Background: Sepsis, the body’s overwhelming systemic response to infection, strikes more than 1 million patients annually in the United States and is known to impact over 48,000 Veterans every year. Over the past decade, sepsis survival has continued to improve through a better understanding of effective therapies, early intervention, and prophylaxis. This has been seen in the private sector and VA with inpatient mortality dropping from 15% in 2008 to 10% in 2012. However, despite this improvement, a patient with sepsis may have up to a 100% increased risk of death at 30 days depending on the hospital to which he/she is admitted in the VA system. Significance/Impact: This proposal is specifically designed to address three priority domains of the HSR&D Service - Healthcare Informatics, Quality and Safety of Health Care, and Virtual Care. Through the unique combination of these three domains we plan to address two significant limitations in understanding sepsis in the VA; 1) that existing reports may or may not provide insight into the distinguishing characteristics of the patients that died with sepsis in what are thought to be similar VA ICUs, and 2) in these reports there are multiple interventions known or suspected to improve outcomes with varying levels of efficacy about which little or no information is offered. Innovation: We can use a novel data source (TeleICU) to directly identify the patient level factors associated with negative outcomes in the septic patient population, quantify the practices of high and low performing units, and subsequently improve the care provided to septic patients in VA ICUs using this information. Specific Aims: Aim #1: Evaluate risk factors of negative outcomes for specificity to sepsis and within sepsis types to determine best strategies for adjustment and calculate risk-adjusted ICU mortality rates. Aim #2: Evaluate the incidence of previously unmonitored elements of sepsis care (hypotension, ventilator management, and antibiotics) and their impact on outcomes cited in SA1a (ICU mortality as primary outcome) in VA sepsis patients. Aim #3: Qualitatively examine the management of hypotension, lung protective ventilation, and antibiotic therapy in sepsis patients at high and low performing sites (based on appropriately adjusted ICU mortality rates). Evaluations will include documented protocols, clinical workflows, and TeleICU support. Methodology: We will conduct a mixed methods investigation by merging data from the TeleICU and Corporate Data Warehouse to first identify high and low performing ICUs in the treatment of sepsis. We will then perform an ethnographic investigation of 3 high and low performing ICUs in the treatment of sepsis. Simultaneously, we will examine the management of hypotension, the use of lung protective ventilation, and the use and efficacy of antibiotics in septic patients in the ICU. Next Steps/Implementation: We will develop real-time clinical decision support, to provide local clinicians with updates on septic patients in the ICU that provide information regarding the state of septic patients and their compliance with metrics that are associated with improved outcomes. We anticipate this will improve the overall survival of sepsis patients and potentially reduce the cost of care.

The outcomes of septic patients in the VA are known to vary depending on facility and region of treatment. Typically, the care of septic patients at high risk of death is provided in the ICU. While there are known treatment modalities that improve the outcomes of sepsis patients, it is unclear what these modalities are and if these modalities are used in the VA consistently. This proposal will use advanced datasets, previously not available, and qualitative measures to examine the care delivered in septic patients in the VA. In this proposal, we describe using data from the TeleICU and other VA sources to develop a better understanding of the care delivered in VA ICUs to look for variation in practice and modalities that are associated with improved outcomes in sepsis.