A novel shift-level approach to optimizing ICU interprofessional team composition for mechanically ventilated adults

About This Grant

PROJECT SUMMARY At the core of critical care delivery is the interprofessional team, comprised of nurses, physicians, and respiratory therapists, who provide around-the-clock care to the most critically ill patients. The COVID-19 pandemic underscored the need for a robust critical care workforce and exposed existing structural inequities in the healthcare system. Research has shown disparities in mechanical ventilation (MV) care and outcomes based on patient characteristics, such as age and ethnicity. One evidence-based practice associated with improved outcomes for MV patients—spontaneous awakening and spontaneous breathing trials (SATs and SBTs)—requires effective coordination and collaboration among the entire team. For decades, researchers have sought to optimize the organization of critical care teams to enhance patient outcomes. While studies within individual professions, such as nurses or physicians, have examined the impact of clinician characteristics (e.g., experience) and scheduling characteristics (e.g., overtime) on outcomes, there remains a dearth of research investigating how variations in clinician and scheduling characteristics across the entire team influences patient outcomes. To address this knowledge gap, this application aims to determine how clinician and scheduling characteristics impact patient outcomes, and explore the moderating effect that clusters of patient characteristics have on the relationship between clinician and scheduling characteristics on patient outcomes. We propose a novel secondary analysis utilizing a unique dataset that links electronic health record data from nearly 17,000 critically ill patients to human resource clinician characteristics of 8,000 clinicians across 5 intensive care units from 2018-2021. Our specific aims are: 1) Test the hypothesis that patients cared for by teams with better clinician characteristics and scheduling characteristics will be more likely to receive SATs and SBTs each shift (when eligible), more likely to be extubated each shift, and less likely to die each shift. Among (n=8,952) MV patients, we will use a patient-level design and a longitudinal, repeat-measures analysis, where patients serve as their own control. 2) Evaluate the moderating effect of patient characteristics on the relationship between clinician and scheduling characteristics and patient outcomes. Among all ICU patients (N=16,940), we will use multiple interactions to identify what clusters of patient characteristics (e.g., age, sex, race, ethnicity) may be most vulnerable, i.e., experience negative outcomes, to variation in clinician and scheduling characteristics of the assigned team. If supported by our findings and replicated in a larger national study, then a patient-centered approach to ICU team assignments could be a cost-effective intervention that does not require additional staffing. This proposal has the potential to significantly transform the way interprofessional teams are assigned to ICU patients.