P191 The impact of changes in sepsis coding on mortality reports

Background Coding practices impact the recorded prevalence of disease. Sepsis, most commonly caused by pneumonia, remains a common cause of mortality worldwide. In April 2017, NHS Digital issued new guidance on coding of septicaemia. Following this, many hospitals reported an increase in mortality from sepsis. We hypothesised that this altered mortality was artefact due to changes in how primary admission diagnosis was recorded during processing of hospital data. Methods Hospital Episode Statistics from the Admitted Patient Care dataset for NHS hospitals in England, from April 2016 to March 2018 were included (twelve months before and after change in coding practices). Adult patients with an International Classification of Diseases 10 (ICD-10) code associated with the Agency for Healthcare Research and Quality (AHRQ) Clinical Classifications Software (CCS) class ‘Septicaemia (except in labour)’ recorded for an inpatient episode were identified. Patient comorbidities were assessed using ICD-10 codes recorded within the admission episode. Results 514,678 hospital episodes with a coded diagnosis of sepsis were studied. After the coding change, there was no increase in the number of episodes where sepsis was coded as a secondary diagnosis, but a significant increase in the number of episodes where sepsis was coded as the primary reason for admission, for all demographic factors studied (sex, ethnicity, social deprivation index). This was sustained for the whole study period. There were significant differences in the case-mix of patients with a primary diagnosis of septicaemia before and after the coding change, including the recorded prevalence of specific comorbidities such as asthma and lung cancer, and an increase in the proportion of patients with a primary diagnosis of sepsis who had any recorded respiratory disease. Reported differences in outcomes could be accounted for by this change in case-mix. Conclusion Changes in coding practices can cause important differences in the types of patients where the diagnosis of interest is recorded, which in turn can influence reported outcomes. The impact of any coding change should be considered carefully when reviewing longitudinal trends in outcomes for any disease where coding practice or diagnostic definitions have changed, within local reporting measures and also in published literature.