By Robert Verkerk PhD
Founder, executive & scientific director
Studies comparing health outcomes in vaccinated and unvaccinated children are often considered the ‘holy grail’ when it comes to understanding possible associations between vaccination and specific health risks. This is especially the case with conditions that affect the neurological and immune systems that appear most sensitive. That of course includes autism. Getting to the bottom of this public health question is relevant to the vast majority of parents who have a right to know about the safety, or otherwise, of childhood vaccines prior to them being given to their most precious chattels. Few members of the public might understand that as long as causative evidence is not available, vaccine protagonists will continue arguing there’s no evidence that vaccines cause autism or other neuro-immune conditions, while ignoring the rapidly growing body of evidence showing associations, both from human and animal studies.
Incredibly few such studies have ever been published – and attempts to publish them have been met with violent reactions that typically result in the papers being retracted. This famously happened to Wakefield’s 1998 Lancet paper. But the process is ongoing, albeit less conspicuously and without justification from a kangaroo court such as the General Medical Council (GMC).
The Mawson papers
In November 2016, a vaccination vs unvaccinated study based on self-reported questionnaires of families of home-schooled children was published online. A flurry of responses on social media caused the journal to rethink its review process. The result was a retraction of the paper.
Retraction is not an option for journal editors if they simply don’t like the findings of research – or, worse than that, they’re worried their journal will be brought into disrepute because others don’t like it. It’s a device that protects the integrity of the scientific process usually where there is clear evidence of unreliable or erroneous scientific outcomes, either because of the author’s misconduct (e.g. data fabrication) or honest error (e.g. miscalculation or experimental error). Or there may be plagiarism, unethical research or other serious breaches of scientific integrity. Retraction is not a device that prevents publication of studies that represent low-quality evidence, that is at odds with the high-quality evidence that comes out of placebo-controlled randomised studies that are viewed as the pinnacle of the evidence-based medicine hierarchy.
Under the stewardship of lead author Professor Anthony Mawson, from the Department of Epidemiology and Biostatistics, School of Public Health, Jackson State University, Mississippi, the paper was re-submitted to another journal, Journal of Translational Science, in March 2017. Prof Mawson is an epidemiologist with what can only be seen as a solid curriculum vitae. The paper was accepted on 21 April 2017 and published online 3 days later – before being retracted soon after.
Professor Anthony Mawson, lead author of two retracted studies on health
outcomes of vaccinated vs unvaccinated children
Centrally, the paper concluded:
“…in a final adjusted model with interaction, vaccination but not preterm birth remained associated with NDD [neurodevelopmental disorders], while the interaction of preterm birth and vaccination was associated with a 6.6-fold increased odds of NDD (95% CI: 2.8, 15.5)…While vaccination remained significantly associated with NDD after controlling for other factors, preterm birth coupled with vaccination was associated with an apparent synergistic increase in the odds of NDD.”
A second paper was also submitted to the Journal of Translational Science. The study was a cross-sectional study of vaccinated and unvaccinated children based on the same survey data, and aimed to also compare outcomes in children who were born prematurely against those born full-term. This second paper was also retracted — and we await detailed reasoning from the journal. 18/05/17 – please see update below.
These studies are the results of analysis of self-reported survey data. By definition, therefore, they are a long way removed from randomised controlled trials (RCTs) which could never deliver the kind of overview of health outcome trends an epidemiologist such as Mawson might wish to investigate. They are what they say on the tin: broad brush strokes of information, a starting point for investigation that is commonplace in science, that helps inform more targeted research in the future. In terms of methodologies, the authors of the studies are not trying to deceive their readers. They used a number of methods to reduce the risk of bias. This was declared in one paper as follows:
“….minimizing potential bias in the information provided by mothers, all communications with the latter emphasized neutrality regarding vaccination and vaccine safety. To minimize recall bias, respondents were asked to use their child’s vaccination records. To enhance reliability, closed-ended questions were used and each set of questions had to be completed before proceeding to the next. To enhance validity, parents were asked to report only physician diagnosed illnesses.”
The greatest weakness of any such study is that it relies on self-reporting by parents. The authors recognised and discussed this weakness, arguing that there was no means to cross-check the results with medical records because the survey was anonymous, presumably to protect the subjects and families and also help reduce biases. ‘Low quality’ findings they may be, ones that result from misconduct, negligence or falsification, they definitely weren’t.
The second paper investigating parents’ self-reported health status of major health outcomes from surveys is of a generally higher quality than the first. However, because it investigates these outcomes in the same vaccinated and unvaccinated samples using the same data set as the first paper, and then singles out those babies that were born prematurely (prior to 40 weeks gestation) the numbers involved within the groups were smaller. However, small sample size is not a reason to not publish – or to retract – a paper. Outcome studies from case series of much smaller numbers (say n = 12) are not uncommon, and ironically that was the number used in Wakefield and co-workers’ retracted Lancet study. But the small number of cases was never in issue.
The number in Mawson’s study evaluating effects on preterm infants were larger. It was the results – not the the numbers – that almost certainly triggered the decision to have the paper retracted, following a cascade of predictable criticism. Here they are: 666 children in total, of whom 261 (39%) were unvaccinated and 51 (7.7%) were born preterm. The part that has been scrubbed from the scientific record is the fact that, of these, 50 children (7.5%), had a neurodevelopmental disorder (NDD).
Given the sensitivities and question marks over copyright ownership of the papers, we’ve summarised them here.
Imperfect science makes for a better science
Clearly, the surveys of homeschooled children is hardly going to be the basis of a definitive work that proves or disproves any link between neurodevelopmental disorders or NDDs such as autism spectrum disorder (ASD) and vaccines. However, it’s a good starting point as a means of examining this much neglected area of health outcomes in vaccinated as compared with unvaccinated children. The forces at work to prevent any publication of adverse findings about the overall vaccine schedule can never be under-estimated.
There are clear limitations of the studies, these were declared, and these are not sufficient basis to have the papers retracted. At most, the journal could have recommended revisions to further emphasise the importance of the limitations. The publication of the papers would likely have triggered further, more in depth, research, the results of which would contribute further to our understanding of the relationship between such factors as the age of vaccination and preterm (premature) status.
As it is, the pro-vaccine lobby have done themselves no favours. The lobby has attempted to erase from the scientific record the results of surveys of vaccinated and unvaccinated children that have seldom been studied, yet are of vital relevance to public health. The effect is likely to cement even further the divide between those who support and are opposed to receipt of the full vaccination schedule. The debate remains as polarised as ever and the main losers are members of the public – and one of our most vulnerable groups in society – our children.
The two papers by Anthony Mawson and co-workers, retracted or not, provide suggestive evidence from the survey data that vaccination appears to be associated with an increased risk of middle-ear infection (otitis media), pneumonia, allergic rhinitis, eczema and neurodevelopmental disorders (including autism).
Professor Anthony Mawson’s greatest sin may have been that he’d previously uttered some considerate words about Wakefield’s Lancet paper of 1998. Regardless of these, his present work needs to be considered on the basis of its own merits. In a free world, we are all entitled to our own opinions, and to interfere with such freedoms does a great disservice to the scientific process on which we have become centrally reliant as a society.
18/05/2017 – update to story: – since publication of this story we have received further information from Prof Mawson stating that the original submission to Frontiers in Public Health was rejected and not retracted as reported by Retraction Watch. The apparent rejection of the two papers published by the Journal of Translational Science was in fact due to the links being taken down whilst the journal checked that a retraction had not taken place. The links to the two papers have now been restored.