A geostatistical framework for the multi-scale boundary analysis of space-time tr

描述（由申请方提供）：尽管健康状况总体改善，但健康结局变化的时间和幅度（例如癌症晚期诊断百分比、不良出生结局）显示出强烈的种族和地理差异。在不同尺度上量化这些差距的程度以及它们如何随时间变化，是了解其根源和跟踪消除差距进展情况的关键指标。拟议的研究将通过以下方式促进实现这些重要目标：1）开发一种地质统计方法，以确定和绘制对应于个体->邻域->区域的嵌套变化尺度，探索它们与协变量的关系（例如，其他健康结果或推定因素），并量化这些空间模式的时间稳定性，和2）深入的边界和多层次分析的地理和社会经济差异的发病率乳腺癌晚期诊断和不良出生结果（死亡率和低出生体重），以及他们的时间变化，在密歇根州。具体而言，本项目将实现三个目标：1。制定新的方法，将边界分析应用于栅格数据（例如图像或疾病风险图），并在多尺度框架内说明不同空间级别的边界存在情况（例如，县、邮政编码和人口普查区），并在时间框架中实施边界分析，以研究地理边界强度和位置的时间趋势。2.探索创新理论的扩散和基于模拟退火的空间聚合算法的使用，用于健康结果的时间趋势的表示和探索性数据分析，以及它们与空间和时间上的假定因素的关系。3.应用该方法来证明该方法及其对调查几种健康结果时间趋势的地理和种族差异的独特益处（晚期乳腺癌、婴儿死亡率和低出生体重），并探讨它们与潜在因素的关系，如接近筛查设施（例如乳房X光检查诊所）、社会经济地位、空气污染和个人因素（例如吸烟、健康保险、年龄）。这些技术和科学创新将彻底改变我们在多个空间尺度和不同时间观察和解释癌症发病率变化的能力，这将有助于为因果关系或影响生存或发病率的风险因素的深入个体研究提出假设，并建立有针对性的癌症控制干预措施的基本原理，包括考虑卫生服务 筛查和诊断测试的需求和资源分配。它还将有助于长期量化现行战略和政策的惠益，以减少已观察到的癌症诊断阶段和婴儿死亡率方面的地理和种族差异

DESCRIPTION (provided by applicant): Despite the overall improvement in health status, the timing and magnitude of changes in health outcomes (e.g. percentage of cancer late-stage diagnosis, adverse birth outcomes) display strong racial and geographical disparities. Quantifying the magnitude of these disparities at different scales and how they change with time are crucial metrics for understanding their origins and tracking progress towards their elimination. The proposed research will contribute to these important goals through: 1) the development of a geostatistical approach to identify and map nested scales of changes corresponding to individual -> neighborhood -> region, explore their relationships with covariates (e.g. other health outcomes or putative factors) through boundary overlap analysis and quantify the temporal stability of these spatial patterns, and 2) an in-depth boundary and multi-level analysis of the geographic and socioeconomic disparities in the incidence of breast cancer late-stage diagnosis and adverse birth outcomes (mortality and low birth weight), as well as their temporal changes, in Michigan. Specifically, this project will accomplish three aims: 1. Develop new methodologies for applying boundary analysis to raster data (e.g. imagery or disease risk maps) and within a multi-scale framework to account for the existence of boundaries at different spatial levels (e.g., counties, ZIP codes and census tracts), and implementing boundary analysis in a temporal framework to allow the study of temporal trends in the strength and location of geographic boundaries. 2. Explore the use of diffusion of innovation theory and simulated annealing-based spatial aggregation algorithms for the representation and exploratory data analysis of temporal trends in health outcomes and their relationship to putative factors in both space and time. 3. Apply the methodology to demonstrate the approach and its unique benefits for the investigation of geographical and racial disparities in temporal trends of several health outcomes (late-stage breast cancer, infant mortality and low birth weight), and the exploration of their relationships with potential factors, such as proximityto screening facilities (e.g. mammography clinics), socio-economic status, air pollution, and individual-level factors (e.g. smoking, health insurance, age). These technologic and scientific innovations will revolutionize our ability to visualize and interpret variation in cancer incidenceat multiple spatial scales and across time, which will help generating hypotheses for in depth individual studies of risk factors that are causal, or impact survival or morbidity, and establishig the rationale for targeted cancer control interventions, including consideration of health services needs, and resource allocation for screening and diagnostic testing. It will also facilitate the long-term quantification of the benefits of current strategies and policies for reducing the observed geographic and racial disparities in cancer stage at diagnosis and incidence of infant mortality