Hoang Cung Truong, Amin Vikesh, Behrman Jere R, Kohler Hans-Peter, Kohler Illiana V
Department of Economics, University of Pennsylvania, USA.
Central Michigan University, USA.
SSM Popul Health. 2023 Jun 19;23:101448. doi: 10.1016/j.ssmph.2023.101448. eCollection 2023 Sep.
We investigate the roles of genetic predispositions, childhood SES and adult educational attainment in shaping trajectories for three important components of the overall health of older adults -- BMI, depressive symptoms and cognition. We use the Health & Retirement Study (HRS) and group-based trajectory modeling (GBTM) to identify subgroups of people who share the same underlying trajectories ages 51-94 years. After identifying common underlying health trajectories, we use fractional multinomial logit models to estimate associations of (1) polygenic scores for BMI, depression, ever-smoked, education, cognition and subjective wellbeing, (2) childhood SES and (3) educational attainment with the probabilities of trajectory group memberships. While genetic predispositions do play a part in predicting trajectory group memberships, our results highlight the long arm of socioeconomic factors. Educational attainment is the most robust predictor-it predicts increased probabilities of belonging to trajectories with BMI in the normal range, low depressive symptoms and very-high initial cognition. Childhood circumstances are manifested in trajectories to a lesser extent, with childhood SES predicting higher likelihood of being on the low depressive symptoms and very-high initial cognition trajectories. We also find suggestive evidence that associations of educational attainment on the probabilities of being on trajectories with BMI in the normal range, low depressive symptoms and very-high initial cognition vary with genetic predispositions. Our results suggest that policies to increase educational attainment may improve population health by increasing the likelihood of belonging to "good" aging trajectories.
我们研究了遗传易感性、童年社会经济地位（SES）和成人教育程度在塑造老年人总体健康的三个重要组成部分（体重指数、抑郁症状和认知能力）轨迹方面所起的作用。我们使用健康与退休研究（HRS）和基于群体的轨迹建模（GBTM）来识别年龄在51 - 94岁之间具有相同潜在轨迹的人群亚组。在确定了共同的潜在健康轨迹后，我们使用分数多项逻辑回归模型来估计以下因素与轨迹组成员概率之间的关联：（1）体重指数、抑郁、曾经吸烟、教育程度、认知能力和主观幸福感的多基因得分；（2）童年SES；（3）教育程度。虽然遗传易感性在预测轨迹组成员方面确实发挥了作用，但我们的结果突出了社会经济因素的深远影响。教育程度是最有力的预测因素——它预测了属于体重指数在正常范围、抑郁症状低且初始认知能力非常高的轨迹的概率增加。童年环境在轨迹中的体现程度较小，童年SES预测了处于低抑郁症状和非常高初始认知轨迹的更高可能性。我们还发现有暗示性的证据表明，教育程度与体重指数在正常范围、抑郁症状低且初始认知能力非常高的轨迹概率之间的关联因遗传易感性而异。我们的结果表明，提高教育程度的政策可能通过增加属于“良好”衰老轨迹的可能性来改善人群健康。
