Managing risk in statistics "Relative Risk"

Abstract

To determine how a disease is being transmitted, identity high-risk groups in the community requiring special prevention measures and to elucidate the most important factors contributing to disease causation, it is essential to move beyond counts of disease occurrence. As discussed in a previous article in this series, it is necessary to transform these tallies into incidence rates (or risk), that is, divide the number of new occurrence of the disease by the population at risk in that geographical location during the specific time-period of interest. However, to determine a particular risk-factor's (or "exposure's" in statistical jargon) importance, once needs to go one step further. That crucial step is to compare the risk of disease occurrence in those people "not exposed" to that factor. This can be done either by dividing the risk in those exposed with the risk in those unexposed to that factor i.e relative risk, or subtracting the risk in those unexposed from the risk in those exposed i.e attributable risk. We will focus on relative risk in this article and attributable risk in a subsequent article.