Murray C J
Int J Epidemiol. 1988 Mar;17(1):122-8. doi: 10.1093/ije/17.1.122.
The infant mortality rate is not a good indicator of overall mortality or health status. Based on new empirical life tables from the UN Population Division, it can only predict life expectancy with 95% confidence to within a 14-year range. Two infant mortality rates must be nearly 80 units apart to be 95% confident that life expectancy in the two communities is different. Life expectancy itself is not an ideal general measure of mortality, because it implicitly weights deaths at different ages in an inconsistent fashion. A measure of potential years of life lost is preferable because it is ethically more consistent.
婴儿死亡率并非总体死亡率或健康状况的良好指标。根据联合国人口司新的经验生命表,它只能在14年的范围内以95%的置信度预测预期寿命。两个社区的婴儿死亡率必须相差近80个单位,才能有95%的把握确定这两个社区的预期寿命不同。预期寿命本身也不是衡量死亡率的理想通用指标,因为它以一种不一致的方式隐含地对不同年龄段的死亡进行加权。潜在寿命损失年数这一指标更可取,因为它在伦理上更具一致性。