Coding practice for sepsis 2008-21.

Background: Sepsis has a high incidence and mortality rate. Accurate data are needed for health service planning and for research, and there is a need to identify coding practices in Norway.
Material and Method: All patients over 17 years of age who had been admitted to Norwegian hospitals with sepsis in the period 2008-21 were identified using diagnostic codes for infection plus organ failure, and specific codes for sepsis, from the Norwegian Patient Registry.
Results: There were 317 705 admissions with diagnostic codes for sepsis, of which 210 391 (66.2 %) were sepsis with a known focus, 77 627 (24.4 %) were of unknown focus and 29 687 (9.3 %) were codes for both a known and unknown focus. The percentage of sepsis episodes coded with a known focus varied between the health regions. The highest percentage was in the Western Norway Regional Health Authority (72.1 %, 95 % confidence interval (CI): 71.8 to 72.5), and the lowest was in the Central Norway Regional Health Authority (59.2 %, 95 %, CI 58.7 to 59.7). The use of codes with a known focus increased each year on average by 3.2 % (95 % CI 2.7 to 3.6, from 47.5 % in 2008 to 82.3 % in 2021), while the use of codes with an unknown focus decreased by 2.3 % (95 % CI -2.7 to -1.9) from 37.8 % in 2008 to 13.0 % in 2021. Known and unknown focus combined also decreased by 0.9 % per year on average (95 % CI -1.0 to -0.8) from 14.3 % in 2008 to 4.1 % in 2021.
Interpretation: The coding of sepsis in Norwegian hospitals has become more uniform.