Further Evidence of MMR Vaccine Safety: Scientific and Communications Considerations

Abstract

It has been 20 years since, in a subsequently retracted study, Wakefield and colleagues (1) reported on 12 children with developmental delay, 8 of whom were diagnosed with autism within 4 weeks of receiving the measles, mumps, rubella (MMR) vaccine. Despite substantial limitations, the study received wide publicity, and the claims published in the article contributed to damaging confidence in the safety of the MMR vaccine, leading to a sharp drop in vaccination rates in the United Kingdom and, possibly, in other countries. Since the publication of the initial report by Wakefield and colleagues, and despite many subsequent studies not finding an association between MMR vaccine and autism, public concerns regarding a potential link between MMR vaccine and the development of autism have persisted. In one of the largest studies to date, Madsen and colleagues (2) conducted a retrospective analysis of 537303 children born in Denmark between 1991 and 1998, representing 2129864 person-years, to assess a potential link between autism and receipt of MMR vaccine. They concluded that MMR vaccine was not associated with development of autism and that the risk for autism in the group of vaccinated children was the same as that in unvaccinated children. This study also demonstrated that there was no association between the child s age at the time of vaccination or the time since vaccination and the development of autism. Moreover, in a recent meta-analysis involving 5 large cohort studies (n= 1256407 children) and 5 casecontrol studies (n= 9920 children), there was no relationship between MMR vaccination and autism (odds ratio, 0.84 [95% CI, 0.70 to 1.01]) (3). In this issue, Hviid and colleagues (4) report another nationwide cohort study from Denmark. They used a Cox proportional hazards regression model to evaluate whether receipt of MMR vaccine increased the risk for autism in children born between 1999 and 2010. They captured 5025754 person-years of follow-up and estimated an adjusted hazard ratio of 0.93 (95% CI, 0.85 to 1.02) among children who received MMR vaccine compared with those who did not. This is one of the largest studies published on MMR vaccine and autism. Hviid and colleagues included 6517 children with a diagnosis of autism in their cohort by using the Danish Psychiatric Central register. This allowed the investigators to define risk for autism in subgroups of children who may be more susceptible to developing this condition, such as those with a sibling history of autism. One critique of previous studies is a supposedly inadequate focus on subgroups of children considered by some, not always on the basis of evidence, to be at a high risk for autism. Therefore, it is important to note that Hviid and colleagues observed no increased risk for autism in several subgroups: children with a sibling history of autism or those who received other childhood vaccines and certain time frames after receipt of vaccine. In the context of this new study, and given that mainstream studies have consistently pointed toward a lack of association between MMR vaccine and autism, a couple of questions arise. First, is there sufficient uncertainty to warrant additional studies? Second, what impact will the accumulating evidence refuting an MMRautism association have on the public perception of vaccine safety? In an ideal world, vaccine safety research would be conducted only to evaluate scientifically grounded hypotheses, not in response to the conspiracy du jour. In reality, hypotheses propagated by vaccine skeptics can affect public confidence in vaccines. Therefore, in some cases, investigators continue to add to the evidence base on safety, even after clinical equipoise has been largely resolved. This may be justified if the cost, including the opportunity cost, of these studies is not too high. For example, analyzing an existing administrative or surveillance databaseas was the case in Hviid and colleagues studyis likely to require substantially fewer resources than a prospective study. Irrespective of the absolute costs, the opportunity cost of this research should be kept in mind: For example, continuing to evaluate the MMR-autism hypothesis might come at the expense of not pursuing some of the more promising leads. Even in the face of substantial and increasing evidence against an MMRautism association, the discussion around the potential link has contributed to vaccine hesitancy. As a result of a successful immunization program, measles was declared eliminated in the United States in 2000 (5). However, misinformation and the reluctance of many parents to vaccinate their children contributed to the 2015 measles outbreak in Disneyland in California (6) and the 2017 outbreak in Minnesota (7). Five outbreaks have already been reported in 2019, in which the majority of cases were unvaccinated (8). Therefore, generating evidence on MMR vaccine safety may be useful but is certainly not sufficient. It has been said that we now live in a fact-resistant world where data have limited persuasive value. So how do physicians and public health officials debunk the MMRautism myth? An approach similar to that proposed by Cook and Lewandowsky (9) may be useful in addressing the misperceptions regarding vaccines and autism. First, any myth should be clearly labeled as such. For example, there is evidence that a misleading headline can induce a reader to remember the inaccurate information while discounting the correct information presented subsequently. Second, while confronting the erroneous information, the focus should be on a few key facts: It is not essential to rebut every piece of misinformation. Finally, an alternative explanation of the perceived phenomenon should be provided. Otherwise, the individual can revert to their original erroneous belief. A recent book provides talking points to clinicians for addressing common vaccine safety questions, including questions around vaccines and autism, based on the latest vaccine safety literature and these and other evidence-based communications approaches (10).