(1398) Discrepancy Between Discharge Diagnosis and Sepsis Identified by a Regular Nursing Floor Sepsis Team

Introduction: While outcomes for patients with present on admission (POA) sepsis have improved over the last two decades, outcomes for patients developing not present on admission (NPOA) sepsis remain poor. Use of a dedicated team to identify patients on regular nursing floors (RNFs) with signs of sepsis represents a potential mechanism to improve early detection of patients with NPOA sepsis. Here we report the discrepancy between the number of likely septic events identified by a sepsis team dedicated to RNF patients and discharge diagnoses assigned to those patients by primary team providers.

Methods: The Sepsis Emergency Response Team (SERT) is active across adult RNFs of a 1400 bed academic center in Cleveland, OH. The program is staffed by Advanced Practice Providers and utilizes an electronic alert triggered by the presence of systemic inflammatory response syndrome criteria and organ dysfunction to screen patients for signs of sepsis. We reviewed cases identified by SERT as having likely sepsis from January through December of 2023 and determined how many of these patients were assigned an ICD10 sepsis diagnosis code by primary team providers at hospital discharge.

Results: SERT providers identified 411 alerts across 399 unique encounters as consistent with likely sepsis. 253 (61.6%) alerts occurred greater than 48 hours after hospital admission. Primary services with the highest percentage of likely septic events were internal medicine (28.7%), hematology/oncology (22.9%), and gastrointestinal surgery (12.7%). 26.3% of likely septic events were associated with transfer to an intensive care unit within 24 hours. 192 (48.1%) encounters received an ICD10 sepsis discharge diagnosis; in-hospital mortality for these patients was 18.8%. Mortality was 4.8% among 207 (51.9%) patients with likely sepsis who did not receive an ICD10 sepsis discharge diagnosis.

Conclusions: 51.9% of hospital encounters featuring at least one likely septic event identified by a dedicated team did not subsequently receive an ICD10 sepsis discharge diagnosis. This suggests inconsistency between clinical concern for sepsis and applied sepsis diagnosis codes. Further studies are needed to better define this discrepancy and its impact on sepsis epidemiology and downstream outcomes.