Daniel, the lions’ den and the earliest clinical trial
Epidemiology (from the Greek epidamia, the prevalence of disease) is the science of populations, but it’s too easy in epidemiology to confuse cause and correlation. So here is a claim of cause:
eat breakfast → eat more → yet lose weight, paradoxically
(or alternatively)
skip breakfast → eat less → yet gain weight, paradoxically
and here is a claim of correlation:
Epidemiological studies that look only at breakfast and weight can easily confuse correlation with cause – except that the science of epidemiology has long generated a ‘hierarchy of evidence’ by which to distinguish them, and it is a theme of this book that epidemiologists have not always been sufficiently rigorous in applying that hierarchy.
The hierarchy of evidence: Conflicts over diet are age-old, and some can be sourced to the Bible. Daniel was a Jew who had been captured by Nebuchadnezzar, the King of Babylon, and who was consequently condemned to various vicissitudes including the lions’ den (from which, happily, he was rescued). Daniel was also instructed to eat the food of the royal court, to which he objected on grounds of observance. Let Daniel I: 12–16 take up the story of how he asked that he and his fellow captives be given:
nothing but vegetables to eat and water to drink. Then [Daniel said] compare our appearance with that of the young men who eat the royal food, and treat your servants in accordance with what you see. So he [the chief official] agreed to this and tested them for ten days. At the end of the ten days they looked healthier and better nourished than any of the young men who ate the royal food. So the guard took away their choice food and the wine they were to drink and gave them vegetables instead [New International Version].
This was a clinical trial! The first to have been recorded. But though it wasn’t too badly controlled, we think we can do better now, and today we understand that some methodologies are more powerful than others and that they can be ranked in a hierarchy of evidence:
Systematic reviews and meta-analyses
Randomised blinded controlled trials
Randomised controlled trials
Cohort studies
Case-control studies
Cross-sectional surveys
Case reports.
Let me briefly look at these, starting with the weakest methodologies.7
7. Case reports: In such a report, the medical history of a patient is told as a story. ‘Mr Joe Blogs has always smoked and he has just celebrated his eightieth birthday, therefore smoking potentiates longevity.’ It doesn’t require genius to understand why case reports provide only weak evidence of cause and effect.
6. Cross-sectional surveys: These are ‘snapshots’. In such studies, people are asked two questions, which might be: what do you eat for breakfast and what is your weight? As I’ve shown above, many breakfast studies fall into this category, which is unfortunate because this sort of snapshot study can be very misleading, i.e. at any one time people may be large and thus skip breakfast while, later, those people may be slim and thus eat it, but it is not the eating of breakfast that makes you slim (and, vice versa, not the skipping of breakfast that makes you fat); rather, it’s being large that encourages people to skip breakfast, and being slim that encourages people to eat it. So cross-sectional or snapshot studies can lead to conclusions that are 100 per cent wrong.
5. Case-control studies: These are not used frequently in breakfast research, so I’ll not describe them here.
4. Cohort studies: These are an attempt at avoiding the problems of a ‘snapshot’. In a cohort study, two groups of people are selected because they either do or do not eat breakfast (say) and then some years later their outcomes are determined. During the 1940s, 1950s and 1960s Bradford Hill and Richard Doll performed their famous cohort study on doctors who either did or did not smoke, discovering that smoking causes lung cancer.fn1
3. Randomised controlled trials: Now we are moving from observations to experiments, where participants are given a drug or some other intervention (such as skipping breakfast or not eating the royal food) and scientists then determine the effect.
Experiments, though, are only as good as their controls: if you give a drug to a group of people and then get an effect, you need to know that those people were not going to produce that effect anyway, so in clinical medicine we do controlled trials, where the responses of subjects to a drug are compared to the responses of subjects who do not receive the drug. But the experimenter mustn’t pick the control subjects, because that might bias the results, so in clinical medicine we do randomised controlled trials, where the two groups of subjects are selected to be as similar as possible, with individuals being distributed between the two groups randomly.
2. Randomised blinded controlled trials: Ideally, to avoid subconscious bias, neither the experimenters nor the subjects of a trial should know who is part of the intervention group and who is part of the control group, but unfortunately I need not explain this any further as we can’t do blinded trials in breakfast: blinding requires that we provide control subjects with placebos, yet we can’t provide placebos for breakfast. Breakfast studies have therefore been deprived of the most robust experimental protocols, but as the science of astronomy illustrates, knowledge can progress without the full panoply of experimental protocols: if we are careful in our observations, we can – in the absence of experiments – show that the earth moves round the sun rather than vice versa, but we do have to view the observations carefully, without preconceptions.
1. Systematic reviews and meta-analyses: These are sophisticated words that describe the sophisticated methods by which the results of many different trials can be pooled, to provide more secure conclusions than any one trial can provide.
Conclusion: Clinical medicine has created a hierarchy of evidence, and in this book I try to show where breakfast epidemiologists have, unfortunately, ignored the hierarchy, to thus confuse correlation with cause.
I have explored the two major explanations for the apparent paradox of breakfast eaters consuming more calories than breakfast skippers while being slimmer; now let me offer five more:
1 Healthily minded people ‘know’ they should eat breakfast
2 People under-report their food intake
3 Breakfast skipping is not properly defined
4 ‘Kick-starting’ metabolism
5 Breakfast skippers are owls, not larks.
Let’s look at these in turn.
1. Healthily minded people ‘know’