A vigorous debate erupted this week in the Sports Medicine and Health worlds after my old colleague and laboratory head during my long past University of Cape Town days, Tim Noakes, published some ideas of his (and his collaborators) that exercise did not work to reduce weight if one did not improve one’s diet concurrently. This idea of Tim’s followed up other controversial ideas he has put forward during the last few years that carbohydrates are bad, and that one should eat a high fat diet and reduce carbohydrates to almost zero to optimize health. He has gone even further and suggested recently that one’s exercise capacity is optimized by a high fat diet as compared to a high carbohydrate diet, and furthermore that children should be ‘weaned’ onto a high fat diet from an early age, if I heard his message right. The problem for a lot of folk is that for more than 20 years before this, Tim strenuously advocated a high carbohydrate diet, and that carbohydrates were ‘king’, before having his epiphany and change of heart on diet, due apparently to himself being diagnosed as suffering from Type 2 diabetes a while ago and trying to alter his own diet as a result of this. Whether Tim is right or wrong with either his ‘old’ or ‘new’ messages, if all this is confusing for us scientists, I can imagine how difficult it must be for the general public to understand what is ‘right’ in the nutrition field, and why most folk often view the messages scientists propose with a high index of suspicion. So why does research and science often seem to confuse any issue rather than clarify it?
Research is defined as an endeavour to discover new, or collate old facts, by the scientific study of a subject or by a course of critical investigations. Research is not a new endeavour, and indeed most of any civilization’s progress must have been based on the testing of a hypothesis, or a change in lifestyle, and then accepting that change as custom / habit if it improved living conditions or quality of life. However, although research and science developed into a distinct entity with its own rules and way of doing things over hundreds of years, until the 20th century no detailed statistical testing was applied to any findings, and researchers and scientists merely reported individual case studies, or pooled data from a number of different case studies, with no attempt to control any variable, and made deductions from them. In the early part of the 20th century, with the advent of inferential statistics, researchers and clinical scientists began to believe that these case report studies were too simplistic, and possibly biased or prejudiced by external influences which were not reported or controlled for. From the 1930’s to the 1960’s, research methodology was technically improved, and studies were required to be well controlled before they were accepted for publication in scientific journals, which is and always has been the principle ‘vehicle’ of how research findings are made available to the scientific community and general public. Participants were required to be chosen randomly, variables well-controlled, and prospective replaced retrospective studies as the ‘gold standard’ for scientific / clinical research trials. Journals are now rated on the quality of research they publish, and studies that randomly select subjects, are ‘double-blind’ (meaning that neither the research subjects nor scientists know about the contents of a particular intervention or drug being tested during the trial), and examined participants in a controlled environment, have a greater chance of being published in a ‘quality’ journal than do studies that do not have these characteristics.
Unfortunately, this rigidity, while commendable, has led to problems of inference to the general public of the findings of such ‘gold-standard’, well controlled studies. In such trials, an individual is taken from their normal environment and turned into a laboratory ‘rat’ for the duration of the study / clinical experiment, and every possible variable is kept the same for each person in the trial, and any transgression of the ‘order’ of the trial leads to the expulsion of the participant from the trial. But human beings are not laboratory rats (and rats I am sure would say that not all laboratory rats are similar!). For example, during the ‘best’ drug or nutrition trials, participants are sometimes required to stay in a hospital or laboratory for the duration of the trial, and must all eat the same food and follow the same daily routine. The results assimilated from these artificially constructed environments are then published, and for example, a new drug or diet, based on the success of the trial, is approved and marketed to the general public. However, all individuals in their normal life will be following their own routines, probably with alcohol and perhaps tobacco consumption and unique diet, and to apply the average results derived from the original trial to this general public may be misrepresentative. All one can say with certainty after a well-controlled ‘gold standard’ trial is that in the specific group studied, in the specific environment where the test occurred, the drug or diet tested produced changes that were different from those produced by the placebo or other drug or diet it was tested against.
A further problem with extrapolating research results to the general public is that researchers generally produces average values (means) to describe the outcomes for the group of individuals involved in a specific trial, and then suggests to the general public that their data averages / the results of their trials are applicable to the general public from which they were originally selected. But human beings are individuals and not averages, and it is therefore problematic to extrapolate averages from clinical or nutritional (or indeed any) trials to larger population groups. For example, to tell a patient or person who follows a particular diet that research has shown that they have a 30% chance of dying younger than those following another type of diet is fraught with potential error – each individual’s date with death could be anywhere along the bell-shaped mortality curve, independent of whatever diet they choose to eat in their daily life. Indeed, they may live to be 90 or 100 whatever they eat if that is what their genetic and physical program entails (and a good degree of ‘luck’ / chance is also perhaps involved given how many things can kill one en route to one’s programmed life end time), despite there being a difference in average age of death between those who eat the diet being tested and that eating another diet. The problem of inference / extrapolation is made even more complex because of the fact that most research relies on the current research techniques / tests available to test a diet or drug at the time the trial occurs, and these tests available may be too simplistic to pick up potentially significant signs of toxic changes occurring in the individuals being tested. For example, 70 year ago, anabolic steroids were routinely used in large doses to cure ‘melancholia’ and depression. A few decades later scientists and clinicians started realizing, as their assessment techniques improved with time, that anabolic steroid drugs in large does could cause liver, heart and lipid profile disorders, and might even lead to dependency and psychological dysfunction. Therefore, the clinicians and researchers who initially recommended the therapeutic use of anabolic steroids were unwittingly placing their patients at risk, because of a lack of adjunctive knowledge and simplistic research techniques available to inform their decision at the earlier historical time points.
Therefore, for all the reasons above, even the best researcher has to be cautious about extrapolating their findings to the general public, due to the inherent weaknesses of even the ‘best’ science, and a good scientist would always be cautious when announcing their findings, and when availing themselves to the press to discuss them. The problem of course is that a scientist’s ego often gets in the way, and they often rush to announce their findings, or perspectives on a finding of another scientist, to get their name in the press, and get themselves ‘in the spotlight’ as much as possible. A lot of science, as nicely put by Ad Lagendijk in an article in Nature a few years, involves aggressive men (this gender focus was specifically used in the article) ‘fighting for their scientific claims to, at best, miniscule advances’ in scientific knowledge, with ‘territorial behavior’ underpinning their claims, and with ‘successful scientists incessantly travelling round the world performing their routine (speeches about their findings and theories) like circus clowns – forcefully backing up their assertions over what their contributions are to the latest scientific priorities and findings’ (sic). While this is a somewhat gloomy / negative assessment of the behaviour of a lot of (male) scientific folk, there is potentially a grain of truth in this, and unfortunately some scientists are perhaps too quick to engage with the public about their own findings, in order to assuage a huge / weak ego, and this can only complicate a research ‘message’ even more than just the methodological issues described above. The chance for a scientist’s ego to be more involved than it should is made even easier in modern time with the advent of the internet and publication vehicles beyond that of the routine scientific peer-review process (though the peer-review process is not without its faults either) where any work published is peer-reviewed and too extreme conclusions are required to be ‘toned down’ before publication, such as Blogs, Twitter, Facebook, and other methods where scientists now can very easily speak directly to the public, and in some ways can cause more ‘harm’ than good potentially by doing so.
So going back to the initial discussion of my old colleague Tim’s ‘flip flop’ on his diet message from high carbohydrates to high fat as being the best diet, and now that exercise may not be of benefit for weight loss (if I have read his latest message right, though it is very opaque what exactly is being said), folk should understand that scientists do have their own opinions and biases, do enjoy being in the limelight and press, and even if they have generated their conclusions on the basis of their own lab data, this data itself may be ‘flawed’ due to the reasons described above to no fault of the researcher themselves beyond not being cautious enough about their data. Therefore, all scientific information received by the public from scientists, particularly those scientists and clinicians that ‘shout the loudest’, should be taken with a strong ‘pinch of salt’ and heard with caution. For scientists themselves, we have to keep on remembering when venturing out of our labs to engage with the public that every word we say is ‘fraught with danger’ / has the potential to be incorrect or proved wrong in the future, and that the goal of every scientist and clinical researcher should perhaps not be to make a big new finding, but to be always dispassionate and cautious, no matter what one finds or concludes from ones time in the laboratory. As a friend and colleague of mine (and great scientist), Dr Angus Hunter once advised me, those that are the most passionate in science are the worst, and those that are most cautious in their message ‘the best’ (or words to that effect!). There is perhaps some truth in these words, at least when engaging with the public as a scientist, or conversely, when the public is trying to work out the validity of what a scientist / science is saying. Eat fat? Eat carbs? As the old wise words always suggest, moderation and balance is surely always best, and this goes not just for eating choices, but also for when scientists and clinicians engage with the public about their research work and findings!