Significance of relative risk measures

Generally speaking RR (and Odds Ratio) values greater than 1 indicates that a disease is more likely to occur in an exposed group as compared to an unexposed group. Conversely, a RR value less than 1 means that a disease event is less likely to occur in an exposed group compared to unexposed group. The confidence that a RR value is a measure of a real increased measurable risk, and not a consequence of chance, is dependent on several factors: 1. the size of the population; 2. the variability in the responding population; and 3. the intensity of the effect.  All of these attributes are accounted for in the calculation of the 95% CI. Thus, to gauge if a RR measure truly does indicate an increase or decrease in risk of disease after exposure, one should examine a RR 95% CI. If the interval overlaps with 1, the RR cannot be considered significant (Box 7).

Box 7.

A longitudinal study was set up to monitor colonies for mortality and other factors as they moved up and down the east coast to pollinate crops. Forty nine colonies were examined in June of 2007, and 20 of them were found to have entombed pollen during the examination. In January 2008, 15 of the colonies that had entombed pollen were dead, as compared to the 6 colonies that died in the cohort without entombed pollen (vanEngelsdorp et al., 2009a).

Box 7 Table

Box 7 Equation

As the RR is greater than 1 and the 95% CI do not overlap with 1, we can say that the increased risk of mortality associated with entombed pollen is significant.  For every colony that died by January that did not have entombed pollen in June, 3.6 colonies died that did have entombed pollen.