ABSTRACT
In epidemiology, we are often interested in evaluating the chance that an individual who possesses a certain attribute also has a specific disease. The most basic epidemiological measure is the probability of an individual becoming newly diseased given that the individual has the particular attribute under consideration. This is called the
risk
of disease; as defined in Section 1.1, the attribute considered is called the
risk factor
. Hence, risk measures the probability of disease incidence. Although risk is a useful summary of the relationship between risk factor and disease,
it is not sufficient by itself for assessing the importance of the risk factor to disease outcome. For instance, we may find that 30% of a sample of women who use a particular type of contraceptive pill develop breast cancer, so the risk of breast cancer is 0.3 for pill users. This would seem impressive evidence implicating the pill, unless it transpired that a similar percentage of nonpill users have also developed breast cancer. As in most procedures in epidemiology, a comparison group is required; the simplest one to take here is the group without the risk factor. This leads to the definition of the
relative risk
(or
risk ratio
) as the ratio of the risk of disease for those with the risk factor to the risk of disease for those without the risk factor. If the relative risk is above 1, the factor under investigation increases risk; if less than 1, it reduces risk. A factor with a relative risk less than 1 is sometimes referred to as a
protective factor
. In most cases, we shall use the general term ‘risk factor’ without specifying the direction of its effect. Sometimes (somewhat unnecessarily) risk is called
absolute
risk
, to distinguish it from relative risk. Computation of the risk and relative risk is particularly simple from a 2
×
2 table (two rows by two columns) of risk factor status against disease status, designated algebraically by Table 3.1. This represents data from
n
subjects free from disease at the outset of the study (the
baseline
). Each individual’s risk factor status at baseline was recorded, as was whether she or he went on to develop the disease during the study.
