The purpose of this project is to develop, test, and refine an automated alert system to notify healthcare providers following failure to respond to sepsis in an appropriate and timely fashion. Because the majority of patients diagnosed with sepsis are elderly, this disease disproportionately impacts individuals with special health care needs, including the disabled, chronically ill, and those facing end-of-life. However, any patient can develop sepsis and require critical care hospitalization. Therefore, this project is relevant t the Agency for Healthcare Research and Quality's mission to improve the quality, safety, efficiency, and effectiveness of health care for all Americans. Background: A major challenge in treating critical care patients is delay in initiating appropriate therapy upon sepsis diagnosis. Further, even when prompt therapy is initiated, it may be incomplete, inadequate, and/or hampered by provider error. Failure to respond to sepsis after diagnosis in a timely and appropriate manner can lead to complications, including organ dysfunction, refractory hypotension, multiple organ dysfunction syndrome, and death. Solution: Timely administration of appropriate therapy after diagnosis of sepsis has been demonstrated to significantly improve patient outcomes. With this in mind, our hypothesis is that an automated alert system for detecting failure to rescue after sepsis diagnosis can reduce time to achieve appropriate response in the ICU setting and thereby improve outcomes. The goal of this proposal is to develop and test an automated alert system for the detection of failure to rescue after sepsis diagnosis. To do this, we propose the following three specific aims: (1) improve the diagnostic performance of an existing severe sepsis "sniffer" (2) test different modes of automated alert delivery in experiment, and (3) pilot real-time testing of an improved severe sepsis alert system using the hospital electronic medical record. Our methods will include algorithm optimization, simulated experiment, and prospective testing in the critical care setting. The long-term goal of this proposal is to exploit this technology to improve patient outcomes and reduce overall adverse occurrence rates. This knowledge is significant as it will lay the foundation for the development of increasingly sophisticated automated detection systems to enhance the ability of providers to improve patient outcomes and reduce patient mortality.
As one of the primary causes of critical care hospitalization, mechanical ventilation, and in-hospital mortality, sepsis has a direct impact on significant patien outcomes and healthcare costs. With more than 1,400 global deaths per day from septic shock, even modest improvements in timely and appropriate response to sepsis can substantially impact these public health outcomes. Because the task of adherence to sepsis treatment protocols is complex, an automated detection and alert system has the potential to improve this process.