I know that this will be the second autism post in a row, but this has been on my mind for a while, so bear with me!!
I want to talk about gluten (and casein and soy....) and the "lack of evidence" regarding the efficacy of a gluten-free, casein-free (GFCF) diet to address some of the symptoms of ASD (and ADHD).
Every parent of a child with autism has heard of the GFCF diet. Anecdotal reports from parents/teachers/therapists claim either miraculous results or no effect -- the diet either works or it doesn't. But -- when it works, it WORKS. Eye contact improves, language skills explode, impulsivity decreases, and attention skyrockets. The distinction between kids who respond and kids who don't is generally explained (based on anecdotal evidence) by the presence or absence of gastro-intestinal and/or atopic conditions.
Basically -- if you have an autistic child with gastrointestinal complaints and/or symptoms of allergies, there is a strong likelihood that you will see an improvement with a GFCF diet. If you have an autistic child without these complaints or symptoms, the diet is less likely to work. Again -- all based on anecdotal evidence.
So, what's up with the anecdotal evidence?
Anecdotal evidence is essentially evidence presented in the form of testimonials. This type of evidence is in contrast to traditional scientific evidence which requires statistical confirmation in order to be considered acceptable.
There is very little traditional, statistically-supported evidence for the efficacy of a GFCF diet to treat the symptoms of autism. Consequently, the vast majority of physicians dismiss its value and actively discourage parents from attempting it. My own experience with Ean's doctors has ranged from benign condescension to an angry admonishment that my son was "....going to grow up to hate me".
Still, no lack of statistically-supported evidence is going to deter a parent who has seen his or her child transform on this diet. Ean's transformation happened both physically and developmentally. Prior to starting the diet, Ean exhibited symptoms of severe allergies (allergic shiners, Denni-Morgan lines, inf lammed nasal membranes, constant mouth breathing, asthma, etc.), all despite the fact that extensive allergy testing revealed no significant allergens. He had idiopathic inflammation in his finger joints -- to the point where we had to tape his fingers straight before bed or else the joints would lock and he would be unable to straighten his fingers. He also suffered from severe and chronic constipation, resulting in bloody stools and hemorrhoids. Every single one of these symptoms resolved within 8-weeks of being GFCF. Every. Single. One. The developmental benefits were no less dramatic. Rich and I both cried the morning, after only 2 full days on the diet, when Ean made eye contact with us. His attention and focus improved, as did his memory and ability to follow directions. His teachers and therapists were uniformly stunned.
When I share my anecdotal evidence with physicians, they pat me on the head and tell me that there is no "real" data to support the diet. They tell me that whatever I am seeing must be a coincidence or my imagination. And I'm not alone in this experience. I have spoken to countless parents who are equally amazed at the effect of the diet, and equally as dismissed by their children's physicians.
These dismissals aren't coming from a place of malice. Physicians aren't trying to thwart our attempts at treatment. They generally want to best for their patients and their families.
So what gives?
The reality is that (with a few exceptions) the studies DON'T support the efficacy of the diet. Given that physicians are trained to practice medicine based on the results of medical research, it makes perfect sense that they would uniformly dismiss treatments that have not been proven (and have actually been disproven) by standard research methodology.
But, the fact remains that thousands and thousands of families report dramatic benefits when they institute a GFCF diet. And if you've ever met a parent of a child with autism, you know that these parents are among the most well-informed researchers and advocates out there. Rich and I are no exceptions -- we are both bright and well-educated. We both have backgrounds in science. We are both well-versed in statistics. We understand research methods and the importance of the scientific process. We are not superstitious, desperate, or grasping at straws. Like thousands and thousands of other families, we know -- unequivocally -- that the GFCF diet has helped our child.
Again -- what gives? Why are there thousands of well-informed parents claiming one thing while thousands of physicians are claiming the opposite? To answer this question we have to know a little about medical research and statistics and a little about the nature of children with autism.
Let's start with statistics and research design. Research studies, particularly research studies using human subjects, are very expensive and very time-consuming. It is for this reason that a significant number of medical studies are sponsored by the pharmaceutical industry (this is the work I used to do). Pharmaceutical companies have the money and manpower to sponsor and/or conduct large-scale, well-controlled studies. No other type of organization can come close to matching their resources. And, because pharmaceutical companies are for-profit organizations, they are only going to make significant investments in studies that will benefit their interests. No judgements here -- it's just the way capitalism works!
So here's our first point -- the only industry realistically capable of conducting a large-scale, well-controlled study has no vested interest in such a study. In fact, one could argue that they have a vested interested in NOT conducting such a study, since positive results may threaten profits by decreasing the number of children on behavior-modifying drugs.
The handful of studies that have been conducted have all be very small -- with the number of participants ranging from 9 to 72. The small number of participants presents a HUGE problem when it comes to a concept called statistical power.
In the simplest of terms, statistical power is the likelihood that a research study and accompanying statistical analysis will detect an effect. Think about a microscope -- if you try to see a single cell on a low-power microscope you will look and look and then report that there's nothing there. It's not until you turn up the power and use a greater magnification that you will see the cell that was there all along. An underpowered study will fail to detect an effect -- even if that effect is actually there.
There are several factors that affect statistical power, particularly number of participants (aka, sample size) and the amount of natural variability in the sampled population. Both of these factors are compromised in the existing GFCF studies.
As discussed above the agencies most interested in investigating the effects of the GFCF diet do not have the financial resources to conduct a large-scale study. Because of these limited resources, the published studies are based on very small sample sizes, resulting in woefully little statistical power.
A lack of power resulting from a small sample size wouldn't necessarily be a problem if there wasn't so much variability among children with autism. Have you ever heard the phrase "if you meet one child with autism, you've met one child with autism"? What that means is that there is no single face of autism. Two children with the very same diagnosis can present so differently that they may as well be from different planets.
What does that have to do with statistical power?
Well -- imagine you were doing a study on the effect of sun exposure on fair- versus dark-skinned people. You wouldn't need a big sample size to determine that fair-skinned people burn more quickly than dark-skinned people, right? You could figure that out pretty quickly. Staying with that logic, if children with autism were a fairly homogeneous group, you wouldn't need a very large sample size to determine the efficacy of a particular intervention. But - children with autism are a VERY diverse group, which means that researchers need to use a sample size that is large enough to compensate for that (through increased power).
It is also worth noting that in addition to the natural diversity seen in the autistic population, a study of dietary interventions automatically lends itself to a significant amount of compliance variability. Removing gluten and casein from a diet is not as simple as remembering to take a pill every day. The complicated nature of the diet means that there will be "user errors" which will further increase the variability within the sample and, thus, the need for a sample size large enough to sufficiently power the study.
The sample sizes in the published studies are minuscule by all medical research standards. In other words, the studies are completely and totally underpowered. An underpowered study will fail to demonstrate an effect -- even if the effect is very, very real.
So, what's my point? Basically, when your well-meaning doctor tells you that the GFCF diet has no scientific merit, he is basing that conclusion on studies that are fatally flawed. The only way you will know if your child will respond to the diet is to try it. Find someone who has done it and pick their brains. If you don't know anyone in your community, go on-line. There are countless groups and countless parents willing to talk you through it. Ask for advice, shopping lists, menus, and support.
And if it works for your child, then be ready and willing to lend a hand (and a statistic text book) to the next unsure mom or dad headed down this road.
Comments