This article is something I've been thinking about writing for quite a while, now. In discussions about vaccines, the question of whether they relate to autism (they don't) and research, the idea of a study looking at vaccinated vs. unvaccinated children often comes up. Typically, the call for such research comes from those who are opposed to vaccines or who are scared or uncertain about vaccines, either completely or just in terms of how they are currently administered. Certainly, such a study would pretty definitively answer the question once and for all. Granted, there would still be some die-hards who, if the study were to show that vaccines do not cause autism, would come up with some excuse. They may say that the study was flawed in some unspecified manner. Perhaps they would assert some conflict of interest that may or may not be real. And I've already heard some state, before such a study has even been done, that the vaccines that the parents received are to blame. That last one could potentially go on and on ("well, it must be the vaccines the grandparents/great-grandparents/etc. received!").
At any rate, there are two basic ways to go about such a study. First, there is a retrospective epidemiological study. This kind of study would look at populations that already received or did not receive vaccines, taking the form of case-control studies or cohort studies. There likely would be no randomization. There would be no placebo control. Individuals would probably be matched by age and a host of other factors. There are a lot of potential variables, though, that could confound the results. As I'm not an epidemiologist, I won't get into the difficulties involved in this sort of study. René Najera over at The Epi Times has a series of posts [Unfortunately, an anti-vaccine activist's actions forced René to take his site down] on epidemiology that may help inform readers of the considerations involved in this kind of research. Prometheus also wrote about vaxed vs. unvaxed over at Photon in the Darkness [Sadly, Prometheus took down his site for financial reasons.] back in 2008, touching on the logistical nightmare of finding enough subjects to enroll.
The second approach would be to conduct a prospective, double-blind, randomized, placebo-controlled trial.
What does all that mean? Well, prospective means that subjects would be enrolled before any interventions (i.e., vaccines) were received; that basically means that kids would be enrolled from birth. Double-blind means that neither the subjects nor their parents, nor the researchers, would know whether the subject receives the actual vaccine or a placebo. And, randomized means that neither the subjects nor the researchers would get to decide which group (vaccine or placebo) they get put into. Basically, it's like flipping a coin to determine which group you're in. That could present a bit of a challenge for enrolling subjects: trying to convince someone who is against vaccines to have a chance that their kid could get one (and be exposed to whatever outcome they fear), and trying to convince those who are for vaccines to have a chance their child might not get one (and thus, be unprotected against a range of diseases). Finally, the placebo in such a study would be a shot, given in the same manner as a vaccine, except without any active ingredients in it (i.e., no antigens, no adjuvants like aluminum salts, no preservatives like thimerosal, etc.). Most likely, the placebo would be saline. Since autism is typically diagnosed around 2-3 years of age, we would need to follow the subjects for at least that long, but preferably at least five years. The subjects would need to come into the study office for scheduled administrations of vaccines (or placebos). There would be regularly scheduled evaluations and checklists to measure the children's development and to check for signs of autism, as well as to keep track of other illnesses the children may get.
Now we know what the study would look like, as far as basic design. There are a few other considerations (e.g., how many subjects to enroll in each group in order to find an X% risk of autism diagnoses in each group), but we have the basics. So what's stopping someone from conducting such a study? I mean, assuming one gets past the logistical hurdles of funding, subject recruitment and enrollment, etc., why isn't this kind of study being done to answer the question, once and for all, whether vaccines and autism are at all linked?
The short answer is: ethics and legalities.
I'll expand on this in a series of posts, which can be found here. A lot of the information is going to come from some of the laws surrounding clinical trials in the United States, as well as some documents outlining ethical guidelines in the conduct of human subjects research.
In the interests of disclosure and so my readers will have an understanding of where I'm coming from, I volunteer on a committee that reviews research protocols to ensure adequate protections for the subjects. I also have a master's degree in Regulatory Affairs. I do not work for any pharmaceutical or medical device company, nor do I derive any profit from financial interests in a pharmaceutical or medical device company.
Addendum: Someone asked me on Twitter a question along the lines of, "What about the possibility of a currently unknown contaminant?" This person added that such a consideration would affect the ethics of the study. While a contaminant could conceivably play some role, when considering ethics, we must go with what we know. Certainly, study design could prepare for that possibility, but we cannot make a go/no-go decision based on complete unknowns. Hopefully this will be clear once I finish this series.
Vaccinated vs. Unvaccinated Part 2 - A History of Medical Research Ethics
Vaccinated vs. Unvaccinated Part 3 - Ethical Evaluation
Vaccinated vs. Unvaccinated Part 4 - The Legal Perspective
Complete Vaccinated vs. Unvaccinated Series