Powell Lisa M, Jones Kelly, Duran Ana Clara, Tarlov Elizabeth, Zenk Shannon N
Health Policy and Administration, School of Public Health, University of Illinois at Chicago, United States.
Health Systems Science, College of Nursing, University of Illinois at Chicago, United States.
Econ Hum Biol. 2019 Aug;34:39-48. doi: 10.1016/j.ehb.2019.05.006. Epub 2019 May 29.
The consumption of ultra-processed foods in the U.S. and globally has increased and is associated with lower diet quality, higher energy intake, higher body weight, and poorer health outcomes. This study drew on individual-level data on measured height and weight from U.S. Department of Veterans Affairs medical records for adults aged 20 to 64 from 2009 through 2014 linked to food and beverage price data from the Council for Community and Economic Research to examine the association between the price of ultra-processed foods and beverages and adult body mass index (BMI). We estimated geographic fixed effects models to control for unobserved heterogeneity of prices. We estimated separate models for men and women and we assessed differences in price sensitivity across subpopulations by socioeconomic status (SES). The results showed that a one-dollar increase in the price of ultra-processed foods and beverages was associated with 0.08 lower BMI units for men (p ≤ 0.05) (price elasticity of BMI of -0.01) and 0.14 lower BMI units for women (p ≤ 0.10) (price elasticity of BMI of -0.02). Higher prices of ultra-processed foods and beverages were associated with lower BMI among low-SES men (price elasticity of BMI of -0.02) and low-SES women (price elasticity of BMI of -0.07) but no statistically significant associations were found for middle- or high-SES men or women. Robustness checks based on the estimation of an individual-level fixed effects model found a consistent but smaller association between the price of ultra-processed foods and beverages and BMI among women (price elasticity of BMI of -0.01) with a relatively larger association for low-SES women (price elasticity of BMI of -0.04) but revealed no association for men highlighting the importance of accounting for individual-level unobserved heterogeneity.
美国和全球范围内超加工食品的消费量都有所增加，且与饮食质量较低、能量摄入较高、体重增加以及健康状况较差有关。本研究利用了美国退伍军人事务部2009年至2014年期间20至64岁成年人的身高和体重测量的个人层面数据，并将其与社区和经济研究委员会的食品和饮料价格数据相联系，以研究超加工食品和饮料价格与成人体重指数（BMI）之间的关联。我们估计了地理固定效应模型，以控制价格中未观察到的异质性。我们分别为男性和女性估计了模型，并按社会经济地位（SES）评估了不同亚人群的价格敏感性差异。结果表明，超加工食品和饮料价格每上涨1美元，男性的BMI单位降低0.08（p≤0.05）（BMI的价格弹性为-0.01），女性的BMI单位降低0.14（p≤0.10）（BMI的价格弹性为-0.02）。超加工食品和饮料价格较高与低SES男性（BMI的价格弹性为-0.02）和低SES女性（BMI的价格弹性为-0.07）的BMI较低有关，但在中等或高SES男性或女性中未发现统计学上的显著关联。基于个体层面固定效应模型估计的稳健性检验发现，超加工食品和饮料价格与女性BMI之间存在一致但较小的关联（BMI的价格弹性为-0.01），低SES女性的关联相对较大（BMI的价格弹性为-0.04），但男性未发现关联，这突出了考虑个体层面未观察到的异质性的重要性。
