J Altern Complement Med. 2002 Jun; 8(3): 379-83.

Teaching evidence-based complementary and alternative medicine: 2. A conceptual approach to causation-part 1.

Wilson K, Mills EJ, Hollyer T, Vohra S, Guyatt G, .

Department of Medicine, University of Toronto, Ontario, Canada.

One of the most common questions that arise in clinical practice is whether a causal relationship exists between two factors. In order to answer this question three steps need to be taken: First an association needs to be demonstrated between treatment/exposure and effect. Next, the possibility of this association being the result of error needs to be determined. Finally additional evidence to support a cause-and-effect relationship needs to be identified. Part 1 of this two-part paper describes how a complementary and alternative medicine provider goes through the first two steps when confronted with the question of whether silicone breast implants cause arthritis to develop. Also covered are the need for a control group when attempting to establish whether an association exists, the potential for systematic error (bias) or unsystematic error (chance) to distort an association, and the susceptibility of different study designs to systematic error.