Trends of infant mortality in Hong Kong (1956–90) and evaluation of preventable infant deaths

Objective: This study describes the time trends for infant mortality in Hong Kong and aims to develop statistical models that can be used to predict changes of infant mortality in places already having low levels of infant mortality.
Methodology: Data on births and deaths of infants in Hong Kong during the years 1956–90 were analysed annually as well as by aggregating the data into seven consecutive quinquennia. To assess the contribution of preventable infant deaths, causes for infant deaths were classified into two broad categories: (i) congenital anomalies; and (ii) preventable diseases. A simple linear regression model was used to analyse the time trend of the mortality rate of the preventable diseases (PIMR) over the seven quinquennia.
Results: During the period 1956–90, the infant mortality rate fell from 60.9 in 1956–5.9 per 1000 in 1990 and the neonatal mortality rate fell from 24.2-3.8 per 1000. There was no clear time trend observed for infant mortality of congenital anomalies. However, the time trend for PIMR (log scale) was very close to a straight line and simple linear regression modelling showed a R² of 0.9970.
Conclusion: As the infant mortality rate (IMR) falls to below 30 per 1000, the further rate of decrease becomes less predictable from the regression model of the IMR. By removing the portion of deaths attributable to congenital anomalies, the further decrease in infant mortality became more predictable down to very low levels of IMR.