The association between sepsis and diagnostic errors: A secondary analysis of the Utility of Predictive Systems for Diagnostic Error study

Abstract Background Diagnostic errors (DEs) are common in patients who die or go to the intensive care unit (ICU), due to gaps in diagnostic processes, but it is unclear if this holds true for sepsis patients. Objectives To determine whether patients with sepsis, as defined by International Classification of Disease‐10 (ICD‐10) code, have a different DE risk and/or underlying diagnostic process faults compared with other patients who die or are admitted to an ICU. Methods We conducted a retrospective cohort study of two thousand four hundred and twenty‐eight adult patients discharged from 29 academic medical centers in 2019 were included in the vizient clinical database and the hospital medicine re‐engineering network, meeting centers for medicare & medicaid services criteria for a “medical” admission. Patients included either died in‐hospital or were admitted to the ICU during the studied admission. Safer Dx and diagnostic error evaluation and research (DEER) taxonomy defined DE and diagnostic process faults, respectively. Adjusted risk ratios (aRRs) for DE were estimated using modified Poisson regression clustered by site. Interaction terms tested differences between patients with and without sepsis. Results Of 1218 sepsis patients (50.2%), DE rates were similar to nonsepsis patients (22.3% vs. 23.9%, p = .734). Adjusted analysis showed no association between sepsis and DE (aRR: 1.06, 95% confidence interval [CI]: 0.91–1.23). Sepsis was less strongly associated with DE if gaps in history‐taking (interaction‐term p = .02) or consultation problems (interaction‐term p = .005) were present. No other DEER interactions were significant. Conclusion Sepsis was not associated with a higher adjusted risk for DEs or individual diagnostic process faults, suggesting diagnostic issues in sepsis are tied to broader clinical care problems rather than sepsis‐specific factors.