TOH – The Ottawa Hospital Academic Medical Organization
TOH-21-014 – Introducing and evaluating Sepsis Advisor: a clinical decision support tool to help predict and prevent future unexpected deterioration of patients with infection, including COVID-19
Intensive care units (ICUs) and Emergency Departments (EDs) are at the forefront of the COVID-19 challenge. There is a need to identify those patients presenting to the ED with the highest risk of deterioration to optimize decisions on where to admit the patient (i.e. disposition decision making). Patients with infection who are discharged home or admitted to the ward, and subsequently deteriorate and require ICU admission have worse outcomes and cost far more than patients admitted directly to ICU from ED. While COVID-19 patients present with mild respiratory symptoms, this can progress rapidly to hypoxemic respiratory failure and/or septic shock, associated with very high mortality. By identifying patients presenting with infection who are at increased risk of short-term deterioration, clinicians could more accurately identify optimal disposition: home, the hospital ward, acute monitoring areas, or the ICU. We have developed a novel tool, Sepsis Advisor (SA), to provide disposition decision support to physicians assessing patients with infection in the ED. The SA predictive model is generated by analyzing the novel dimension of heart rate variability (HRV) combined with clinical and laboratory data, to identify high-risk patients with increased risk of deterioration, and low-risk patients with decreased risk of deterioration. Previous analysis of the model in 832 ED patients presenting with infection identified a high-risk cohort with a 4.3-fold (95% CI, 3.2–5.4) increased risk of deterioration (i.e. occurs with a prevalence of 35%), and low-risk group with 0.2-fold (95% CI 0.1–0.4) risk of deterioration (prevalence of 2%). Although there are no data regarding the utility of SA and its predictive models for COVID-19 patients, we hypothesize that our predictive model will perform well, as it is based on a physiological association between loss of variability and increased illness and loss of physiological reserve that should hold true for viral infection. Thus, this observational study provides the first opportunity to assess the feasibility of implementing SA in the ED and lays the foundation for a future randomized controlled trial. We aim to evaluate SA in 100 patients presenting to the ED with suspected infection and interview 15-30 physicians and 10-20 nurses to identify their opinions regarding SA. We initiated this study at The Ottawa Hospital in August 2021 and have enrolled 34 patients as of November 2021, including 2 COVID-19 positive. SA implementation has been technically feasible, with waveform data captured and an SA report generated >90% of the time. Interviews have been completed with 8 ED physicians, all of whom reported positive opinions associated with the SA tool and provided valuable feedback for future iterations of the SA report. The SA report template has been updated twice, with the most recent version shown in subsequent interviews. We anticipate that enrollment for this study will be complete in early 2022. We will expand interviews to include internal medicine and critical care physicians, as their feedback will also be essential to generating an SA report that supports disposition decision making in a clinical setting. Integration of the SA tool in the ED will be expanded to include nursing staff in the wave form data capture and data entry process. Upon completion of this study, we aim to confirm that SA implementation in the ED is feasible and that initial perceptions are positive. A future randomized controlled trial would evaluate improvements to the quality and efficiency of care for patients with suspected infection, including COVID-19, by optimizing disposition from ED to ensure best clinical outcomes are achieved, minimizing unexpected deterioration at home or on the ward requiring harmful, resource-intensive, and costly ICU transfer. While critical to COVID-19 patients, results would also more widely apply to all patients with ED infection.
Methods of Risk Reduction and Mitigation
Primary Project Lead for Contact
Secondary Project Lead for Contact