I am not a proponent of many supplements; the fact is that human trials with adequate subject numbers is lacking in most cases. Too often supplements are justified with in vitro, animal, and observational evidence which then fail when they enter interventional trials. It is difficult to explain this to lay people effectively, who prefer generalized, simple advice.
So I am excited to see that the latest data visualization from InformationIsBeautiful is on the scientific evidence on a number of supplements from only randomized, placebo-controlled human studies. An interactive chart is available here, or a static image is below:
According to the website, the researchers looked at over 1500 abstracts on pubmed.org and cochrane.org for their data, then put all of the information into a Google Spreadsheet here, where all of the sources are available. The bubbles are created from data on the spreadsheet, so changes are easy to make. Several have been already based on comments on the site.
Clearly many of these are very debatable, and just examining abstracts isn’t very reliable. Dose, context, and other study design features are also difficult to consider in a visual form like this without over-cluttering. But I think it is a great start, and it gives people a quick and simple guide, better the majority of supplement sources on the web.