• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>International Conference on Information Technology Interfaces >Aspects of hierarchical regression modeling in health services and outcomes research
【24h】

Aspects of hierarchical regression modeling in health services and outcomes research

机译：健康服务和结果研究中分层回归建模的方面

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

The explosive growth of Bayesian methods and computational techniques during the last decade has brought hierarchical regression modeling into the mainstream of applied statistics (Gilks et al, 1996; BUGS manual, 2001). Hierarchical regression models (HRMs) provide a natural framework for the analysis of multi-level, clustered data (Lindley and Smith, 1972; Wong and Mason 1991). Such data are encountered routinely in health care research. For example, patients can be clustered by health care provider, such as physician or hospital, or by study, in the case of meta-analysis; health care providers can be clustered by health care systems, market areas, and geographic areas. The databases typically include information on covariates, measured at various levels of aggregation. For example, patient-level covariates may include socio-demographics and information on patient severity and comorbidity; provider-level covariates may include information on provider organization, resources, and processes. The analysis of these data addresses a broad range of scientific and policy questions. In particular, the analysis of variations in health care process and outcomes across units, such as hospitals or geographic areas, examines the question of whether comparable patients receive similar treatments across these units; whether they experience similar outcomes; and whether differences in unit characteristics and practice patterns are related to differences in observed outcomes. The comparison of the performance of health care providers (usually referred to as profiling analysis), derives and compares provider-specific measures of health care utilization and outcomes, adjusted for patient case-mix. A usual goal in profiling analyses is to identify of aberrant providers and to establish processes for controlling quality of care and generating appropriate information for individual and policy decision making. (Gatsonis et al, 1995; Goldstein and Spiegelhalter 1996; Normand et al, 1997; Daniels and Gatsonis, 1997, 1999).
机译：贝叶斯方法和计算技术在过去十年中的爆炸性增长将分层回归建模带入了应用统计的主流（Gilks​​等，1996; Bugs Manual，2001）。分层回归模型（HRMS）为分析多层次，集群数据（Lindley和Smith，1972; Wong和Mason 1991）提供自然框架。这些数据通常在医疗保健研究中遇到。例如，在Meta分析的情况下，患者可以由医疗保健提供者（如医师或医院）或通过研究聚集;医疗保健提供者可以由医疗保健系统，市场区域和地理区域聚集。数据库通常包括在各种聚集水平下测量的协变量的信息。例如，患者级协变量可能包括社会人口统计数据和有关患者严重程度和合并症的信息;提供商级协变量可能包括有关提供商组织，资源和流程的信息。对这些数据的分析解决了广泛的科学和政策问题。特别是，跨部门或地理区域的单位的医疗保健过程和结果的变化分析，研究了可比较患者是否接受这些单位的类似治疗的问题;他们是否经历过类似的结果;以及单位特征和实践模式的差异是否与观察结果的差异有关。对医疗保健提供者的表现（通常称为分析分析）的比较，并衍生出特定的医疗保健利用和结果的措施，调整患者案例混合。分析分析中的通常目标是识别异常提供者，并建立控制护理质量的流程，并为个人和政策决策产生适当的信息。 （Gatsonis等，1995; Goldstein和Spiegelhalter 1996; Normand等，1997; Daniels和Gatsonis，1997,1999）。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号