Sepsis has been one of the leading causes of death across the world for many years. The 2021 Surviving Sepsis Campaign provides guidelines for early detection, recognition and intervention. Our healthcare system tried several ways to treat patients with sepsis. We tried sepsis screening on paper, a variety of alerts in the electronic medical record (EMR), and recently embarked on a journey to use tele-ICU, also known as tele-critical care, to proactively screen patients for sepsis. Through trial and error, we created a program that benefits patients and nurses.
Tele-ICU Sepsis Surveillance Objectives
- Increase sepsis screenings and actions taken.
- Reduce the alerts and workloads of bedside nurses.
- Create a flexible staffing model.
- Use innovative technologies such as artificial intelligence (AI).
- Improve patient outcomes for sepsis.
Tele-ICU Sepsis Nurse
Initially, bedside nurses were asked to complete the sepsis screens for patients who received Best Practice Alerts. We found that this work led to alert fatigue for the nursing staff. The alerts were often dismissed because the staff was completing other tasks. Our leadership team decided that a dedicated screening nurse could reduce the number of alerts the bedside nurses received and increase the number of screens completed. We employed light-duty critical care nurses who could work virtually using their critical care expertise while recovering from various injuries. This process provided evidence-based care while decreasing the workload at the bedside.
Integrating AI Into the EMR
We created a list in the EMR that displayed all the high-risk patients with sepsis in our health system. This list was generated by an algorithm that identified patients who had been risk stratified by the predictive model and narrowed the list to those at highest risk. Our team was involved in developing the list, which displays completed tasks such as lactates, antibiotics and blood cultures. The list allows our team to quickly visualize what was completed and prioritize patients who are not currently on antibiotics.
Implementation of the New Nursing Practice Change
The following steps outline our implementation of the new nursing practice on sepsis screening:
- A small group of ICU nurses worked with tele-ICU staff to begin screening the sepsis list to reduce the number of alerts the bedside nurses received.
- Our virtual sepsis nurses were educated through webinars and classes that Sepsis Alliance offered. We also attended unit staff meetings to educate bedside staff about our process, the role of the virtual sepsis nurse and the role of bedside nurses.
- The tele-ICU nurse screened a list of patients who had already been triaged by the algorithm. The algorithm helped screen 500 patients to create a smaller group of 50 patients.
- Then the tele-ICU nurse reviewed the chart to complete the sepsis screen. A patient was screened as positive if they had a possible source of infection, met SIRS criteria and developed organ dysfunction.
- A nurse-driven protocol directed nurses to place an order for a lactate blood draw for positive patients.
- Once the lactate was ordered, the tele-ICU nurse communicated with the physician and bedside staff.
- The tele-ICU nurses helped bedside staff clarify any unclear expectations about the nurse-driven protocol and place new orders.
- We created documentation pathways to alert providers that a patient was at high risk for sepsis and needed follow-up.
Evaluation of Outcomes
Our initial objectives were met, and our team of nurses provided sepsis surveillance for hundreds of patients per week without having to involve bedside staff. As the number of screenings increased, risk-adjusted mortality decreased. Virtual sepsis surveillance quickly became a crowd-pleaser among bedside nurses and physicians. Patients received individualized care from multidisciplinary teams, thus improving outcomes. By decreasing alert fatigue, interruptions and cognitive burden for bedside staff, the program helped our system achieve its patient safety goals.
Expansion and Evolution of Sepsis Surveillance
This program began as a pilot project at one of our community hospitals that consisted of 100 beds. We later transitioned the program to include our largest facility with 700 beds, one unit at a time. As we expanded, we thought we would need to increase staff. Surprisingly, the number of sepsis alerts decreased, and no additional staff was needed. We have provided coverage for two hospitals with approximately 800 inpatient beds and two emergency departments with the same number of virtual nurses. Our unit guarantees that patients will be screened from 9 a.m.-3 p.m., because we are able to staff that timeframe easily with our available resources.
When expanding to a new unit or facility, we include a presentation to unit leadership about the project, data analysis and education. We use LEAN methodology in project planning, and pre-post outcomes are shared with unit leaders during the project cycle.
The goal of our project was to provide support for our healthcare system, with a priority to increase awareness of sepsis, while improving screening volume, accuracy and actionable interventions, and also decreasing alert fatigue. Alert to Action: Implementing Artificial Intelligence-Driven Clinical Decision Support Tools for Sepsis explores additional aspects of our story from the perspectives of other team members.
Our hope is to eventually support bedside staff on every campus in our health system while ensuring they remain educated and aware of the signs of sepsis.
