Critical illness represents an extraordinarily high burden on the health system. Between 4 and 7 million Americans are admitted to an intensive care unit (ICU) each year, and the incidence of critical illness syndromes like acute respiratory failure, acute lung injury and sepsis is expected to rise dramatically with the aging of the US population. As a result, the ICU is an increasingly important area for quality improvement initiatives and comparative effectiveness research designed to improve patient outcomes. However, these efforts are limited by the lack of a robust measure of hospital performance for critically ill patients. Existing risk-adjusted mortality measures are limited in several key ways. First, they only focus on patients admitted to the ICU, neglecting severely ill hospitalized patients who are not admitted, ignoring the role of ICU admission decisions in patient outcomes, and failing to reward hospitals for high-quality care for sick patients on the hospital ward. Second, they exclude patients transferred in from other hospitals and do not account for variation in discharge practices across hospitals, neglecting the role of care transitions in outcomes following critical illness and failing to reward hospitals for high-quality care of comple transfer patients. Third, due to the inherent unreliability of outcome measures they lack sufficien precision to be useful in clinical or health policy decisions. As a result, current risk-adjusted mortality measures may fail to accurately identify high-performing hospitals, hindering both comparative effectiveness research and efforts to translate clinical evidence into practice. This project will address each of these problems through the development and validation of novel methods for measuring hospital-specific risk-adjusted mortality rates for critically ill patients. e base our approach on a conceptual model of critical care quality that emphasizes the entire episode of critical illness, not just the episode of care within an ICU. Using clinical and administrative data from Pennsylvania hospitals, we will apply state-of-the-art Bayesian techniques and an innovative marginal structural modeling approach to create and test hospital- specific critical care mortality rates that account for variation in ICU admission practices and inter- hospital transfers. Then we will develop a new composite measure of critical care mortality rates that combines information about the structure and outcome of care to increase the precision of outcome assessment. Our results will provide clinicians and policy makers with novel measurement tools for assessing critical care performance on a national scale, as well as provide researchers with new measures to test the effectiveness of system-wide clinical and policy innovations designed to improve outcomes for patients with acute respiratory failure and other forms of critical illness.
Over 4 million Americans are admitted to an intensive care unit each year, and critical illness syndromes such as acute respiratory failure and acute lung injury claim hundreds of thousands of lives annually. By determining the optimal way to measure critical illness outcomes across hospitals, this project will provide clinicians, researchers and policy makers with important tools to improve the organization and management of critical care services in the United States.
|Barbash, Ian J; Zhang, Hongwei; Angus, Derek C et al. (2017) Differences in Hospital Risk-standardized Mortality Rates for Acute Myocardial Infarction When Assessed Using Transferred and Nontransferred Patients. Med Care 55:476-482|
|Barbash, Ian J; Le, Tri Q; Pike, Francis et al. (2016) The Effect of Intensive Care Unit Admission Patterns on Mortality-based Critical Care Performance Measures. Ann Am Thorac Soc 13:877-86|