The following is a summary of “Adjudication of codes for identifying sepsis in hospital administrative data by expert consensus,” published in the October 2024 issue of Critical Care by Garland et al. 

Researchers conducted a retrospective study to refine the administrative data definition of patients with sepsis admitted in hospitals, encompassing severe and less severe cases.  

They assessed Calgary’s adult population (1.033 million), of which 61,632 eligible hospitalizations were compared to different algorithms, used 1,928 infection codes and 108 organ dysfunction codes to develop a new algorithm, AlgorithmL, which required at least 1 infection and 1 organ dysfunction code to be considered sepsis. AlgorithmL with 4 existing algorithms were compared with the Centers for Disease Control’s Adult Sepsis Event (ASE) definition to assess the performance in identifying sepsis cases in ICU and non-ICU settings.  

The results showed AlgorithmL with 720 infection codes and 50 organ dysfunction codes, while comparing algorithms with variations from 42 to 941 infection codes and 2 to 36 organ codes. There was a significant overlap between AlgorithmL and the comparators. Annual sepsis incidence rates (per 100,000 population) based on AlgorithmL were 91 in the ICU and 291 in the non-ICU cohort, compared to 28 to 77 for ICU and 11 to 266 for non-ICU cohorts using comparison algorithms. Hospital sepsis mortality rates based on AlgorithmL were 24% in ICU and 17% in non-ICU cohorts, while comparison algorithms showed rates of 27 to 38% in ICU and 18 to 47% in non-ICU cohorts, 41% met the ASE criteria, compared to 42 to 82% for the comparator algorithms of AlgorithmL-identified cases.   

They concluded that AlgorithmL codes more infection and organ dysfunction codes than other algorithms and identified more sepsis cases with lower mortality rates. 

Source: journals.lww.com/ccmjournal/fulltext/9900/adjudication_of_codes_for_identifying_sepsis_in.389.aspx