My blog posts don't usually get a lot of comments. Many posts don't get any comments at all, so when I get a handful of comments on one post, I start to think that I have hit a resonant note with people. Well, either that or that I have had the same post at the top of my blog for almost a month. That DOES tend to allow more time for people to comment on it. Anyway, my last post on the misunderstood nature of an undiagnosed celiac, or undiagnosed anything, has prompted more shouts of "Amen, Brother!" than anything I have written in the past year.
As I write this, I am taking a break from writing experimental details for my doctoral dissertation, which I defend in July. Soon I will become Dr. ElwoodCity, and move to Home Town and join the Faculty of Science and Technology at Home Town College. For the past several years I have been working on the development of smarter imaging agents for the diagnosis of breast cancer. I mention this so you can see my comments in the context of someone who is clearly not "anti-science".
This is going to be a long post, because this is a big topic.
The thing is, there a limitations on diagnostic medicine, and those limitations contribute to the number of people who are undiagnosed for celiac disease. This is not in itself a problem. The problem is that far too many people don't understand what those limitations are, and so fail to even see them. If you don't see any limitations for diagnostic medicine, you are also going to be blind to the limits of what Scientists and Doctors to know about your health. Whatever they say must be both accurate and precise.
This was my main complaint with Kate Murphy's article. It implied over and over that if you have a positive test, you should be on a gluten free diet, because the test determines whether you have celiac disease. If you have a negative test, the only reason you would go gluten free is because you are buying into the newest fad diet. I'm sorry, but that just reveals who doesn't know what about science, and it isn't good.
The main thing that science tries to accomplish is the removal of uncontrolled variables. We figure out everything that may be affecting the outcome, and change them one at a time, keeping the others constant. That way we can say how each factor influences the outcome. In my field, that isn't more complicated than figuring out how you want to change a recipe. Cooks optimize recipes much the same way I optimize a chemical reaction. It is pretty simple when all the variables are things you control, like ingredients and temperatures. In the human body, it is just a tad bit more complicated.
The genetic differences in human populations, as well as lifestyle choices, will occasionally make the result of any diagnostic test worthless for one individual. This is why science preaches against relying on anecdotal evidence. Think for example if we were trying to decide upon a new set of dietary recommendations, and someone asked you what you ate in order to feel healthy? Would your answers be applicable at all to the people who live next door? To the majority of the population? No, because the people who frequent this blog have a completely different set of variables controlling how they interact with food than the majority of the population. Diagnostic tests are designed for the majority, not the individual.
The solution to this is to decide upon an acceptable threshold for false positives and false negatives. You take a population, divide them into groups of yes and no, responders and non-responders, and you give them the new test. Then you decide how to interpret the results so that the most people with a yes result are found, without including too many people from the no group. The result is that a certain number of people who should be yes are missed - a false negative. Some people who should be no are included in the yes group - a false positive. Every test has false positives and false negatives.
The current method of predicting response to hormone therapy for breast cancer patients has only a 5% false negative rate. That is really good. It has a false positive rate of 60%, though, meaning that if the test predicts that you will respond to hormone therapy, you have only a slightly better than 50/50 chance that you actually will. This is why I have a research topic. Shouldn't there be a way to improve that?
So here is the question - What is the false negative rate tests for celiac disease? Does anyone know? The gold standard test is endoscopy, which implies that the blood tests, while good, have either higher false negatives or higher false positives. To figure it out, we need to take a population and divide it into people who have celiac disease and people who don't, then perform endoscopies on all of them to see how the endoscopy results correlate with who has celiac disease, and who doesn't. Now, aside from the fact that that is a complicated study that no one cares enough to do, how are you going to do the initial division into groups? How are you going to select celiacs from "normal" people to have something to correlate your data to?
The only way I can think of is to put people on a gluten free diet, and see who feels better and who doesn't. Wait a minute... I think I've heard that somewhere before...
Oh yeah! It was in an article in the New York Times, by Kate Murphy. "The final proof is reversal of symptoms on a gluten-free diet." Thanks, Kate, for helping us know how we can know definitively whether we should be on a gluten free diet or not. The irony of the whole situation is that the people who claim authority for the diagnostic tests because they come from "science" understand the process less well, and is less scientific, than someone who just stopped eating wheat, barley, rye and oats (sometimes) to see if they would feel better.
So here's to you, undiagnosed celiac scientists!
(And this doesn't even address the issue of medical doctors, and their ability to think outside the box. That could be another whole post on its own, but I need to get back to work.)