Epidemiology/Introduction

It can be seen that the two odds ratios worked out in these two ways above are identical.

This second way of working it out is very close to the 3 other measures of effect where one tries to calculate the effect by dividing the occurrence of disease in the exposed by the occurrence of disease in the unexposed. Except that there is no meaningful measure of occurrence in a case control study because one selects the number of cases and controls oneself and there is no denominator to which to relate the cases in the numerator. So this way of looking at the OR makes no sense.

On the other hand, the first way does make sense, as it is possible to compare the exposure odds of those diseased with the non-diseased.

Fortunately, as they turn out to be the same thing as a ratio one can interpret the OR in the same way as the other 3 measures of effect which are based on directly measurable occurrence measures in the population of interest.

Example:

In a case control study of wood dust exposure and nasal cancer, the following data were obtained:

	D+	D-
E+	53	53	106
E-	43	85	127
	96	138	233

Since the proportion of cases exposed (53/96 = 55.2%) is higher than the proportion of controls exposed (53/138 = 38.4%), there appears to be an association between exposure and disease.

The OR measures this and = 53 x 85/43 x 53 = 1.98.

This means that the measure of effect is estimated (by the use of the OR) to be about two times higher among the exposed than among the unexposed.