Methodology in medical research; the need for controlled clinical studies.

Too many medical researchers vitiate their work by ignoring the problem of uncontrolled variables. They therefore publish clinical impressions "dressed up" in scientifically meaningless numbers. A prototypical example of this practice is contrasted with a controlled study, each employing the same (small) number of patients. It is shown how the use of controls can convert a meaningless experiment into one that has assessable scientific significance.A survey of current literature revealed that in only 21 of 100 articles studied were adequately controlled experimental conditions employed. Since they usually deal with very complex systems, it is urged that medical researchers exercise more scientific rigor with regard to control problems.