Information is Beautiful. Visualizing the Evidence for Health Supplements.

21 03 2010

In a world driven by data, we need a simple means of digesting it all. Visualization of data may help to coop with the information overload. Good visualizations enable people to look at vast quantities of data quickly.

Bram Hengeveld at Geriatric Care (geriatricare.wordpress.com) told me of Snake Oil, a fantastic visualization of scientific evidence for popular health supplements. A well chosen name too, because Snake oil  is both a traditional Chinese medicine, as a  term for “medicines” that are fake, fraudulent, quackish, or ineffective. The expression is also applied metaphorically to any product with exaggerated marketing but questionable or unverifiable quality or benefit. (Wikipedia).

Snake oil is just one visualization at Information is Beautiful (link), the site created by David McCandless, a London-based author, writer and designer who wrote for The Guardian, Wired and others, and nowadays an independent data journalist and information designer. His passion: visualizing information – facts, data, ideas, subjects, issues, statistics, questions – all with the minimum of words (see about).

When you see snake oil you intuitively understand it all.

The image is a “balloon race”. The larger the bubble the higher its popularity in terms of number of Google hits. Orange bubbles look promising but have (yet) a low evidence.

The higher a bubble, the greater the evidence for its effectiveness. But the supplements are only effective for the conditions listed inside the bubble. Evidence is only shown for supplements, taken orally by an adult with a healthy diet.

Some supplements may be represented by multiple bubbles, one for each condition:  after all, the evidence may vary across conditions. For example, there’s strong evidence that Green Tea is good for cholesterol levels. But evidence for its anti-cancer effects is conflicting.

Another nice thing about Snake oil is that it is interactive. You can show (filter) the results for specific conditions or supplement types. Below I selected cardio. Most bubbles disappear. The evidence seems strong for green tea, fish oil and red yeast rice and low for vitamin E and omega-3. When you move your mouse over a bubble it pops up and you can read the supplements name and the condition to which the evidence applies.

Truly amazing.

One might ask how GOOD are the data on which these bubbles are based?

Well I haven’t checked, but the visualization generates itself from this Google Doc. The Google spread sheet shows all the data on which the visualization is based. These can be PubMed Records, Cochrane Systematic Reviews, Medline Plus or a full text paper. The image is automatically regenerated when the google doc is updated with new research that has come out.

The only thing that strikes me as a information specialists is that the way the evidence is retrieved is not stated. Probably this isn’t done in an evidence based way, because each piece of evidence is based on ONE article only. The choice of the paper seems rather random. And some supplements are rather vague. What is meant with “anti-oxidants?” Many of the supplements have anti-oxidant activity for instance.

But the idea in itself is great. Suppose we could gather the evidence in a more evidence based way, share it in Google Docs, appraise it and visualize it. Wouldn’t that be wonderful?