The Collison Newsletter December 2010



Gaining an understanding of the different types of research, and which types of research are most trustworthy, can help you decide how much weight to give reported studies on food and health.


How frequently we read in the press, and see on television, the latest breakthroughs in research that are going to cure almost every disease known to mankind. An exciting new finding makes head-lines, often creating a misleading impression that the results are definitive, when it may only be some preliminary study in rats and when, and if extrapolated to humans, it will be 5, 10 or more years before it may have a therapeutic application.


Good science nurtures hunches and observations. Even the most promising ideas must work their way through a hierarchy of studies before the researchers and scientists can draw firm conclusions. To be considered reliable, findings must be reproduced by other studies and in different groups of people.


How does the relevance of the evidence and findings progress to clinical application?  There are laboratory studies, observational studies and experimental studies.

Laboratory Studies 

Experiments carried out in test tubes, or laboratory trials involving animals, can suggest how and why the underlying biochemistry might work, but the findings do not automatically apply to people.

Observational Studies 

These studies are on people. Researchers follow large groups of people, often for prolonged periods up to a decade or more. These are epidemiological studies. Examples include the well known Framingham Heart Study, which has been going for decades. Using questionnaires, and other methods including face to face interviews, scientists collect data at regular intervals as thousands of participants simply live their lives.


Most risk factors cannot be directly tested in people, but in these prospective studies, by comparing those who stay healthy with those who become ill or develop disease states, scientists try to identify factors that could account for the difference. There are various types of epidemiological studies called cohort, longitudinal, prospective and case-control studies.


Retrospective studies can also supply valuable information.

Experimental Studies 

These are studies in people where the researchers control what happens. In the field of nutrition and health studies, this usually means testing a diet or behaviour change. These are called ‘clinical trials’.


In clinical trials, experimental studies start with a small number of participants and, if successful, are repeated with more and different groups of people.


Within this group are randomised controlled trials. If correctly carried out, these are considered the gold standard - the most credible studies of all. Volunteers taking part in these trials are randomly assigned either to a group that tests an experimental drug, a dietary supplement, or another treatment or assigned to a control group where they receive a standard treatment such as a proven drug or a placebo (an inactive ‘look-alike’) or a diet for comparison. If possible, both the volunteers and the researchers are ‘blinded’, meaning they do not know who is in which group until the end of the study. A ‘double-blind crossover’ study means that half way through the study the groups cross over, for example if on a placebo in the first half of the study, the active substance is given in the second half, again with neither the volunteer or research worker knowing until the end of the study.


There are a number of factors that can give valuable information or provide cautions in understanding and interpreting the research procedures outlined above and the results that are subsequently recorded and eventually made public, either in reputable publications (ie peer-review journals) or sensational news-releases.

Systematic Review

This is a helpful analysis where researchers conduct comprehensive searches of previously published studies, evaluate the better quality evidence and summarise large amounts of information. If the studies are similar enough, researchers may include a meta-analysis, which combines and re-analyses the data from several studies. An excellent example of a meta-analysis is the topic of my March 2008 newsletter Prevention of Cancer. Food, Nutrition, Physical Activity and the Prevention of Cancer.

Recall Bias

Observational studies often suggest a link between a nutrient and a certain disease. But it cannot be said that the nutrient actually caused or prevented the disease. Beware of headlines that indicate otherwise: “Soy protects heart”, “Red meat causes cancer” and so on. These studies can be tainted by recall bias. This is when patients with a disease process are asked to describe their diet - they often recall a worse diet than the one they actually followed. Healthier people tend to see their diet through rose-coloured glasses. Routine questioning (ideally face to face) of the same people over time can avoid recall bias. These studies require large numbers, take a long time and sometimes rely on imprecise questionnaires.

Metabolic Studies

These are a type of clinical trial and typically involve a small number of volunteers who eat specially prepared meals for short periods and are tested at regular intervals. These studies are rigorous and closely controlled. They show effects on risk factors like blood pressure and cholesterol, but are usually too brief to show actual prevention of disease.


How well do the participants follow instructions?  Randomised controlled trials can be extremely difficult to conduct because researchers have to control the behaviours of large numbers of people, often for many years. If participants do not stay with the assigned treatment, findings could be blurred or even lost.

Quality of Studies

Systematic reviews often pool results of all available observational or experimental studies, but like other forms of research, they vary in quality. They are only as good as the studies that have been included. In the Cancer Prevention study referred to above, almost a half a million publications were reduced to 7000 reputable, and hence acceptable, ones and these were used for the meta-analysis.

Potential Risks

Finally, it is also important to consider whether the possible benefits to your health outweigh potential risks.

Before making a change in diet or life-style, or deciding to take supplements or medications, obtain as much information as you can and, with the above overview, recognise what to look for in making your decision.


*Copyright 2010: The Huntly Centre.

Disclaimer: All material in the website is provided for informational or educational purposes only. Consult a health professional regarding the applicability of any opinions or recommendations expressed herein, with respect to your symptoms or medical condition. 


Back to the list  Print friendly version