Dietary disclosures: how important are non-financial interests?BMJ 2018; 361 doi: https://doi.org/10.1136/bmj.k1451 (Published 12 April 2018) Cite this as: BMJ 2018;361:k1451
All rapid responses
The contention that scientists must not allow their personal dietary
choices to affect their research (1) is flawed. It assumes that
despite the now vast body of literature that exists on the subject of
nutrition, we still know nothing about what the optimal diet for human
health and longevity actually is.
Individuals interested in this subject are often well versed: it is
right that we should hold opinions based on our personal appraisal of
the evidence to date, and we should apply these to our own dietary
choices as well as our selection of future study topics.
Since it is unrealistic to expect that we are one day going to arrive
at a unified conclusion with which everyone agrees (accompanied
perhaps by a blinding flash of light), we must address the question of
how new knowledge can be generated against the backdrop of that
We must also begin to address the slightly more tricky psychosocial
question of what might happen – what might be happening already in
fact – if the evidence starts to weigh in a direction most of us don’t
Schwab has quoted Ioannidis and Trepanowski’s five examples of dietary
views that could present conflicts (2), which include strict veganism.
He asked: “Are all diets conflicts of interest or only some? In
research about the vegan diet, should only vegan researchers
Think about it this way: we all have a 2000-calorie-per-day bucket,
and we have to decide what to put in it. If we choose 100 calories of
steak, we aren’t choosing 100 calories of tofu. If we choose 100
calories of tofu, we aren’t choosing 100 calories of steak. There’s an
opportunity-cost involved in choosing either.
This may sound obvious, but just because more people choose steak than
tofu doesn’t make them right, and it doesn’t mean they aren’t making a
choice just because they occupy the majority position of the day.
Some people choose steak because they believe it’s healthy, or because
they believe it’s okay to kill and eat animals. Some religious groups
advocate that God placed animals on earth for human consumption. These
are considered positions, which create the potential for bias.
Equally, some people choose tofu because they believe it’s healthy, or
because they believe it’s not okay to kill and eat animals. Vegans
advocate that it’s wrong to kill animals or cause them to suffer
unnecessarily, and this has been recognised as a protected
philosophical belief, unrelated to any particular religion but
nevertheless cogent and serious. These are considered positions, which
create the potential for bias.
The only other group is people who eat what is put in front of them,
or what they most desire from what is put in front of them:
availability, habit, instinct and addiction are drivers here. Either
they’ve never considered health or morality, or their heads are in the
sand. They are not the ones who should be doing research.
As F. Scott Fitzgerald famously said: “the test of a first-rate
intelligence is the ability to hold two opposed ideas in mind at the
same time and still retain the ability to function”. The possibility
that what’s healthy may not be moral, or that what’s moral may not be
healthy, is an example of two such opposed ideas.
To suggest that nutrition science can occupy morally neutral territory
is unrealistic. The growing debate around veganism is illuminating in
this respect. Justification for the consumption of animal products has
long been predicated on their necessity for health. Vegan diets don’t
have to be shown to be superior to omnivorous diets in order to have
significant implications for our food system. They only have to be
shown to be equal for the conclusion to follow that the only remaining
reason for human consumption of animal products is pleasure.
I hypothesise that contemplating this possibility results in a level
of cognitive dissonance most omnivores find intolerable: a whole range
of psychological defences come into play. Labelling vegans as
“dangerous” and “other”, for example, and definitely not acknowledging
the data or commissioning any more studies that might confirm what we
are already beginning to know.
The greatest biases in nutrition science at the present time are not
related to the ways in which data is being collected and interpreted,
although these are the ones we read about because they relate to
studies actually being done. The greatest biases lie in the hypotheses
themselves: the questions being asked and more importantly, not asked.
The holy grail of researcher objectivity is enshrined in the null
hypothesis of no difference. However the null hypothesis of no
difference between the effects on health of the standard western diet
and the Mediterranean diet is likely to court significantly less
controversy than the null hypothesis of no difference between the
effects on health of the standard western diet and the low-fat raw
vegan diet, though both are equally valid scientific questions.
I’m a vegan paediatrician with a master’s degree in public health.
When I became vegan I had no idea what a pit of controversy I was
naively walking into. I just thought it made sense to extend the
dictum “do no harm” to non-human beings, and I knew enough healthy
vegans to know I wasn’t going to die.
I also have twin vegan daughters who are growing normally: they have
shiny hair, immaculate skin and perfect teeth. Their existence is a
challenge to anyone who wants to assert that animal products are
necessary for human survival, or essential for health.
Mozaffarian and Forouhi wrote: “the most extreme voices often drown
out the well informed” (3). This suggests that those with extreme
views are not well informed, or that the well informed will of course
hold moderate views. But moderation may not be best for human health
and longevity: so in addressing the bias in relation to studies being
done and not done, it is necessary to ask and answer the question,
what will happen if the vegans turn out to be right?
1) Schwab T. The whole truth: why non-financial biases matter. BMJ 2018;361:k1451
2) Ioannidis JPA, Trepanowski JF. Disclosures in nutrition research: why it is different. JAMA. 2018;319(6):547–548
3) Mozaffarian D and Forouhi N. Dietary guidelines and health – is nutrition science up to the task? BMJ 2018;360:k822
Competing interests: I have been a vegan for eight years. I have a certificate in plant-based nutrition from eCornell University and I write a blog on the subject: www.kidseatplants.com