Abstract
Although prospective studies directly estimate the influence of a factor on the onset of a disease, case-control sampling is often cheaper, faster, and sometimes the only feasible option. In these studies, a random sample of individuals with the disease (cases) is taken, and the proportion of them, P(1), who were exposed to the factor is evaluated. A random sample of individuals without the disease (controls) is also taken, and the proportion of them, P(0), who were exposed is evaluated. From these two proportions, the respective odds and their ratio, the odds ratio (OR), are calculated. When the incidence of the disease is small, say less than 10%, this OR value closely approximates the relative risk (RR), and therefore tells us, with good approximation, how much greater the risk of having the disease is if one is exposed to the factor. In all cases, case-control studies tend to have more confounding factors, which are not easily controllable, than prospective studies. However, under certain circumstances, they have a much higher statistical power than the prospective design.