Hyping it up over fluoridation

In my time as a scientific researcher, honest scientists used to condemn colleagues who over-hyped their science. To our mind there should have been a special place in hell for scientists who misrepresented their findings or dishonestly described their significance.

That sort of self-promotional behaviour is probably understandable for reasons of ambition – or even the attempt to secure future funding. And these self-promoting scientists usually moved on rather quicker into higher-paid administrative jobs. Not exactly to that special place in hell – but maybe their promotion away from active research reduced the damage their personal self-hyping could do to science in the long run (although I did often wonder about the damage they did to science with their administrative decisions).

A recent article (Hype isn’t just annoying, it’s harmful to science and innovation) got me thinking of this problem again – and to realise we are facing a classic case of this self-promotional over-hyping in recent science related to community water fluoridation (CWF).

Readers may pick up that I am referring to the behaviours of a north-American research group which has been indulging in a wave of self-promotion – a promotion which involves misrepresenting of their own findings and the significance of those findings. I have discussed the research findings of this group in a number of posts – including the following:

More recently they have produced a video promoting and misrepresenting the significance of their work. A video which is being gleefully used by anti-fluoride activists in their propaganda (there has been a bit of a dance over this video which has been roundly criticised scientifically and taken down or moved several times so links often don’t work. But a recent appearance was on the New Zealand anti-fluoride Facebook page).

Group members have also attacked, in a very unprofessional way, fellow scientists who have critiqued their work (see for example When scientists get political: Lead fluoride-IQ researcher launches emotional attack on her scientific critics). On social media, they have attempted to close down any critical discussion of their work – and in a similar manner, they purposely ignore, or even attempt to hide, studies that do not support their claims. (At the personal level I have had a member of this group refuse to fulfil their prior undertaking to do a peer review on a draft paper of mine – presumably because on reading it she became aware that my paper discussed flaws in their work).

In support of my contention that this group is over-hyping their findings, and unprofessionally using this misrepresentation to give support to anti-fluoride activists, I will briefly list below what their findings were.

No effect of CWF on child IQ

While claiming their findings support the claim that CWF is harmful to the brains of children they actually refused to even discuss their own reported data which shows this is not the case. In fact, the data in their two main papers (Green et al 2019 & Till et al 2020) show no effect of CWF on the IQ of children. This confirms the finding of Broadbent et al (2015) – the only other study comparing IQs of children from fluoridated and unfluoridated areas (see Canadian studies confirm findings of Broadbent et al (2015) – fluoridation has no effect on child IQ).

The table below lists their data together with that of Broadbent et al (2015)

A comparison of IQ for children and adults living in fluoridated and unfluoridated areas in countries where CWF is used

I think it unprofessional for this group to ignore their own data while at the same time lending support to activists who are claiming that CWF harms child’s brains. Perhaps they assume that this finding could not be hyped to promote their standing and ambitions. So, instead, they have diverted attention to another part of their work – the relationship between child IQ and measures of fluoride consumption.

Occasional weak relationships of child IQ with fluoride intake

While ignoring some other data – which is unprofessional in itself – they have devoted their promotional material to just one part of their findings – the few cases when they are able to demonstrate a relationship, albeit only a weak relationship, of child IQ with fluoride intake as measured by drinking water fluoride content, estimated fluoride intake or urinary fluoride levels.

I have discussed problems with this approach in my articles listed above but will stress here that the relationships are usually not statistically significant, or very weak when significant (explaining only a few per cent of the variance in IQ), and suffer from inadequate consideration of possible important confounders or other risk-modifying factors. A common problem with the sort of “fishing expedition” involving statistical searching of existing databases in an attempt to confirm a bias.

The figure below shows the relationships considered in the two studies. Most simply are not statistically significant. In a recent article (see Perrott, K. W. (2020). Health effects of fluoridation on IQ are unproven. New Zealand Medical Journal, 133(1522), 177–179) I describe it this way:

“Multiple measures for both cognitive factors and of fluoride exposure are used producing many relationships. Only four of the ten relationships reported by Green et al were statistically significant (p<0.5). Similarly, only three of the twelve relationships reported by Till et al were statistically significant. There is a danger that reported relationships could be misleading – as the proverb says, “If you torture your data long enough, they will tell you whatever you want to hear.” “

Relationships of cognitive measures with exposure to fluoride obtained by linear regression analyses using Canadian MIREC database. Red data points statistically significant (p<0.05) and green data points not significant. Bars represent 95% confidence intervals.

Even if the reported relationships correctly reflected reality (being a “fishing expedition” the chances are they don’t) their concentration on such weak relationships (explaining only a few per cent of the data) could be actively diverting attention away from the factors which are more important. Although this group has been very shy about making their data available for other researchers to check, the data they have published indicate that regional and ethnic differences may be making a much bigger contribution to child IQ.

A big problem (always glossed over by those promoting this work) is that the studies are exploratory, using existing data bases rather than experiments specifically designed to answer the relevant questions. Reported relationships may support preconceived beliefs but it is easy to ignore important confounders or risk-modifying factors (which properly designed experiments would attempt to minimize).

I highlighted the problem of inadequate consideration of other factors in my article critiquing an early paper from this group (see Perrott, K. W. (2018). Fluoridation and attention deficit hyperactivity disorder a critique of Malin and Till (2015). British Dental Journal, 223(11), 819–822). In this case, I showed that when regional factors (in thas case elevation) were included in the statistical analysis the relationship of ADHD prevalence with the extent of fluoridation that Malin & Till (2015) reported simply disappeared.

It is worth adding that in subsequent reports from this group my critique has been completely ignored and they still reported the flawed Malin & Till (20915) as being reliable. I think that is very unprofessional but it does align with the tactics of self promotion and over-hyping of their work.

The down-side of self-promotional hype in science

The article I introduced at the beginning (Hype isn’t just annoying, it’s harmful to science and innovation) finished by concluding:

“Acting this way has a cost. It’s not just about allowing people to feel awe: it’s about empowering those who are not professional scientists or technologists to be able to participate, instead of being spoon-fed a whizz-bang watered-down version of science as cheap entertainment. Hype doesn’t just obscure the reality of what’s going on in science and technology – it makes it less interesting. It’s time we start to look past it and delight in what lies beyond.”

So as honest working researchers we were right to resent self-promotional hype and, perhaps, to wish that a place in hell was reserved for these ambitious self-promoters.

But, looking back, I can recognize that scientists are human and, like everyone else, fallible. It is easy to see how people will place ambition over the truth and why that should resort to hyping their science for ambitious reasons. I can also recognise, as Ioannidis (2015) reported, that “Most Published Research Findings Are False.” I believe Ioannis is basically correct and there are big problems with the scientific literature which contains reports from so many studies based on an exploratory statistical analysis of the sort indulged in by this North American group.

It’s inevitable that such poor science will be seized on by those with political, commercial and ideological agendas to support their claims. This has been done by the anti-fluoride activist groups. For the rest of us it is matter of reading the scientific literature intelligently and critically. And I mean all the literature, not just that related to fluoridation, vaccination and similar “hot topics.”

And, in the end, the truth will out. Poor science and self-promoting ambition and hype do get exposed. The faults in the promotional messages do get exposed. And, new research and data usually provide context for a proper evaluation of the claims made by those which currently hype their work.

Similar articles

One response to “Hyping it up over fluoridation

  1. Making inappropriate claims about statistically non-significant results
  2. Making inappropriate recommendations for clinical practice that were not supported by study results
  3. Attributing causality when that was not possible
  4. Selective reporting, such as emphasising only statistically significant or subsets of data in the conclusions
  5. Presenting data in a more favourable light than was warranted, for example writing overly optimistic abstracts, misleadingly describing the study design and underreporting adverse events.
  6. These strategies are all common to the behaviour of this group of anti-fluoride researchers.

    Unfortunately, they are also common to another published studies – as the article says: “More than a quarter of biomedical scientific papers may utilise practices that distort the interpretation of results or mislead readers so that results are viewed more favourably.”