Statistical Methods in the Comparison of Rates

In decision making, when comparing rates to determine which of them is the highest, itxs important to consider the use of a statistical method: the confidence interval. Objective: To compare oral cancer mortality rates using statistical methods considering confidence intervals. Methods: The cause-specific death rate is defined as the number of deaths assigned to a specific cause devided by the population at risk, and this result is expressed per 10n. It was calculated the oral cancer rates in brazilian regions according to sex, in the periods 1996-1998 and 1999-2001 by point and by 95% confidence interval. Results: The mean oral cancer mortality rates were 4.33 per 100,000 for males and 0.99 per 100,000 females, leading to an intersex ratio of R=4.37 (CI95% = 1.23-4.51), showing a highly significant difference between sexes, over the whole study period. High mortality rates among males were observed in all geographical regions, although this difference varied from one region to another. Conclusion: Although the traditional tests of significance for comparing rates can be used, the confidence interval for the ratio between two rates is simple to apply. In this study the male/female proportion was approximately 4 to 1, and the difference was significant.

Financial Support: CNPq grant # 304140/2002-3