Supplementary Material for: Divergence and Convergence: How Do Income Inequalities in Mortality Change over the Life Course?

<b><i>Background:</i></b> Do inequalities in health by income increase or decrease with age? The empirical evidence is not conclusive and competing theories arrive at different conclusions. <b><i>Objective:</i></b> This study examined inequality in mortality by income over the adult life course with longitudinal data on people aged 30–99 between the years 1990 and 2009. Each person was followed for 19 years. <b><i>Methods:</i></b> We used Swedish total population data with 5,011,414 individual observations. We calculated the probability of having died for ages between 31 and 99. This approach to calculating death risk incorporates selective mortality during the follow-up period into the measure. Age and year standardized income positions were calculated for all individuals. Inequality was assessed by comparing the top 10% income group and the bottom 10% income group. Relative inequality was measured by risk ratios (RR) and absolute inequality by percentage point differences. <b><i>Results:</i></b> The results showed that the highest relative income inequality in mortality was at age 56 for men (RR: 4.7) and at age 40 for women (RR: 4.1) with differing patterns across the younger age categories between the sexes. The highest absolute income inequality in mortality was found at age 78 for men (19% difference) and at age 89 for women (14% difference) with similar patterns for both sexes. Both measures of inequality decreased after the peak, with small or no inequalities above age 95. Income inequality in mortality remained in advanced age, with larger absolute inequalities in older ages and larger relative inequalities in younger ages. <b><i>Conclusion:</i></b> The results for absolute and relative measures of inequality differed substantially; this highlights the importance of discussing and making an active choice of inequality measure. To explain and understand the patterns of inequality in mortality over the adult life course, we conclude that the “age-as-leveler” and “cumulative disadvantage” theories are best applied to an absolute measure of inequality.

**背景**：按收入划分的健康不平等随年龄增长是加剧还是缓解？当前实证证据尚未形成定论，不同竞争性理论得出了相异结论。
**研究目标**：本研究依托1990年至2009年间30至99岁人群的纵向数据（longitudinal data），探究成人生命周期内按收入划分的死亡率不平等状况，对每名研究对象开展了长达19年的随访。
**研究方法**：本研究使用瑞典全国人口数据集，共包含5011414条个体观测数据。我们计算了31至99岁各年龄层的死亡概率，该死亡风险测算方法将随访期内的选择性死亡（selective mortality）纳入衡量维度。我们为所有研究对象计算了经年龄与年份标准化后的收入位次。通过对比收入最高10%组与最低10%组的差异评估不平等程度：相对不平等采用风险比（risk ratios，RR）量化，绝对不平等则以百分点差值作为衡量指标。
**研究结果**：结果显示，男性死亡率的收入相对不平等峰值出现在56岁（RR=4.7），女性为40岁（RR=4.1），且两性在较年轻年龄层的不平等模式存在显著差异。死亡率的收入绝对不平等峰值分别为男性78岁（差值19%）、女性89岁（差值14%），两性在该维度的变化模式较为相似。两类不平等指标均在达到峰值后逐步下降，95岁以上人群的不平等程度微弱甚至不存在。高龄群体仍存在死亡率收入不平等：老年群体绝对不平等程度更高，年轻群体相对不平等更为显著。
**研究结论**：本次研究中，绝对与相对两类不平等衡量指标的结果存在显著差异，这凸显了讨论并主动选择合适不平等衡量方法的重要性。为解释并阐明成人生命周期内死亡率收入不平等的变化模式，本研究认为“年龄平抑效应（age-as-leveler）”与“累积劣势（cumulative disadvantage）”理论更适用于绝对不平等的衡量维度。