should put what the patient needs to know front and center.

Frustration: More studies need to be done using the same outcome measures and the same dosages. Ambien (zolpidem) is often prescribed in lower doses than 10 mg. For the elderly and women, and even most men now, it may be only 5 mg. The dosage may be even lower for many people now. See your doctor, the full package insert, and FDA warnings. Figuring out Treatment Scores is a great example of why we need patient generated research. Why don't patients, doctors, nurses, and pharmacists get together and fill in the gaps?

Frustration: It's actually hard to come up with a diagnosis that is specific enough to find enough good data. What I finally did was use the outcome measure "Insomnia Severity Index," and a time frame of 2 weeks to 3 months, and adult patients. This is what doctors are doing in real life in their heads. They are combing all kinds of data from different time periods, and that are reported on different scales. No wonder few people really know what is going on, and no wonder there is so little transparency.

Frustration: There are so many possible side effects there is not room to list them all. Some are fairly trivial, some are rare but are life and death important. Ambien (zolpidem) has a rare side effect of "sleep driving." It has been reported that people actually do things under its influence and don't know they are doing them. I found that side effect next to impossible to quantify.

Here is a nice open source, full text review article you can read about Ambien (zolpidem): "Zolpidem Is an Effective Option with a Reduced Risk for Dependence in the Treatment of Insomnia," by Alam Shadab, MD, et al. (I have nothing to do with the company or companies that make Ambien or generic zolpidem).

One more thing about Ambien (zolpidem). Before Ambien, many sleeping pills were in the benzodiazepine family. That family of medications, which includes Valium, causes a lot of side effects. Benzodiazepines cause sedation, impaired cognitive function, risk of abuse, addiction, withdrawal syndrome, and dependence on the medication. I'm not saying that Ambien does not cause some or all of the same things, but most of the medical literature I reviewed has concluded that the non-benzodiazepine sleeping pills we have today are superior to the benzodiazepine pills in terms of side effects.

Sleep Apnea: Getting the diagnosis figured out is critical, but can be extremely confusing. While reading the medical literature for this essay, it seemed like I was seeing a "sea change" in the diagnosis of insomnia over time. By using sleep laboratories using machinery, such as polysomnography, I realized that the literature seems to be saying that the underlying cause of insomnia for many people, or even most people, is actually sleep apnea. Therefore, when studying treatments in the future, we will need to study patients who do not have sleep apnea versus patients who do have sleep apnea.

I plan to return to review insomnia many times and in many different ways. I have seen a commercial on TV for Belsomra  (suvorexant). The US FDA approved Belsomra (suvorexant) on August 13th, 2014 for insomnia according to one of its press releases, and it needs to be studied in this same fashion, with quantification.

The future of quantifying medical treatments is bright! Quantification will save time, lower costs, and create transparency. Our system will become more and more automated; people will be able to share their work with each other, and each step will be able to be validated and verified.

Follow Treatment Scores:

DISCLAIMERS: You must consult your own licensed physician, or other licensed medical professional, for diagnosis, treatment, and for the interpretation of all medical statistics including Treatment Scores. Treatment Scores are for educational purposes only. Treatment Scores may be incomplete, inaccurate, harmful, or even cause death if used for treatment instead of consulting a licensed medical professional. No medical advice is being given. We DO NOT CLAIM to cure, treat, or prevent any illness or condition. Nor do our services provide medical advice or constitute a physician patient relationship. Contact a physician or other medical professional if you suspect that you are ill. Call emergency services (call 911 if available) or go to the nearest emergency room if an emergency is suspected. We are not responsible for any delays in care from using our website, our services, or for any other reason. We are not responsible for any consequential damages of any nature whatsoever. We make no warranties of any kind in connection with our writings or the use of or Treatment Scores are about what happened to patients studied in the past; they do not predict the future.

Copyright © 2015 Treatment Scores, Inc.

No comments:

Post a Comment