Better Surgical Quality Indicators for the Elderly

描述（由申请人提供）：老年患者在接受住院手术的患者中占越来越大的比例，其发病率和死亡率明显高于年轻患者。鉴于人们普遍认识到手术质量参差不齐，向患者提供有关医院表现的信息的需求日益增长。不幸的是，大多数程序缺乏可靠的质量指标。对于大多数手术，手术量不能可靠地预测单个医院的死亡率。作为目前大多数绩效薪酬计划的基础，过程衡量标准通常与次要结果有关。最后，直接的结果测量（例如，风险调整死亡率）过于“嘈杂”，无法反映大多数医院的表现。基于这些原因，我们提出了一种新的、综合的质量测量方法，该方法可以最佳地利用所有与医院死亡率相关的信息。我们的应用有两个具体目的：1 .制定手术表现的综合措施。利用1998-2001年全国医疗保险人口的数据，我们将为10种常见或高风险住院外科手术中的每一种制定综合绩效指标。模型输入将包括来自不同质量域的度量，包括结构变量（例如，手术量）、护理过程和患者结果（包括感兴趣的手术和相关手术）。模型输出将是医院和特定程序对“真实”死亡率的估计，以最佳方式过滤掉统计噪声。2。评估综合措施对未来表现的预测效果。考虑到患者希望如何使用它们，手术质量指标最好通过历史指标能否预测未来的表现来判断。在此背景下，我们将再次使用医疗保险数据来评估我们的综合措施（1998-2001年数据）相对于单个质量指标（例如，历史死亡率或单独的数量）预测2002-2003年医院死亡率的效果。

DESCRIPTION (provided by applicant): Elderly patients, who comprise a growing majority of patients undergoing inpatient surgery, experience markedly higher rates of morbidity and mortality than their younger counterparts. In light of widespread recognition that surgical quality varies widely, there is growing demand for providing patients with information about hospital performance. Unfortunately, reliable quality indicators are lacking for most procedures. Procedure volume does not reliably predict mortality of individual hospitals for most procedures. Process measures, the basis of most current pay for performance plans, generally relate to secondary outcomes. And, finally, direct outcome measures (e.g., risk-adjusted mortality) are too "noisy" to reflect performance at most hospitals. For these reasons, we propose a novel, integrative measure of quality that makes optimal use of all available information related to hospital mortality with a specific procedure. Our application has 2 specific aims: I. To develop integrative measures of surgical performance. Using data from the national Medicare population (1998-2001), we will develop integrative performance measures for each of 10 common or high-risk inpatient surgical procedures. Model inputs will include measures from different quality domains, including structural variables (e.g., procedure volume), process of care, and patient outcomes (both for the procedure of interest and related operations). Model outputs will be hospital- and procedure-specific estimates of "true" mortality that optimally filter out statistical noise. II. To assess how well integrative measures predict future performance. Given how patients hope to use them, surgical quality indicators are best judged by whether historical measures predict future performance. In this context, we will again use Medicare data to assess how well our integrative measures (from 1998-2001 data) predict hospital mortality rates in 2002-2003, relative to individual quality indicators (e.g., historical mortality or volume alone).