Informing national guidelines on diet patterns that promote healthy pregnancy outcomes

摘要 美国育龄妇女的饮食质量比过去50年中的任何时候都要差。糟糕的饮食 质量与不良妊娠结局有关，这会导致婴儿死亡并造成 巨大的社会负担。然而，关于促进健康的饮食模式的正式建议 妊娠结局是缺乏的。美国国会最近要求对怀孕的饮食建议是 包括在下一版美国人饮食指南中-这是一份主要的营养政策文件， 为促进健康提供饮食建议。美国农业部/卫生和公众服务部怀孕工作组，其中包括皮丽莎 Bodnar负责总结支持健康怀孕的饮食模式的现有知识 结果，以告知专门针对妊娠的指南。他们确定了一个证据基础，完全是 他指出，这不足以得出经验性建议，并呼吁进行研究，以填补这一关键的知识缺口。 我们的目标是产生经验证据，为国家饮食模式指南提供信息。 以促进健康的妊娠结局。我们假设我们的结果将建议饮食 对孕妇的建议将与主流的营养建议不同。我们预料到了这一点 分歧是因为我们的创新方法将适应食品之间的复杂协同作用 节食。利用在8个美国学术中心注册的7995名美国女性的大规模预期队列，我们将 量化饮食模式对不良妊娠结局(早产)风险变化的影响 37周、小于胎龄儿、妊娠期糖尿病和先兆子痫)。我们将使用机器 允许饮食成分之间复杂相互作用的学习技术。然后，我们将泛化 我们样本中向美国孕妇群体推荐的饮食模式使用尖端技术 “可移植性”方法是在因果推理文献中发展起来的。最后，我们将开发机器 学习算法，将确定谁将从饮食模式建议中受益最大的子组。 该项目的成功完成将为膳食指南科学咨询委员会提供 关于促进健康妊娠结局的理想饮食模式的经验性数据。我们的 创新的方法将成为其他健康领域的营养流行病学家的模板 适用于他们的数据，导致了对饮食指南的广泛影响。开发实用的数据驱动 优化妊娠结局的饮食建议将有助于降低高昂的经济和社会负担 减少不良妊娠结局的负担，改善母亲及其子女的健康。

SUMMARY The diet quality of U.S. childbearing aged women is worse now than any time in the last 50 years. Poor diet quality has been linked with adverse pregnancy outcomes that contribute to infant mortality and pose a tremendous societal burden. Nevertheless, formal recommendations on the diet patterns that promote healthy pregnancy outcomes are lacking. The US Congress recently mandated that dietary advice for pregnancy be included in the next edition of the Dietary Guidelines for Americans—the major nutrition policy document that provides dietary advice for health promotion. The USDA/HHS Pregnancy Work Group, which included PI Lisa Bodnar, was charged with summarizing existing knowledge on diet patterns that support healthy pregnancy outcomes to inform the pregnancy-specific guidelines. They identified an evidence base that was entirely insufficient for deriving empirical recommendations and called for research to fill this critical knowledge gap. Our objective is to generate empirical evidence that will inform national dietary guidance on the diet patterns that promote healthy pregnancy outcomes. We hypothesize that our results will suggest dietary recommendations for pregnant women that will diverge from prevailing nutrition advice. We expect this divergence because our innovative approaches will accommodate the complex synergy among foods in the diet. Using a large, prospective cohort of 7995 U.S. women enrolled at 8 U.S. academic centers, we will quantify the contribution of dietary patterns to variation in risk of adverse pregnancy outcomes (preterm birth <37 weeks, small-for-gestational-age birth, gestational diabetes, and preeclampsia). We will use machine learning techniques that allow for complex interactions among dietary components. Then, we will generalize recommended dietary patterns in our sample to the U.S. population of pregnant women using cutting edge “transportability” methods developed in the causal inference literature. Finally, we will develop machine learning algorithms that will identify subgroups who will benefit most from dietary pattern recommendations. The successful completion of this project will provide the Dietary Guidelines Scientific Advisory Committee with empirically-derived data on the ideal dietary patterns for promoting healthy pregnancy outcomes. Our innovative methodologies will serve as a template for nutritional epidemiologists in other areas of health to apply to their data, leading to a broad impact on the Dietary Guidelines. Developing practical data-driven dietary recommendations to optimize pregnancy outcomes will help to reduce the high economic and societal burden of adverse pregnancy outcomes and improve the health of mothers and their children.