Identification of Postoperative Pulmonary Complication Risk by Phenotyping Adult Surgical Patients who Underwent General Anesthesia with Mechanical Ventilation

点击翻译按钮获取中文摘要

ABSTRACT Science: One in five patients who develop a postoperative pulmonary complication (PPC) dies within 30 days of surgery. PPCs are the second most frequent surgical complications and lead to increased admission to intensive care units, longer hospital length of stay, and high resource utilization. Ventilator induced lung injury (VILI) secondary to intraoperative mechanical ventilation is a risk for PPCs. Lung protective ventilation, which entails lower tidal volume, sufficient positive end expiratory pressure, optimal inspiratory time and an alveolar recruitment maneuver, has been adopted for intraoperative use to protect pulmonary parenchyma against VILI and ultimately reduce PPC incidence. However, we still do not know the optimal ventilator parameters to yield the lowest incidence of PPCs, because what is best varies from patient to patient. Personalized ventilator parameters are a potential solution to solve this problem. A retrospective study leveraging electronic health records (EHRs) is proposed to identify PPC risks by phenotyping adult surgical patients who underwent general anesthesia with mechanical ventilation. The specific aims of this project are to: (1) Examine the incidence of PPCs in the overall study population and phenotype patients based on nonmodifiable patient, surgical, and anesthesia characteristics; and examine the incidence of PPCs within each phenotypic subgroup; (2) Determine the optimal modifiable intraoperative ventilatory parameters associated with the lowest severity of PPCs within each phenotypic subgroup; and (3) Explore machine learning algorithms for predictive models of the incidence of PPCs on patient, surgical, and anesthesia characteristics as well as intraoperative ventilator parameters. The goal of this aim is to gain knowledge and training in machine learning to lay a foundation for postdoctoral training. Training: My long-term training goal is to become a leading nurse scientist in precision health using data science to improve patient outcomes following surgery, such as reducing PPCs. To achieve this goal, I have three short-term goals during my fellowship training: (1) gain knowledge and skills in research design to enhance precision health in anesthesiology to, (2) gain knowledge in advanced analytic techniques for conducting research using big data, and (3) gain an advanced understanding of pulmonary physiology and pathophysiology that influence anesthesia and patient surgical outcomes. This fellowship will allow me protected time to reach my training goals and to build a foundation for my long-term career goals. During the next twenty-six months as a trainee, I will obtain additional training in (1) research methods and design, (2) advanced statistical methods, (3) precision health, and (4) advanced pulmonary physiology and pathophysiology.