A large health system in the state of Washington is making a major push to combat sepsis, the potentially life-threatening condition that affects about 1.7 million Americans each year and a leading cause of hospital readmissions.
CHI Franciscan, a system based in Tacoma, WA, with eight acute care hospitals in the region, launched its Inpatient Code Sepsis program in spring 2019 to identify patients at high risk for sepsis and to put them on antibiotics faster. The health system saw success with a similar program in its ED before rolling it out systemwide.
The sepsis program uses predictive analytics in the electronic health record (EHR) to identify patients who are at risk of developing sepsis, explains Griffith Blackmon, MD, MPH, a critical care physician with CHI Franciscan Health. The tool runs constantly in the background of the EHR and provides an alert to bedside staff when a certain risk of developing sepsis score is identified, he says.
The initiative carries great potential for improving quality of care and reducing admissions, Blackmon says, because research indicates 80% of sepsis-related deaths could be prevented with rapid diagnosis and treatment.1 Educating staff about sepsis risks and best practices also is key, Blackmon adds. (Editor’s Note: For information on educating clinicians about sepsis, see the sidebar at the end of this article.)
“Sepsis is a big problem in healthcare, with the CDC estimating that about 1.7 million Americans develop sepsis every year, and upwards of 270,000 Americans die of sepsis every year,” Blackmon says. “As much as one-third of all hospital deaths are associated with sepsis.”2
The ED program that eventually spawned the Inpatient Code Sepsis program at CHI Franciscan was initiated in June 2017 with a process that used a paper-based screening tool in triage to identify patients at high risk for sepsis. Once identified, the process targeted those patients for prevention and treatment efforts throughout their stay at the hospital.
“That proved to be very successful and widely embraced throughout our system. We have data for two of our hospitals that showed a reduction of more than 30 minutes in the time from triage to receiving antibiotics,” Blackmon says. “That was within six months of implementation. It clearly showed that we could improve the care we deliver with regard to this very serious problem.”
CHI Franciscan wanted to take that ED success with sepsis and apply it to the inpatient setting. Most sepsis patients in the hospital presented that way, but some develop sepsis once admitted, Blackmon says. Part of the challenge is to identify those at risk and to recognize the issue once it begins.
The health system leveraged a predictive analytics tool from its EHR vendor that works in the background of the electronic record, analyzing 85 parameters to calculate a score for risk of sepsis. The tool is completely automated and requires no input or intervention by physicians and staff, Blackmon explains.
“The risk of sepsis score updates every 30 minutes for every patient who is admitted,” Blackmon explains. “We validated the risk of sepsis score on about 15,000 patients before we went live and used that validation data to determine the points that we could consider a critical score. We made that decision based on a balance between specificity and sensitivity, wanting the screening tool to be sensitive enough to find patients, but specific enough to reduce false alerts.”
That is a challenge with any screening tool, Blackmon notes, so the health system had to decide on the right balance of the two goals. CHI Franciscan settled on parameters that generate an alert to the bedside nurse when a patient is at risk for sepsis. The risk assessment score also is available to all clinicians on their worklists, updated constantly, and color coded for the level of risk. The tool reports the sepsis risk assessment as green for very low, yellow for intermediate, and red at the threshold for triggering an alert requiring further evaluation.
“That alert triggers a response developed in collaboration with frontline staff from across our system, all six hospitals. The workflow is that the bedside nurse seeks a second set of eyes, and who that is can depend on the hospital, because the resources are different at each site,” Blackmon says. “One of the nurse’s colleagues joins to take a look at the data and together they determine if the sepsis has been recognized and if it is being treated by the attending providers.”
If it is clear that sepsis is known and being treated, and the patient condition has not changed, the nurse does not contact the attending provider and continues plan of care. If there is ambiguous information in the medical record and the patient is not currently being treated for sepsis, the provider is contacted. Nurses use a sepsis-specific SBAR communication to alert providers that the patient may be developing sepsis.
The physician determines whether to initiate a Code Sepsis. If the provider believes the sepsis risk needs more attention, the team initiates a Code Sepsis, which sets off several actions throughout the hospital, Blackmon says. The code notifies the lab to send a phlebotomist to the bedside and allows the nurse to generate an order set for blood draws, blood cultures, chemistries, and other needs. The code also notifies the pharmacist that the physician may be entering orders that should be triaged to the top of the verification list, particularly for changes in the antibiotic therapy.
Blackmon says the Code Sepsis program was designed to avoid creating unnecessary interruptions or burdens for clinicians, seeking instead to provide a “thoughtful pause” that can allow them to improve care for their patients.
“We’re offering a screening tool to help people recognize the presence of a potentially fatal problem, but we don’t want that tool to automatically initiate a cascade of interventions that may or may not be appropriate,” Blackmon says. “We intentionally made the decision to initiate a Code Sepsis the responsibility of the attending physician so that we left a significant human decision-making component in the process.”
It also is important to make sure the correct provider is consulted about possible sepsis, Blackmon says. For instance, a surgeon and a hospitalist co-manage many patients. Involving the right physician can be key to either addressing possible sepsis or avoiding an unnecessary response.
“Many of the warning signs of sepsis are normal occurrences in the postoperative period and not necessarily indicative of sepsis,” Blackmon says. “When we are asking if these symptoms are indicative of sepsis and whether we need to act further, we want to be sure we are asking the right provider. In our case, we made the decision with those co-managed patients that if the patient was within 48 hours of surgery, it would be the surgeon who was engaged in that conversation rather than the hospitalist.”
The results have been encouraging, Blackmon reports. Within 60 days of implementation, there was a 10% reduction systemwide in the observed-to-expected mortality for sepsis.
Data gathering with the project began with measuring compliance with the sepsis bundle and mortality codes, both at the system level and the individual ICU level, Blackmon says. For determining the observed-to-expected mortality for sepsis, CHI Franciscan used its data warehouse platform, explains Melissa Morris, RN, MSN, CEN, sepsis coordinator with the health system.
The data warehouse provided mortality data, and the health system continues monitoring patients who triggered a sepsis alert, nursing response, and physician response, Morris says. Two of the system’s hospitals employ sepsis coordinators who also monitor ED time to antibiotics for all sepsis patients, she adds.
The sepsis program coordinator is a registered nurse who works in collaboration with the medical and hospital staff to provide multidisciplinary care to sepsis patients. The coordinator provides clinical, professional expertise for patients, families, and staff members. The coordinator collects and analyzes data surrounding the sepsis patient. They also assist in sepsis program development as it relates to the care of all sepsis patients. The sepsis coordinator researches and drives implementation of evidence-based practice as related to sepsis care. Sepsis coordinators also monitor ED time to antibiotics for all patients diagnosed with sepsis in the ED.
A checklist built into the EHR also aids data gathering. It allows nurses to not only walk through the steps of the Code Sepsis response, but also to document when actions were taken and the data gathered through laboratory tests and other measures, Blackmon notes.
“We are able to generate reports that pull information out of that flow sheet. That’s a strategy widely used in many EHRs. If you have data in a flow sheet, you’re able fairly easily to generate a report that is useful in how it compiles the information from many patients,” he says. “We built the flow sheet with the idea that it would be the source of that information when we wanted to track the response of the Code Sepsis program.”
One challenge was getting the necessary information for working with the Code Sepsis program to the bedside nurse, Morris says. With more than 1,000 nurses in the health system, it was not possible to sit with each of them, so CHI Franciscan had to educate all of them without interfering too much with their already busy work schedules, she says.
The health system sent out practice updates to bedside nurses, and sepsis experts attended various nursing meetings, Morris says. The process continues with additional practice updates to all nurses, she says.
Another challenge was trying to keep the project on task and on schedule, says Laura Johnson, a process improvement consultant with CHI Franciscan Health. There were technical challenges with implementing the predictive analytics tool that the CHI Franciscan team wanted in the EHR, Johnson recalls.
“If you’re trying to merge and operationalize the protocol with an electronic component, be aware of that potential difficulty up front. There were a lot of things with our EHR vendor, as far as when they were going to roll things out, that didn’t align with our original timeline,” Johnson explains. “Some of that was out of their control, but it’s a reminder that if you’re going to be implementing a standard work protocol and you’re integrating it into an electronic record system, you need to try to stay in sync as much as possible. In our case, we did have a delay, but our team worked to keep our physicians and nurses engaged during that delay.”
One key to success with the program was how CHI Franciscan tailored the Code Sepsis program to the resources available at each individual hospital or other facility, Johnson says. The standard protocol was developed for the health system, but then team leaders went to each facility to operationalize it based on their resources, she explains. “This was one way we kept everyone engaged while we had delays on electronic integration of the analytics. We worked with each facility to complete a 90-, 60-, and 30-day readiness assessment plan, keeping them engaged to think about how they were going to react to specific components of the sepsis plan based on what they had available at their sites,” Johnson says. “To keep people engaged, you have to be in contact with people frequently and give them tasks so that they feel like a part of the program and want to help you push it forward.”
Once the program went live, CHI Franciscan held weekly check-and-adjust sessions to gather feedback from each facility, Blackmon notes. Those sessions prompted several changes to the workflow within the first couple of weeks. The feedback solicitation continued with 30-, 60-, and 90-day checkups. “That interaction was critical to maintaining momentum and keeping the program alive. It gave users the opportunity to provide meaningful feedback and fine-tune the system,” Blackmon says.
- Kumar A, Roberts D, Wood KE, et al. Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med 2006;34:1589-1596.
- Centers for Disease Control & Prevention. Sepsis. Data & Reports. Available at: . Accessed Nov. 1, 2019.
- Griffith Blackmon, MD, MPH, CHI Franciscan Health, Tacoma, WA. Phone: (888) 825-3227.
- Melissa Morris, RN, MSN, CEN, ED Nurse, CHI Franciscan Health, Tacoma, WA. Phone: (888) 825-3227.
- Laura Johnson, Process Improvement Consultant, CHI Franciscan Health, Tacoma, WA. Phone: (888) 825-3227.