Abstract：Objective To reconstruct the infant mortality rate from the liberation to the late 1970s in China. Methods According to the time sequence, based on the adjusted infant mortality rate of 1981-1990s and national monitored area infant mortality rate of 1991-2010s, we brought in infant mortality rate of 1949-1980s to establish the relative time series for curve fitting. We preliminary screened the multiple prediction models of infant mortality of 1949-2010s, then secondary screened better dynamic model by using the curve fitting of variance analysis, and at last, we found the best prediction model according to the curve fitting accuracy index, for further retrospective prediction of the infant mortality of 1949-1980. Results Eleven dynamic models of infant mortality rate in China were selected, then we screened out the poor fit models. We found out the exponential form, compound functional, growth functional and logistic curve model, were consistent with the predicted results and the standard error of the estimated value was smallest. They were determined the best prediction model. By prediction the infant mortality rate of 1951, 1956, 1961, 1966, 1971, 1976, and 1980 were 170.5‰,140.7‰, 183.0‰, 95.8‰, 79.1‰, 65.2‰, and 55.9‰ respectively. Conclusion Reconstructed infant mortality rate could not replace the actual level, but more close to the actual level.
李鸿斌. 重新构建解放初期至20世纪70年代末中国婴儿死亡率[J]. 中国医院统计, 2015, 22(5): 349-351.
Li Hongbin. Reconstruction of the infant mortality rate from the liberation to the late 1970s in China. journal1, 2015, 22(5): 349-351.
