When Clinicians Disagree on Shock Cause, Septic Shock Patients Face Higher ICU Death Risk

When intensive care teams disagree about what is causing a patient’s shock, that disagreement may itself signal a higher risk of death. A multicenter retrospective cohort study published in Critical Care on 29 October 2026 examined 2,233 patients with septic shock who met Sepsis-3 organ-dysfunction criteria across eight acute care hospitals between 2018 and 2023. The researchers, led by Saieesh A. Rao and Anne M. Stey with colleagues at multiple institutions, found that 213 patients, or 9.5 percent, showed early disagreement among clinicians about the etiology of the shock, and that this disagreement was associated with increased ICU mortality in a specific subgroup of patients.

The study’s method rested on what clinicians actually wrote in the medical record. Shock etiologies documented by clinicians within the first 48 hours of ICU care were extracted using regular expressions and compared using Cosine Similarity Scores, a text-similarity measure in which identical documentation yields a score of 1. Disagreement was defined as a score below 1, meaning the documented explanations for the shock diverged. The researchers then tested the association between disagreement and mortality using an inverse-probability-of-treatment-weighted multivariable logistic regression, adjusting for demographics, comorbidities, SOFA score, blood culture positivity, and ICU specialty. Additional multivariable Fine-Gray models assessed whether disagreement affected how quickly patients received antibiotics, bolus fluids, or vasopressors, and a mediation analysis tested whether differences in those treatment times explained any mortality association.

The pattern of disagreement itself varied by hospital type. Disagreement was documented for 13.3 percent of patients at the academic medical center compared with 7.2 percent at community sites, a difference that was statistically significant (p < 0.001). The mortality association, however, was concentrated among patients with positive blood cultures. Effect modification of the disagreement-mortality association by blood culture positivity was significant (OR 2.37, 95% CI 1.03–5.45, p = 0.042). Within the blood culture-positive subgroup, disagreement was associated with a 1.99-fold increase in adjusted odds of ICU mortality (95% CI 0.98–4.01, p = 0.056), and in linear-probability terms an absolute risk increase of 16.9 percentage points (95% CI +1.1 to +32.7, p = 0.036). The confidence interval for the odds ratio crossing 1 means this stratified estimate should be read as borderline rather than definitive, even though the overall effect-modification test reached significance.

Perhaps the most consequential finding is what did not explain the association. Disagreement was not associated with different rates of antibiotic, fluid, or vasopressor administration in the blood culture-positive subgroup, and the mediation analysis found that the mortality association was not explained by time to antibiotics, bolus fluids, or vasopressors. This challenges the study team’s original hypothesis, which proposed that early disagreement would lead to treatment delays and thereby raise mortality. Instead, the authors conclude that early inter-clinician disagreement on shock etiology functions as a prognostic marker of increased ICU mortality among Sepsis-3 septic shock patients with positive blood cultures, independent of the timing of standard septic shock therapies.

The clinical implication is that the Surviving Sepsis Campaign’s emphasis on early antibiotics, fluids, and vasopressors presumes an earlier step: recognizing sepsis as the cause of shock in the first place. Septic shock mortality remains high despite widespread dissemination of those guidelines, and this study suggests that when care teams document divergent explanations for a patient’s shock in the first two ICU days, that discordance carries prognostic information beyond treatment timing. As a retrospective observational study, it establishes association rather than causation, and the mortality signal was statistically robust only in the overall effect-modification test, with the stratified odds ratio narrowly including the null. The authors note the analysis was limited to patients meeting Sepsis-3 criteria at eight hospitals between 2018 and 2023, and the automated text-comparison approach captures documented disagreement rather than every instance of clinical uncertainty. Even so, the finding points toward a measurable, record-based signal that could identify higher-risk septic shock patients early in their ICU stay.

Subject of Research: Inter-clinician diagnostic disagreement on shock etiology and its association with ICU mortality in septic shock patients

Article Title: Association of inter-clinician diagnostic disagreement with increased mortality among patients with septic shock: a retrospective cohort study

Article References: Rao, S. A., Zhan, T., Garg, R., Yost, C. C., Furmanchuk, A., Nadig, N. R., Huml, J., Shamshuddin, L., Chou, S., Al-Qamari, A., Perigo, E., Vorovich, E., Slocum, J., Berry, A. B. L., Rafferty, M. R., Holl, J. L., & Stey, A. M. (2026). Association of inter-clinician diagnostic disagreement with increased mortality among patients with septic shock: a retrospective cohort study. Critical Care. https://doi.org/10.1186/s13054-026-06384-0

Image Credits: AI Generated

DOI: 10.1186/s13054-026-06384-0

Keywords: septic shock, diagnostic disagreement, ICU mortality, Sepsis-3, retrospective cohort study, blood cultures, Surviving Sepsis Campaign, antibiotics, vasopressors, mediation analysis, Association, inter-clinician