Blinded by science, pragmatism forgotten.

Some of the best brains in North America have worked for over a year to produce the Institute of Medicine's (IOM) report on vitamin D (1) and so we expect expertise and wise words. However, despite great care in sifting evidence, careful application of method and a desire for transparency, the report has turned out to be very controversial. Its wisdom and method are now being questioned, as can be seen from the letters published in this issue of Public Health Nutrition. Of course expertise by its nature is always limited in scope and deep rather than wide. Experts often suspect the expertise of others because they understand the limitations and shortcomings of specialist knowledge. So in forming a committee of experts we generally try to include people who are not all expert in the subject in question. This, we hope, will provide a certain breadth of vision. Experts on the IOM's committee on vitamin D may well have been selected using such criteria but inevitably there were gaps. The committee has shown a marked preference for conclusions based on randomised controlled trials. We can all understand a bias of this kind because clinical trials, when they can be done, may produce very clear conclusions and settle issues of causality unequivocally. However, much of the information that has accumulated linking insufficient vitamin D and sunshine with disease is observational. This work has been built up over many years and provides a largely consistent story. The observational evidence has accumulated strength that goes beyond randomised trials because very different surrogate variables all tend to point to the same conclusion. These variables include exposure to sunlight, colour of skin, latitude, studies of disease in migrants, diet, and vitamin D assays of blood as well as laboratory studies and animal work. Latitude, for example, provides a measure of sun exposure throughout a lifetime and so might be expected to pick up long-term effects that cannot be discovered in a randomised trial that must be limited in duration. Evidence that we already have suggests strongly that insufficient vitamin D in pregnancy and early life is a cause of multiple sclerosis (MS); perhaps not a unique cause but one without which MS is unlikely to occur (2). This is supported by revealing experiments which show that vitamin D interacts directly with numerous sites on the genome, providing us with a new understanding of the nature–nurture …