Rate and rate ratio.

What are they? In human health it is generally the case that the more people studied, and the longer the surveillance of them, the greater the number of new occurrences of any specified condition there are likely to be. This is because of the increase in total time ‘at risk’ of having an event, that is, the increased persons-and-time of observation. A rate is an arithmetic way of quantifying the number of events occurring in relation to both persons and time exposed to risk of the event. In order to avoid very small and hence difficult decimal fractions, the rate will often be expressed in terms of some multiple of the person-time unit. For example, Vessey and Yeates report a rate of first hospitalisations for cervicitis, among women aged 40–44 years, of 33 per 10 000 woman-years.1 This is a simpler way of expressing the underlying but exactly equivalent rate of 0.0033 cases per (single) woman-year of observation. A rate ratio (RR) quantifies a relative comparison between the rates (of occurrence of an event) in two groups, most usually of a group exposed to some risk factor, and a group not exposed.2 The RR is simply obtained by dividing the rate in the one group by the rate in the other. So, for example, the RR for first hospitalisations for cervicitis, in women aged 40–44 years, relative to the youngest age group (i.e. 25–34 years), is 33/11 = 3, which indicates that the older group have an event rate three times that of the younger women.1