Almost 800,000 critically ill patients require mechanical ventilation every year and three quarters of the survivors suffer from persistent disability, which poses a major public health problem as critical care becomes more widely utilized and available. Although early mobilization, which engages patients in physical activity during mechanical ventilation, is a promising evidence-based intervention that may prevent disability, less than ten percent of pa- tients ever get out of bed. This proposal aims to apply precision medicine to identify patients who are most likely to benefit from early mobilization and elucidate how it can be implemented successfully to extend the benefits of early mobilization to critical care survivors at greatest risk for long-term disability. I hypothesize that this re- source-intensive intervention can be applied with greater precision to a subset of patients most likely to bene- fit, and that implementation science strategies can be devised to successfully drive adoption of this interven- tion beyond a clinical trial setting. I will test my hypothesis in three aims:
Aim 1) I will identify the optimal critical illness phenotype for implementation of early mobilization by using cutting-edge machine learning methods;
Aim 2) I will determine the effect of early mobilization on long-term functional disability to incentivize adoption of this practice;
Aim 3) I will determine the barriers and facilitators of implementation of early mobilization across five institutions to identify the contextual features associated with successful implementation to inform strategies that can bridge the gap between evidence base and clinical practice. My long-term goal is to mitigate the com- plications of critical illness with clinical trials using precision-based methods to identify at-risk and yet apt-to- benefit populations paired with implementation science methodologies to illuminate how to bring these interven- tions to the bedside. To accomplish this, I have assembled an exceptional interdisciplinary team of mentors (Drs. Vineet Arora, Matthew Churpek, and John Kress) and advisors (Drs. Shyam Prabhakaran, Donald Hedeker, Laura Damschroder, and Matthias Eikermann) who have a track record of NIH-funding and successful mentor- ship of post-doctoral candidates. I intend to build on my foundation as an accomplished clinical trialist and have formulated an in-depth career development plan to gain expertise in machine learning methods to identify differ- ential treatment effects (Churpek and Prabhakaran), longitudinal data analysis, (Arora and Hedeker), and imple- mentation science methods (Arora, Prabhakaran, and Damschroder) to craft strategies that bring complex mul- tidisciplinary interventions from clinical trials (Kress and Eikermann) to everyday ICU care. Completion of this proposal will train me to fill an unmet need defined by a recent National Academy of Medicine publication which indicated that identification of differential treatment effects must be paired with rigorous implementation to help transition evidence base to routine clinical care. Equipped with advanced statistical skills and implementation science approaches, I will be able to design hybrid effectiveness-implementation trials to target and implement complex multidisciplinary interventions to vulnerable populations in future R01 level applications.
Three-quarters of survivors of mechanical ventilation will develop persistent disability. Early mobilization engages patients in physical activity during mechanical ventilation to prevent disability, but has poor adoption into clinical practice. This proposal aims to apply precision medicine to identify patients who are most likely to benefit from early mobilization and elucidate how it can be implemented successfully to extend the benefits of early mobili- zation to patients at risk for long-term disability.