The continuing series of health claims rejections is really raising the question of what’s going on here. Is EFSA really impossible, as has been suggested more than once? Is nutrition science all that impossible to come to terms with when it comes to providing convincing evidence for a physiological effect?. Or do we still have to get rid of a mind-set that feeds expectations for health claim approval that are based e.g. on ex-vivo studies only, studies that do not show any effect or just one pilot study?

In a way, health claim substantiation is a rather straightforward process. If a food (ingredient) has a beneficial physiological effect (relevant primary endpoint), a pilot study should indicate what effect size to expect, preferably in the target population. Expected effect size (and variation) drives the power calculation of the full scale trial.

Of course there still may remain methodological and statistical pitfalls. But by only using one’s common sense and investigative mind quite some sensible solutions may be provided for.

As a rule:

If your gut feeling tells you not to be too sure of the dossier – don’t submit!

If it needs a lot of discussions to decide whether or not to submit – don’t submit!

………. – don’t submit!

Only if you feel more than fairly confident about the dossier, then you may submit.

If you don’t trust yourself, have an independent,  no-nonsense, presubmission dossier check!



  1. Hello Stoffer
    Clearly the need for pilot study data is critical for any sensible power size calculation and pivotal trial design but have you in your experience ever come across enough market data about the product to be able to calculate effect size? Your insights about European experience will be appreciated. I can ask my team for our experience so far.

    • stoffer says:

      Dear Jayesh,

      First I’d like to apologize for the late response. Apparently the alert did not come through.

      I’m not quite sure what you mean by ‘market data’. I interpret market data as data on sales, number of buyers, consumer satisfaction, recurrent buyers, etc. At best, some subjective measure of efficacy can be derived or inferred from these data.
      Power size calculation is about estimating the size (and variation) of the anticipated physiologically beneficial effect. That has to do with biomarkers measured by validated methods/techniques. I doubt whether market data, in my interpretation of the term, could serve that purpose.

      Do you have experience with or ideas about that?