Identifying Areas to Improve ICU Outcomes through Provider Variation

PROJECT SUMMARY/ABSTRACT Over 1 million Americans require mechanical ventilation each year and mortality estimates range from 30-35% with significant variability at the patient, hospital, and regional level. More recently, physicians were found to be associated with mortality in mechanically ventilated patients, adding yet another level of variation. There is little research as to what individual physicians do to cause this variation in outcomes and whether provider practice patterns can be modified to improve outcomes. To date, morbidity and mortality reduction interventions in this population involve ventilation strategies and harm reduction therapies. Despite these interventions being evidence-based and guideline-supported, their use remains highly variable. The potential link between provider-level variability in mortality and population-level variation in evidence-based practice represents a critical knowledge gap. Specifically, there is a need to understand how physicians’ adherence to evidence- based practices vary, the degree to which this results in harm, and what barriers are modifying practice patterns. The candidate's prior work developed a method, using electronic health record metadata, to assign a provider retrospectively and reliably to each patient for each intensive care day. Using these data, provider- level adherence to low tidal volume ventilation strategies significantly varied across a 12-hospital health system. This Career Development Award builds from that work and seeks to achieve three aims, that together, will promote the candidate's long-term goal of developing and testing strategies to increase adherence to evidence-based care and improve acute respiratory failure survivors' outcomes. Specifically, in this award, the candidate seeks to (1) quantify variation among physicians in adherence to evidence-based practices, (2) generate hypotheses about barriers and facilitators to evidence-based practice adherence, and (3) develop and pilot an intervention targeting barriers to evidence-based care. To complete these aims the candidate will perform a retrospective cohort study (Aim 1), perform a mixed-method study enrolling 40 ward-based physicians that care for mechanically ventilated patients (Aim 2), and use an intervention mapping framework to develop and pilot an intervention targeting barriers to low tidal volume ventilation. The hands-on experience he will acquire in completing these three studies using different methods will be complemented by carefully selected didactic coursework and structured mentoring by senior investigators from multiple disciplines. This work will lead to pilot grants and R01 to support a randomized clinical trial of this intervention. Completing this research will build upon the candidate’s past training and will provide him with the protected time and experience to achieve his career goal of becoming a leading, independently funded outcomes researcher focused on conducting studies within learning health systems to improve outcomes among survivors of ARF and other critical illnesses.