Department of Public Health, Aarhus University, Denmark.
Scand J Public Health. 2018 Nov;46(7):767-773. doi: 10.1177/1403494817749050. Epub 2018 Jan 31.
The standardized mortality ratio (SMR) is a widely used measure. A recent methodological study provided an accurate approximate relationship between an SMR and difference in lifetime expectancies. This study examines the usefulness of the theoretical relationship, when comparing historic mortality data in four Scandinavian populations.
For Denmark, Finland, Norway and Sweden, data on mortality every fifth year in the period 1950 to 2010 were obtained. Using 1980 as the reference year, SMRs and difference in life expectancy were calculated. The assumptions behind the theoretical relationship were examined graphically. The theoretical relationship predicts a linear association with a slope, [Formula: see text], between log(SMR) and difference in life expectancies, and the theoretical prediction and calculated differences in lifetime expectancies were compared. We examined the linear association both for life expectancy at birth and at age 30. All analyses were done for females, males and the total population.
The approximate relationship provided accurate predictions of actual differences in lifetime expectancies. The accuracy of the predictions was better when age was restricted to above 30, and improved if the changes in mortality rate were close to a proportional change. Slopes of the linear relationship were generally around 9 for females and 10 for males.
The theoretically derived relationship between SMR and difference in life expectancies provides an accurate prediction for comparing populations with approximately proportional differences in mortality, and was relatively robust. The relationship may provide a useful prediction of differences in lifetime expectancies, which can be more readily communicated and understood.
标准化死亡率（SMR）是一种广泛使用的指标。最近的一项方法学研究提供了 SMR 与预期寿命差异之间准确的近似关系。本研究检验了在比较四个斯堪的纳维亚人群的历史死亡率数据时，理论关系的有用性。
对于丹麦、芬兰、挪威和瑞典，获取了 1950 年至 2010 年每五年一次的死亡率数据。以 1980 年为参考年，计算了 SMR 和预期寿命差异。通过图形检查了理论关系背后的假设。理论关系预测了对数（SMR）和预期寿命差异之间的线性关系，斜率为[公式：见正文]，并比较了理论预测和计算得出的预期寿命差异。我们分别检验了出生时和 30 岁时的预期寿命的线性关系。所有分析均针对女性、男性和总人口进行。
近似关系准确预测了实际预期寿命差异。当年龄限制在 30 岁以上时，预测的准确性更高，如果死亡率的变化接近成比例的变化，预测的准确性会提高。线性关系的斜率通常在女性中约为 9，在男性中约为 10。
SMR 和预期寿命差异之间的理论关系为比较死亡率大致成比例差异的人群提供了准确的预测，并且相对稳健。该关系可以提供对预期寿命差异的有用预测，这更容易传达和理解。
