Lost in translation: sepsis bundle implementation gaps in medical wards and implications for a hospital-wide Sepsis Team

Background: Early recognition and prompt management of sepsis are essential to improve patient outcomes. While evidence-based sepsis bundles are widely recommended, their implementation in non-critical care settings remains challenging. This study aimed to assess adherence to the sepsis bundle across medical wards identify barriers to implementation, and provide evidence supporting the development of a hospital-wide multidisciplinary Sepsis Team. Methods: We conducted a cross-sectional point-prevalence study at the University Hospital of Modena between April and May 2025. Adult patients admitted to medical wards with sepsis or septic shock were included. The sepsis bundle was defined as completion of serum lactate measurement, blood cultures before antibiotic administration, timely empirical antimicrobial therapy, and intravenous fluid resuscitation. Bundle completion within 3 and 6 hours from sepsis recognition was evaluated. The primary composite outcome was 30-day all-cause mortality and/or unplanned hospital readmission within 90 days. Secondary outcomes included hospital length of stay, intensive care unit (ICU) admission, length of stay, 90-day mortality and readmission. Results: Seventy-one patients were included; 68.6% presented with sepsis and 22.5% with septic shock. The median Charlson Comorbidity Index was 5 and clinical outcomes were poor, with 32 in-hospital deaths (44.4%), 6 ICU admissions (8.3%), and 20 readmissions (27.8%) during the study period. Overall, the sepsis bundle was completed in 47.1% of cases, while only 37.1% and 45.7% achieved bundle completion within 3 and 6 hours, respectively. Lactate measurement represented the major implementation gap, being performed in only 57.7% of patients. Bundle completion occurred more frequently in patients with greater illness severity, including hypotension, septic shock, higher NEWS2 scores, and elevated SOFA scores. Internal Medicine and Infectious Diseases wards demonstrated higher adherence compared with oncological and surgical settings. Bundle completion was not independently associated with improved clinical outcomes, likely reflecting confounding by indication. Multivariable analysis identified age ≥75 years (aOR 15.6, p=0.006) and culture-negative sepsis (aOR 9.37, p=0.006) as independent predictors of the composite outcome, whereas urinary tract infection was protective (aOR 0.13, p=0.028). Conclusions: Sepsis bundle implementation in medical wards was suboptimal and heterogeneous, with significant diagnostic and organizational gaps, particularly lactate measurement. Performance was largely driven by clinical severity rather than standardized processes. The high mortality and readmission rates underline the clinical impact of delayed or incomplete management. These findings support the establishment of a hospital-wide Sepsis Team to promote early recognition, improve bundle adherence, standardize sepsis management, and strengthen collaboration between emergency, medical, microbiology, intensive care, and infectious diseases services.