Trends in birth weight-specific and -adjusted infant mortality rates in Taiwan between 2004 and 2011

Background: A yearly increase in the proportion of very low birth weight (VLBW) live births has resulted in the slowdown of decreasing trends in crude infant mortality rates (IMRs). In this study, we examined the trends in birth weight-specific as well as birth weight-adjusted IMRs in Taiwan. Methods: We linked three nationwide datasets, namely the National Birth Reporting Database, National Birth Certification Registry, and National Death Certification Registry databases, to calculate the IMRs according to the birth weight category. Trend tests and mortality rate ratios in the periods 2010–2011 and 2004–2005 were used to examine the extent of reduction in birth weight-specific and birth weight-adjusted IMRs. Results: The proportion of VLBW (<1500 g) infants among live births increased from 0.78% in 2004–2005 to 0.89% in 2010–2011, thus exhibiting a 15% increase. The extents of the decreases in birth weight-specific IMRs in the 500–999, 1000–1499, 1500–1999, 2000–2499, and 2500–2999 g birth weight categories were 15%, 33%, 43%, 30%, and 28%, respectively, from 2004–2005 to 2010–2011. The reduction in IMR in each birth weight category was larger than the reduction in the crude IMR (13%). By contrast, the IMR in the <500 g birth weight category exhibited a 56% increase during the study period. The IMRs were calculated by excluding all live births with a birth weight of <500 g. The birth weight-adjusted IMRs, which were calculated using a standard birth weight distribution structure for adjustment, exhibited similar extent reductions. Conclusion: In countries with an increasing proportion of VLBW live births, birth weight-specific or -adjusted IMRs are more appropriate than other indices for accurately assessing the real extent of reduction in IMRs.