Putting Implementation into context to advance the management of mechanically ventilated patients.

PROJECT SUMMARY Over one million Americans undergo invasive mechanical ventilation (IMV) annually in intensive care units (ICUs). A few interventions have improved patient-centered outcomes in IMV patients, including sedation minimization, corticosteroid administration in selected patients, and low tidal volume ventilation and prone positioning in the subgroup patients with acute respiratory distress syndrome (ARDS). However, variability exists across ICUs in processes of care, clinical outcomes, and resource utilization. ICUs are highly complex interprofessional environments, and knowledge gaps remain regarding how contextual elements influence implementation determinants in different ICUs. The coronavirus disease 2019 (COVID-19) pandemic created unprecedented surges of patients requiring IMV and catalyzed rapid and dynamic changes in critical care delivery, facilitating some evidence-based practices after many years of persistently low penetration, and impeding others. Furthermore, new evidence, such as corticosteroids for COVID-19 and pneumonia, may have later influenced practice changes for IMV patients more broadly. Thus, this major disruption in critical care delivery is a unique opportunity to gain new knowledge about implementation in the ICU context. The overall objective of this proposal is to better understand the interplay between contextual factors and intervention features in care delivery for IMV patients. Guided by implementation frameworks, we will employ state-of-the-art causal inference methods and innovative qualitative approaches to conduct three specific aims: (1) Quantify penetration of four evidence-based treatments - sedation minimization, corticosteroid administration, low tidal volume ventilation, and prone positioning - over time among IMV patients and specific subgroups; (2) Develop a novel conceptual model of ICU implementation incorporating relationships between determinants and contextual elements; and (3) Apply implementation mapping to create adaptable menus to facilitate evidence-based practices. We will apply state-of-the-art causal inference methods to analyze a diverse, multicenter retrospective cohort of IMV patients admitted to more than 30 ICUs in 16 hospitals before, during, and after the height of the COVID-19 pandemic. In addition, we will use cutting-edge qualitative techniques and implementation research frameworks and will engage an array of stakeholders to develop implementation menus for each treatment that not only provide strategies to promote utilization of evidence- based treatments but also guidance for tailoring those strategies to different contexts. This project will advance the science of care delivery in a high-stakes setting by increasing the evidence regarding how contextual elements interact with treatment characteristics in critical care delivery. It will generate direct preliminary data for future implementation studies to test strategies to promote evidence-based care of IMV patients. Finally, it will create a conceptual model with potential application to delivery of critical care interventions in different ICU contexts more broadly.