THE DETERMINANTS OF INFANT MORTALITY: STATISTICAL MODELS

从历史上看，美国婴儿死亡的总体风险已经下降。然而，在过去的20年里，下降的速度已经放缓，种族和民族差距扩大了。因此，婴儿死亡率仍然是一个严重的公共卫生问题。此外，种族和族裔差异的原因仍然难以捉摸。这项建议的目的是开发必要的方法工具，以便更好地了解影响婴儿死亡率的生物、环境、经济和社会因素以及婴儿死亡率方面的种族和族裔差异。我们认为，现有有效的理论工具，如婴儿和儿童死亡率的最接近决定因素模型，但没有充分运作这些模型的统计工具。理想情况下，这些统计工具应该能够适应直接和间接(通过最接近的决定因素)的影响，允许非线性影响，适当地将最接近的决定因素(其中最重要的是出生体重和胎龄)参数化，并适应出生队列中不可观察到的异质性。我们的初步工作表明，混合模型和Logistic回归的组合可以实现所有这些目标。到目前为止，我们已经开发并测试了出生体重和胎龄(有和没有协变量结构)的混合模型，并对出生体重和胎龄与死亡率的关系进行了建模。我们将在这里扩展这些结果，以测试混合模型的可选参数，并开发将出生体重和胎龄与死亡率相结合的多变量混合模型。我们计划进行能量分析，以确定实施婴儿死亡率近似性决定因素模型所需的最佳样本量。其结果将是一个有据可查的统计方法，充分运用婴儿和儿童死亡率的最接近决定因素模型。这对于检验有关婴儿死亡率减速下降的原因以及婴儿死亡率种族和民族差异增加的原因的理论将是非常宝贵的。最后，我们将使用该模型进行说明性分析，以评估母亲年龄和产次对婴儿死亡率的影响，包括通过出生体重和胎龄的直接决定因素和间接影响。

Historically, the overall risk of infant mortality in the United States has declined. However, the rate of decline has slowed over the last 20 years and racial and ethnic disparities have increased. Consequently, infant mortality remains a serious public health problem. Moreover, the cause of the racial and ethnic differentials has remained elusive. The objective of this proposal is to develop the methodological tools necessary to better understand the factors, biological, environmental, economic and social that influence infant mortality and racial and ethnic disparities in infant mortality. It is our view that effective theoretical tools, such as the proximate determinants model of infant and childhood mortality, are available, but that the statistical tools to fully operationalize these models are not available. Ideally, these statistical tools should be able to accommodate direct and indirect (through the proximate determinants) effects, allow for non-linear effects, appropriately parameterize the proximate determinants (the most important of which are birthweight and gestational age), and accommodate unobservable heterogeneity in the birth cohort. Our preliminary work indicates that a combination of mixture models and logistic regression can achieve all of these ends. To date we have developed and tested mixture models of birthweight and gestational age (with and without covariate structures), and modeled the relationship of both birthweight and gestational age with mortality. We will extend these results here to test alternative parameterizations of the mixture model and develop multivariate mixture models that combine birthweight and gestational age with mortality. We plan to conduct a power analysis to determine the optimum sample sizes necessary for operationalizing the proximate determinants model of infant mortality. The result will be a well-documented statistical method that fully operationalizes the proximate determinants model of infant and childhood mortality. This will be invaluable for testing theories concerning the causes of the decelerating decline of infant mortality, as well as, the increases in racial and ethnic differentials in infant mortality. Finally, we will conduct an illustrative analysis using this model to evaluate the effects of maternal age and parity on infant mortality including the direct and the indirect effects of maternal age and parity through the proximate determinants of birthweight and gestational age.