The VHA AMPREDICT Decision Support Tool: Translating Success to Point of Care

慢性肢体威胁性缺血（CLTI）患者的截肢水平决策具有挑战性。的 在这一人群中，截肢水平的选择可以深刻地影响手术死亡率，再住院， 截肢、功能性活动以及最终的生活质量。其中最重要的影响因素 截肢水平的决定是保持活动能力，因为它与生活质量有关。但是，在这方面， 可能无法实现更远端截肢的潜在移动性益处， 愈合受损，需要持续的伤口护理，最终截肢到更高的水平。额外 在这一人群中，决策的关键因素是死亡风险高于大多数癌症 诊断。这种有限的生存率使得外科医生和患者都必须做出决定， 最好确保病人的剩余寿命符合他们的价值观和优先事项。退伍军人健康管理局 数据显示，2005-2014年间，经跖骨（TM）截肢的比例从2005年的2005年增加到2014年的2006年的2007年的2008年的2009年增加到2014年的2009年。 所有CLTI截肢的10%至30%，经胫骨（TT）截肢的比例相应降低， 经股动脉（TF）截肢。TM截肢的增加可能是由于 血运重建，更好的团队合作，以及保留踝关节将使上级 然而，不清楚在决定中如何考虑不愈合的风险。平衡这些风险 是医生和患者之间复杂的共享决策（SDM）过程的核心， 确定截肢的“最佳”水平。成功实施SDM的重要资源 临床决策支持工具（DST）。它们为临床医生提供了关键结局的患者特定风险， 真实的时间。为了让医生拥有这样一个工具，我们利用我们先前的预测模型来创建 AMPREDICT DST.这是一个在线DST，包括主页、预测页面和患者结果页面 特定的一年死亡率和截肢风险以及达到基本独立水平的概率 在三个截肢水平（TM TT，TF）中的每一个上的移动性。它已经经历了VHA医生的成功测试 全国范围DST在临床护理中的成功实施也需要用户群体的购买 需要克服潜在的实施障碍。我们对DST在医生用户中的评价 提供了重要的见解，将在本提案中加以讨论。许多供应商讨论了 将DST集成到EHR中的重要性，并自动填充预测因子，以最大限度地减少实施 负担此外，血管外科医生建议考虑额外的血管诊断和治疗 在临床决策过程中，在预测模型中被视为重要的预测因素。这是 在最近的文献中得到了前沿发展的支持。因此，这项赠款有三个目标。 首先，我们将通过以下方式提高AMPREDICT预测模型在更现代人群中的质量： 用可用的类似预测因子替换EHR中不容易获得的预测因子， 其他血管预测因素。我们将通过一个系统的发展过程来实现这一目标， 人口统计学、社会支持、合并症、健康因素、既往血管手术、实验室检查值， 药物，以及截肢前EHR中可用的心理健康预测因素。二是 执行必要的规划和计划，以将增强的AMPREDICT DST集成到VHA中 EHR通过CPRS和Cerner促进自动填充。这将通过合作来实现 与截肢者护理系统联合Cerner合作，并与一家经验丰富的公司合作， 在VHA EHR中实施风险计算器。第三，我们将评估增强的可用性和效果 电子健康档案中的决策支持系统对医患决策冲突、患者决策后悔和满意度的影响。[半- 外科医生的结构化定性访谈将评估影响实施的临床/机构因素。 总之，这将导致一个重要的临床进步，在一个最改变生活的临床决策。

Amputation level decision making in patients with chronic limb threatening ischemia (CLTI) is challenging. The choice of amputation level in this population can profoundly affect risk of operative mortality, re-hospitalization, reamputation, functional mobility, and ultimately quality of life. One of the most important factors influencing the amputation level decision is the preservation of mobility because of its association with quality of life. However, the potential mobility benefits of more distal amputations may not be realized because of the increased risk of compromised healing, need for ongoing wound care, and ultimately reamputation to a higher level. An additional factor critical in decision making in this population is a mortality risk which is higher than the majority of cancer diagnoses. This limited survival creates an imperative for both surgeon and patient to make decisions that will best ensure the patient’s remaining life years conform to their values and priorities. Veteran Health Administration data suggest that between 2005-2014, the proportion of incident transmetatarsal (TM) amputations tripled from 10% to 30% of all CLTI amputations with a corresponding decrease in the proportion of transtibial (TT) and transfemoral (TF) amputations. The increase in TM amputations may be driven by the improvement in revascularizations, greater teamwork, and the assumption that preserving the ankle joint will enable superior mobility; however, it is unclear how the risks of not healing are considered in the decision. Balancing these risks is at the core of a complex shared decision-making (SDM) process between physicians and patients as they determine the “best” level of amputation. An important resource that enables successful implementation of SDM are clinical decision support tools (DSTs). They provide clinicians with patient specific risks for key outcomes in real time. To equip physicians with such a tool, we leveraged our prior prediction models to create the AMPREDICT DST. It is an online DST that includes a home page, predictor pages, and a result page with patient specific one-year mortality and reamputation risks and probability of achieving a basic level of independent mobility, at each of three amputation levels (TM TT, TF). It has undergone successful testing by VHA physicians nationwide. The successful implementation of the DST in clinical care requires buy in from user groups as well as the need to overcome potential implementation obstacles. Our evaluation of the DST in physician users has provided important insights that will be addressed in the current proposal. Many of the providers discussed the importance of DST integration into the EHR with auto-population of the predictors to minimize the implementation burden. Further, vascular surgeons recommended considering additional vascular diagnostic and therapeutic predictors that are viewed as important in the clinical decisional process, in the prediction models. This is supported by cutting edge developments in the recent literature. The goals of this grant are therefore threefold. First, we will enhance the quality of the AMPREDICT prediction models in a more contemporary population by replacing predictors not readily available in the EHR with similar predictors that are available, and the evaluation of additional vascular predictors. We will accomplish this through a systemic development process considering demographics, social support, comorbidities, health factors, prior vascular procedures, laboratory values, medications, and mental health predictors available in the EHR just prior to the amputation. Second, we will perform the necessary planning and programming to integrate the enhanced AMPREDICT DST into the VHA EHR to facilitate the auto-population through CPRS and Cerner. This will be accomplished through collaborating with the Amputee System of Care joint Cerner workgroup and collaborating with a firm with significant experience implementing risk calculators within the VHA EHR. Third, we will evaluate the usability and effect of the enhanced DST within the EHR on physician and patient decisional conflict, patient decisional regret and satisfaction. [Semi- structured qualitative interviews of surgeons will evaluate clinical/institutional factors that affect implementation.] Together this will result in an important clinical advancement in one of the most life changing clinical decisions.