How do epidemiologists discover the causes of disease?
There are generally five criteria's accepted for the causality of any disease, and these five criteria were articulated in a famous surgeon general's report regarding whether or not smoking is a cause of lung cancer. For example, the five criteria include the strength of association, and let's look specifically at the association between smoking and lung cancer. The data suggested that the more people that smoked or the greater the number of cigarette smoke, then the greater the likelihood of developing lung cancer. This was an example of strength of association. Another criterion was time sequence: whether or not the occurrence of the exposure factor occurs before the development of the outcome, and in the case of smoking and lung cancer, this was demonstrated in the surgeon general's report. Consistency and repetition is a third criterion and this refers to the ability to demonstrate the association in a number of different settings and in a number of different ways. Specificity, perhaps, is one of the weakest criteria. Of course, smoking and lung cancer, means that smoking causes lung cancer and only lung cancer. However, when we examine the evidence, we find that smoking is associated with other forms of cancer and other diseases as well as lung cancer, so the association tends to be less specific than would be ideal. And then, finally, the fifth criterion of association is called coherence of explanation, and really, this criterion means: does the association observed make scientific sense? I believe that the surgeon general's report argued that it did make sense; that if one is exposed to irritating substances found in cigarette smoking, that it is plausible that lung disease and other health effects could occur as a result of cigarette smoking.
