|Table 1: Trends in age-adjusted HIV/AIDS mortality rate per 100,000 population according to socioeconomic deprivation quintiles, United States, 1987–2009.|
|Source: Data derived from the US National Vital Statistics System, 1987–2009. |
Death rates are age-adjusted by the direct method to the 2000 US standard population. The standard error (SE) of the age-adjusted mortality rate can be approximated by the rate divided by the number of deaths. For example, SE for the mortality rate in the most-deprived group in 2005–2009 = 6.28/SQRT (16,066) = 0.05. The standard normal Z test can be used to evaluate the statistical significance of the difference in mortality rates between any two groups at one point in time or to test for change in mortality rates between two time points for a specific group. If the absolute value of the Z-statistic = (R1 − R2)/SQRT [(SE(R1))2 + (SE(R2))2] ≥ 1.96, then the difference is statistically significant at the 0.05 level. If the test statistic ≥ 2.58, then the difference in the rates is significant at the 0.01 level, where R1 is mortality rate for the first demographic group, R2 is mortality rate for the second demographic group, and SE(R1) and SE(R2) are the standard errors associated with R1 and R2, respectively. For example, in 1995 the HIV/AIDS mortality rate in the most-deprived socioeconomic group (22.27) was 2.8 times higher than the corresponding rate in 1998 (8.04), and this difference was statistically significant at the 0.01 level (computed Z-value = 57.36).
The 95% confidence interval for a specific mortality rate can be constructed by using the formula = rate ± 1.96 * SE.