Peripheral Artery Disease: Long-term Survival & Outcomes Study (PEARLS)

背景：外周动脉疾病（PAD）是一种常见且高度病态的疾病。近25%的患者 在诊断后3年内死亡，可能是由于心血管（CV）事件的高发生率：心肌梗死 (MI)或者中风因腿部疼痛、行动不便和截肢而残疾的比例要高得多。 仅与PAD相关的医院护理费用就超过210亿美元。然而，关于长期生存的研究， PAD的CV和肢体结局以及现有治疗的影响仍然有限，这在很大程度上是由于 PAD诊断代码的准确性。我们的团队开发了一种使用自然语言处理的新方法 (NLP)在退伍军人健康管理局（VHA）内以高准确度识别PAD患者。 意义：外周动脉疾病：长期生存和结局研究（PEARLS）研究将 以多种方式推进PAD的科学知识。使用我们新颖的NLP工具来识别患有PAD的退伍军人， 我们将研究长期生存和临床结果的轨迹，评估推荐的 治疗（药物、风险因素控制和血运重建）以及上述治疗与 上述成果。我们还将按种族和民族检查PAD护理和结果的差异， 确定这些差距在多大程度上是由于获得高质量护理造成的。我们的工作将 填补PAD研究中的重要空白，并为改善这一高危人群的护理提供提供见解。 创新：使用基于信息学的方法，将一组新诊断的PAD患者聚集在一起， 一个大规模的综合保健系统具有很强的创新性。我们相信，我们的队列识别方法将 变革和促进大数据分析用于研究，改善医疗服务和未来的临床试验。 具体目标：A1。检查PAD长期结局的轨迹，并评估种族和族裔 差距。A2.检查退伍军人健康中PAD的医疗和侵入性管理模式 施用A3.确定医疗和侵入性管理与长期结局的关系 方法学：我们将实施我们的NLP算法，以在VHA中识别新的PAD诊断患者， 2015-2020年，并获得临床和治疗相关变量的数据。我们将纵向跟踪我们的队列， 死亡率、CV事件（MI、卒中）和肢体事件（截肢）。我们将检查PAD治疗的使用情况 和风险因素控制，使用多水平模型识别患者水平和医院水平的治疗预测因子。 我们将使用边缘结构模型来评估PAD治疗与长期结果的相关性。 执行/今后的步骤：主要的可交付成果将包括：a）评估发展和民主政策的长期成果， 确定护理和结果中的种族差异;（B）确定PAD治疗对长期 对决策有用的长期结果，以及c）对治疗中站点水平差异的评估 模式.我们设想，我们的研究结果将有助于我们制定全面的疾病管理计划， 提高护理质量，减少有效治疗使用方面的差距。

Background: Peripheral artery disease (PAD) is a common and highly morbid condition. Nearly 25% of patients die within 3 years of diagnosis, likely due to a high incidence of cardiovascular (CV) events: myocardial infarction (MI) or stroke. A significantly larger proportion experience disability due to leg pain, poor mobility and amputation. The cost of PAD-related hospital care alone exceeds $21 billion. However, research regarding long-term survival, CV, and limb outcomes in PAD and the impact of existing treatments remain limited in large part due to the poor accuracy of PAD diagnosis codes. Our team has developed a novel approach using natural language processing (NLP) to identify PAD patients with a high degree of accuracy within the Veterans Health Administration (VHA). Significance: The Peripheral Artery Disease: Long-term Survival & Outcomes Study (PEARLS) study will advance scientific knowledge for PAD in several ways. Using our novel NLP tool to identify Veterans with PAD, we will examine the trajectory of long-term survival and clinical outcomes, evaluate utilization of recommended treatments (medications, risk factor control and revascularization) and the association of above treatments with the above outcomes. We will also examine disparities in PAD care and outcomes by race and ethnicity and determine the extent to which these disparities our due to access in high quality care. Collectively, our work will address important gaps in PAD research and yield insights for improving care delivery in this high-risk population. Innovation: The use of an informatics-based method to assemble a cohort of newly diagnosed PAD patients in a large integrated health system is highly innovative. We believe that our approach for cohort identification will be transformational and promote big data analytics for research, improving care delivery, and future clinical trials. Specific Aims: A1. Examine the trajectory of long-term outcomes of PAD and assess racial and ethnic disparities. A2. Examine patterns of medical and invasive management of PAD in the Veterans Health Administration A3. Determine the association of medical and invasive management with long-term outcomes Methodology: We will implement our NLP algorithm to identify patients with new PAD diagnosis in VHA during 2015-2020 and obtain data on clinical and treatment related variables. We will follow our cohort longitudinally for mortality, CV events (MI, stroke) and limb events (amputation). We will examine utilization of PAD treatments and risk factor control, identify patient-level and hospital-level predictors of treatment using multi-level models. We will use marginal structural models to evaluate the association of PAD treatments with long-term outcomes. Implementation/Next Steps: Key deliverables will include a) an assessment of long-term outcomes in PAD and identifying racial disparities in care and outcomes; (b) determining the relative impact of PAD treatments on long- term outcomes which can be useful for decision-making and c) an assessment of site-level variation in treatment patterns. We envision that our findings will help us develop comprehensive disease management program to improve quality of care and reduce disparities in use of effective treatments.