Strengthening implementation science in Acute Respiratory Failure using multilevel analysis of existing data

项目摘要 多达100万美国人经历急性呼吸衰竭（ARF），需要机械通气， 重症监护室每年。研究一再表明，经证实有效的， 有时候，这些患者需要挽救生命的循证实践（EBP），目前还不清楚如何选择 最佳实施战略，可以弥合证据与实践之间的差距。共同 选择方法有其固有的局限性。例如，概念映射和实现映射 严重依赖利益相关者的观点，是劳动密集型的，可能会关注利益相关者的偏好 而不是潜在影响最大的战略。定量方法也具有挑战性，因为 实践的重要决定因素--如个人动机和组织文化--很难 按比例测量。 实施的一个重要目标是减少可归因于下列因素的基本商业惯例吸收的差异： 临床医生和环境设置。虽然临床实践应根据患者因素而变化， 实施方案试图克服临床医生和环境因素（例如， 知识或资源），限制EBP吸收。适用综合执行框架 研究（CFIR）对这个概念模型，个人和内部设置的领域应该有最小的 重症监护实施项目成功后对EBP依从性的影响。我们假设 当患者的个体和内部环境的CFIR域的变异性较低时， 与现有证据支持治疗的患者相比， 治疗效果较弱。我们的总体目标是展示1）如何建立多层次建模技术 可用于估计可归因于CFIR结构域的EBP使用中的变异比例 的内部设置和个人的特点，以及2）如何产生的信息可以告知选择 实施战略并评估其有效性。作为概念的证明，我们将研究两个证明有效的 干预措施-低潮气量通气治疗急性呼吸窘迫综合征和袋式面罩 在插管期间进行通气。我们将使用来自低潮气量通用数据库的现有多中心数据集。 支持：干预性试验（LOTUS-FRUIT）队列研究招募的可行性和3例随机化 收集使用袋式面罩通气数据的试验。

PROJECT SUMMARY Up to 1 million Americans experience acute respiratory failure (ARF) and require mechanical ventilation in an intensive care unit annually. Studies repeatedly revealed incomplete penetration of proven-effective, sometimes life-saving, evidence-based practices (EBP) for these patients, and it is unclear how to select optimal implementation strategies that can bridge the gap between evidence and practice. Common approaches to selection have inherent limitations. For example, concept mapping and implementation mapping rely heavily on stakeholder perspectives, are labor-intensive, and may focus on stakeholder preferences instead of strategies with the greatest potential impact. Quantitative approaches are also challenging because important determinants of practice - such as individual motivation and organizational culture - are difficult to measure at scale. One important goal of implementation is to reduce variability in the uptake of EBPs attributable to clinicians and the environmental setting. While clinical practice should vary in response to patient factors and preferences, implementation programs try to overcome clinician and environmental factors (e.g. insufficient knowledge or resources) that limit EBP uptake. Applying the Consolidated Framework for Implementation Research (CFIR) to this conceptual model, the domains of Individuals and Inner Setting should have minimal influence on adherence to EBPs after a successful critical care implementation program. We hypothesize that variability attributable to the CFIR domains of Individuals and Inner Setting is lower among patients when a treatment is supported by high-quality evidence compared to patients for whom the existing evidence for a treatment is weaker. Our overall objective is to demonstrate 1) how established multilevel modeling techniques can be used to estimate the proportion of variation in the use of EBPs that is attributable to the CFIR domains of Inner Setting and Characteristics of Individuals, and 2) how the resulting information can inform selection of implementation strategies and evaluate their effectiveness. As a proof of concept, we will study two proveneffective interventions - low tidal volume ventilation for acute respiratory distress syndrome and bag mask ventilation during intubation. We will use existing multicenter datasets from the Low Tidal Volume Universal Support: Feasibility of Recruitment for lnterventional Trial (LOTUS-FRUIT) cohort study and from 3 randomized trials that collected data on the use of bag-mask ventilation.