IMPROVING THE EVIDENCE FOR UNSTABLE ANGINA GUIDELINES

本项目旨在改善AHCPR不稳定心绞痛实践指南的证据基础，该指南挑战临床医生考虑对有急性心脏缺血（ACI）症状的低风险患者进行门诊管理。利用两项ED决策的临床有效性试验资源，即急性心脏缺血时间不敏感预测仪器（ACI-TIPI）影响试验和即将完成的Sestamibi扫描临床试验，该项目将确定与分诊指南建议的一致性是否与降低短期不良结果（非致死性心肌梗死和死亡）发生率和降低短期利用率(ED复诊、再入院、冠状动脉造影和血运重建术)。通过图表回顾，我们将根据AHCPR分诊指南和其他不稳定心绞痛风险分层模型，根据短期不良结局风险对9191例研究患者进行必要的分类。对于每个指南风险组，我们将在分层逻辑回归模型中检验分诊倾向对不良结果预测的独立贡献，控制患者水平、医生水平和医院水平的不良结果预测因子。由于这些预测因子的不平衡分布可能会对“符合指南”与“不符合指南”的分诊倾向组的不良结果进行偏倚比较，因此我们将研究两种病例组合调整方法：倾向评分和ACI-TIPI评分。由于除了指南之外，还有其他一些现有的工具可以改善急性心脏护理的决策，因此，与其他可用的ACI预后模型相比，我们还将计算指南风险组在区分有和没有不良短期结果的患者方面的敏感性和特异性。拟议的研究将为政策制定者提供预测指南实施对结果、成本和不稳定型心绞痛患者使用医疗服务的影响的能力。

This project aims to improve the evidence base for the AHCPR Unstable Angina Practice Guideline, which challenges clinicians to consider outpatient management for low-risk patients with symptoms suggestive of acute cardiac ischemia (ACI). Capitalizing on the resources of two clinical effectiveness trials of ED decision making, the Acute Cardiac Ischemia Time Insensitive Predictive Instrument (ACI-TIPI) Impact Trial and the nearly completed Sestamibi Scan Clinical Trial, this project will determine whether agreement with guideline recommendations for triage is associated with decreased short-term rates of adverse outcomes (non-fatal myocardial infarction and death) and decreased short-term utilization (ED revisits, readmissions, coronary angiography, and revascularization). Through chart review, we will abstract items necessary to categorize 9191 study patients according to their risk of adverse short term outcomes, based on AHCPR triage guidelines and other unstable angina risk stratification models. For each guideline risk group, we will test the independent contribution of triage disposition to the prediction of adverse outcomes in hierarchical logistic regression models, controlling for the patient-level, physician-level, and hospital-level predictors of adverse outcomes. As unbalanced distribution of these predictors may bias comparisons of adverse outcome in "guideline- concordant" versus "guideline-discordant" triage disposition groups, we will examine two approaches for case-mix adjustment: propensity scores and ACI-TIPI scores. As several other existing instruments, in addition to guidelines, may improve decision-making in acute cardiac care, we will also compute the sensitivity and specificity of the guideline's risk groups in discriminating between patients with and without adverse short-term outcomes, compared to that of other available prognostic models for ACI. The proposed research will provide policymakers with the ability to project the impact of guideline implementation on outcomes, cost, and the use of health care services for patients with symptoms suggestive of unstable angina.