Reporting Timelier Infant Mortality Rates

The tracking and reporting of infant mortality rates (IMR) often lags eight months or more after the infant deaths occur. This delay, due to the process of collecting birth and death certificate data used to calculate IMR, can prevent timely identification and response to changes in infant mortality. In Marion County, Indiana, on average, 99% of all birth certificates and death certificates will have been received within 1.5 and 7.5 months of the birth or death, respectively. As is common practice, we do not report the IMR until vital records completeness is at least 99%, meaning that the IMR for the prior year is not released until mid-August. In this presentation, we propose a method to generate more timely IMRs. Using six years of data, we calculated a year-to-date IMR (ytdIMR) at the end of each quarter, based on the data that was available one month after that date. To assess how well the ytdIMR estimated the year-end IMR (yeIMR), we calculated the difference between each ytdIMR and the corresponding yeIMRs based on complete data. We repeated these calculations for 364 “pseudo-years”, progressively treating each day of the year as though it were March 31, June 30, September 30, or December 31. The result was over 1800 ytdIMR-yeIMR differences per prediction date, providing a basis for an empirical confidence interval for how well quarterly ytdIMRs calculated one month after a date predict the yeIMR. This approach may allow program administrators to be timelier in addressing increases in the IMR.