Automated Delineation of Cancer Service Areas

摘要 20多年来，美国的医疗保健服务一直是通过创建“服务领域”的方法来提供信息的，例如 达特茅斯地图集项目的医院服务区(HSA)，以评估卫生保健资源是如何分配的 以及这对健康结果有何影响。政策制定者已经使用这些单位来评估区域 在卫生保健利用和质量方面的差异，以设计改进卫生保健系统的战略。送货 在美国，癌症护理代表着一套独特的患者、技术、临床专业化和患者- 以视角为中心，区别于其他患者群体。医学研究所和美国癌症学会 临床肿瘤学最近注意到在癌症护理提供方面存在一种“危机”，并强调有必要 评估质量的有意义的方法。我们建议开发一种新的方法来生成癌症服务区(CSA)- 地理空间单位，类似于HSA，但专门针对癌症护理--以便创建一个框架，以评估区域 癌症护理服务的提供、质量和结果。基于通过所有付款人索赔获取的医疗保健利用率以及 医疗保险声称，我们将扩展和完善达特茅斯人血清白蛋白模型。派生的CSA有几个关键区别 从现有的服务区域划分：a)侧重于癌症特定的患者群体/诊断；b)纳入门诊 除住院患者外，还声称要获取连续护理；c)完善复杂的基于网络的社区 在获得CSA的地理邻接性的同时考虑患者护理的空间模式的检测方法；以及d) 在地理信息系统(GIS)环境中创建适应用户定义的自动化程序 服务、诊断或临床表型的集合。我们的具体目标是：1)发展癌症服务区(CSA)- 独特的癌症特定地理医疗保健利用单位，通过改进的方法评估癌症护理 方法；2)评估CSA与达特茅斯HSA，以评估其对感兴趣人群的空间特异性； 以及3)展示CSA作为相对于癌症人群的独特空间单位的效用。CSA的创建是 迫切需要政策制定者(如国会)、决策领导人(如ASCO)、医疗保健系统，并最终 寻求关于高质量癌症护理的可靠、可报告的信息的患者。这是迈向这些目标的第一步，而且 承诺同时服务于广泛的服务领域方法论。

ABSTRACT For over 20 years, health care delivery in the U.S. has been informed by methodologies that create “service areas”, such as Hospital Service Areas (HSAs) of the Dartmouth Atlas Project, to evaluate how health care resources are distributed across the population and how that impacts health outcomes. Policy makers have used these units to assess regional variation in health care utilization and quality to design strategies for improving health and health care systems. Delivery of cancer care in the United States represents a unique set of patients, technologies, clinical specialization, and patient- centered perspectives, distinct from other patient populations. The Institute of Medicine and the American Society of Clinical Oncology have recently noted that there is a “crisis” in cancer care delivery, and highlighted the need for meaningful ways to assess quality. We propose to develop a novel method to generate Cancer Service Areas (CSAs) – geospatial units analogous to HSAs, but specific to cancer care – in order to create a framework for assessing regional cancer care delivery, quality, and outcomes. Based on health care utilization captured through all-payer claims and Medicare claims, we will extend and refine the Dartmouth HSA model. The derived CSAs have several key distinctions from existing service area delineations: a) focus on cancer-specific patient population/diagnoses; b) inclusion of outpatient claims, in addition to inpatient, to capture continuum of care; c) refinement of a complex network-based community detection method to account for spatial patterns of patient care while attaining geographic contiguity of the CSAs; and d) creation of an automated program in a Geographic Information Systems (GIS) environment that adapts to user-defined sets of services, diagnoses, or clinical phenotypes. Our specific aims are to: 1) Develop Cancer Service Areas (CSAs)- unique, cancer-specific geographic units of healthcare utilization to evaluate cancer care through a refined methodologic approach; 2) Evaluate the CSAs versus Dartmouth HSAs to assess their spatial specificity to the population of interest; and 3) Demonstrate the utility of CSAs as unique spatial units with respect to the cancer population. Creation of CSAs is an urgent need for policymakers (e.g. Congress), decision leaders (e.g. ASCO), health care systems, and ultimately patients who seek reliable, reportable information on quality cancer care. It is a first step towards these goals, and promises to serve broad service area methodologies at the same time.