Why You Can't Trust Medical Research Results (blog.healthkismet.com)

Sample sizes are too small, and later studies usually negate the originally observed effect. It also means most headline news about health is useless.