Accuracy of real-time SNOMED-CT coding by clinicians in an urban tertiary emergency department: A retrospective cohort study. (September 2022)

• Record Type: Journal Article
• Authors: Jones, Peter; Ho, Jess; Burbridge, Hayley; Hill, Dane; Khalil, Rana; Raumati, Inia
• Is Part Of: International journal of medical informatics. Volume 165(2022)
• Publication Date: 2022-09
• Subjects: Emergency medicine -- SNOMED-CT -- Ethnicity -- Inequities; Medical informatics -- Periodicals; Information science -- Periodicals; Computers -- Periodicals; Medical technology -- Periodicals; Medical Informatics -- Periodicals; Technology, Medical -- Periodicals; Computers; Information science; Medical informatics; Medical technology; Electronic journals; Periodicals; Electronic journals; 610.285
• DOI: 10.1016/j.ijmedinf.2022.104813
• Languages: English
• ISSNs: 1386-5056

Abstract

Graphical abstract: Highlights: Clinician recorded diagnosis using SNOMED-CT accurately reflects clinical notes. Recording of procedures using SNOMED-CT was inadequate. Inequities for Māori were found with respect to reporting discharge diagnoses. Inequities were due to systemic bias in the procedure for extracting diagnosis for reporting. Abstract: Objective: This study aimed to determine the accuracy of the reported diagnoses and procedures to the National non-Admitted Patient Collection (NNPAC) from Auckland City Hospital Adult Emergency Department, and whether there were disparities between Māori and non-Māori patients. Methods: We audited 5788 (n = 594 Māori, 5194 non-Māori) visits in February 2021 to determine whether diagnoses and procedures were recorded and whether these were recorded differently for Māori compared to non-Māori. A random sample of case notes, stratified by five common chief presenting complaints (n = 114) were selected to compare clinician recording of diagnoses and procedures in real time, to those derived from the clinical notes by auditors blinded to the actual diagnosis and patient name and ethnicity. The New Zealand Emergency Department SNOMED-CT reference set was used to code diagnoses. Results: Māori were less likely to have a diagnosis recorded when discharged from the ED compared to non-Māori, relative risk 1.48 (1.08, 2.04), p = 0.016 (n = 3045). Failure to record diagnoses was due to flaw in the system for extracting diagnoses from electronic notes, rather than failure to make a diagnosis. There was agreement in 111/114 cases for diagnosis: 53/56, 94.6% (95 %CI 85, 99) for Māori, and 58/58, 100% (95 %CI 93, 100) for non-Māori; p = 0.115. There was agreement in 60/114 cases for procedures completed: 31/56, 55.4% (95 %CI 42, 66) for Māori, and 29/58, 50% (95 %CI 38, 62) for non-Māori; p = 0.567. Conclusion: Māori were less likely to have a diagnosis recorded at discharge due to systemic bias in how we captured diagnoses electronically. Our system should change to remove this inequity. The diagnoses recorded using SNOMED-CT were mostly an accurate reflection of clinician's notes, while recording of procedures was poor.