Geographic modeling of very low birth weights around and near Texas federal super

描述（由申请人提供）：出生时体重低于1500克的婴儿中有15%在出院前死亡，2 - 5%在两年内死于并发症。那些存活两年的婴儿往往是残疾的，一生中容易受到多种健康风险的影响。这些婴儿被标记为极低出生体重（VLBW），但也被认为是早产。综合条件影响非洲裔美国婴儿的两倍多，通常是白人婴儿。原因是多因素的，包括遗传易感性，环境暴露和个人行为风险因素。所有这三类因素都被认为是造成种族差异的原因。一个关键的需求存在于解析这些混淆的组件。我们提出了一种方法来模拟这些影响，已成为可能的贝叶斯风险建模的马尔可夫链蒙特卡罗实现的最新发展。该研究将评估VLBW属于更广泛的不良出生结局（ABO）集群的程度，并将确定高风险位置和种族间变异性的任何空间模式。这些结果对于设计和证明即将到来的R 01应用程序中要评估的位置和疾病至关重要。我们将提出一项采用全基因组关联研究（GWAS）的研究，以进一步将个体风险因素解析为存在分层或地理风险的遗传和行为因素。该应用程序的目的是使用现有的数据库来评估极低出生体重是否是可能与婴儿癌症相关的不良出生结局。中心假设是VLBW与德克萨斯州联邦超级基金站点周围的儿童癌症具有相似的风险模式。这一假设将通过两个具体目标进行检验：具体目标1将模拟德克萨斯州47个联邦超级基金地点周围极低出生体重的地理风险。具体目标2将模拟德克萨斯州47个联邦超级基金地点周围极低出生体重和儿童癌症风险之间的空间相关性。这个建议是创新的，因为它将利用两个新的发展，完全条件分层建模和多变量建模。极低出生体重和婴儿癌症之间的相关性的证明将验证未来全基因组关联研究的有力扩展。根据这项研究的结果，未来的GWAS可以将个人风险因素解析为遗传和行为，当与地理风险因素的混杂和相互作用是可能的。 . 公共卫生相关性：这项拟议中的研究意义重大，因为将一种常见的疾病，即极低出生体重，以及相对罕见的儿童癌症纳入一组常见疾病中，将大大加强未来的研究。一个关键的需要存在于解析混杂的组成部分，基于种族的遗传学，基于种族的暴露和种族行为差异。我们提出了一种方法来模拟这些影响，已成为可能的贝叶斯风险建模的马尔可夫链蒙特卡罗实现的最新发展。这些进步包括多变量和层次建模。这些结果对于设计和证明在即将到来的应用中要评估的位置和疾病至关重要。

DESCRIPTION (provided by applicant): Fifteen percent of babies born with a weight less than 1500 g die before being discharged from the hospital and 2 to 5% die within two years from complications. Those infants who survive two years are frequently disabled and prone to a lifetime of multiple health risks. These babies are labeled as very low birth weight (VLBW) but are also considered preterm. The combined conditions affect African-American babies more than twice as commonly as Caucasian babies. The causes are multi-factorial and include genetic susceptibility, environmental exposures and personal behavior risk factors. All three groups of factors are considered as contributory to the racial disparity. A critical need exists in parsing these confounded components. We propose an approach to model these effects that has been made possible by recent developments in Bayesian risk modeling as implemented by Markov Chain Monte Carlo. The study will evaluate the extent for which VLBW belongs in a broader cluster of adverse birth outcomes (ABO) and will identify high-risk locations and any spatial patterns of race-to-race variability. The results will be crucial for designing and justifying the locations and diseases to evaluate in a forthcoming R01 application. We will be proposing a study that employs a genome wide association study (GWAS) to further parse individual risk factors into genetic and behavioral in the presence of hierarchical or geographic risks. The objective of this application is to use an existing database to evaluate very low birth weight as an adverse birth outcome that is potentially correlated to infant cancers. The central hypothesis is VLBW has similar risk patterns to childhood cancer around federal superfund sites in Texas. This hypothesis will be tested by two specific aims: Specific Aim 1 will model the geographic risks for very low birth weights around the 47 federal superfund sites in Texas. Specific Aim 2 will model the spatial correlation among risks for very low birth weight and childhood cancer around the 47 federal superfund sites in Texas. This proposal is innovative because it will exploit two new developments, fully conditional hierarchical modeling and Multivariate modeling. Demonstration of correlation between very low birth weights and infant cancer would validate a powerful extension to a future genome-wide association study. With the results of this study, a future GWAS could parse personal risk factors into genetic and behavioral when confounding and interaction with geographic risk factors is possible. . PUBLIC HEALTH RELEVANCE: The proposed research is significant because the inclusion of a common condition, very low birth weight, with relatively rare childhood cancers in a common cluster of diseases would greatly strengthen future investigations. A critical need exists in parsing the confounded components of race-based genetics, race-based exposures and racial behavioral differences. We propose an approach to model these effects that has been made possible by recent developments in Bayesian risk modeling as implemented by Markov Chain Monte Carlo. These advances include both Multivariate and hierarchical modeling. The results will be crucial for designing and justifying the locations and diseases to evaluate in a forthcoming application.