A way to audit compliance with the Surviving Sepsis Campaign bundles

Introduction

A prospective multicentre nationwide study on community-acquired sepsis (CAS) in Portuguese ICUs was carried out with two main objectives: to promote good practice through implementation of the Surviving Sepsis Campaign (SSC) recommendation and to know the epidemiology, which is essential in the treatment of sepsis. In this presentation we intend to show how clinical practice changed over time.

Methods

Seventeen units came together in this project. It lasted from 1 December 2004 until 30 November 2005. Data collection included epidemiological characteristics and comorbidities, the CAS episode (locale of infection, responsible organism, first intention antibiotherapy and associated organ dysfunction) and the compliance with the SSC bundles and recommendations – following a detailed protocol that contained a summarized description of all SSC guidelines. We consider for this poster the first 10 months of the study and compare the compliance of the SSC bundles in the first month with the last 2 months of this period (December 2004/January 2005 vs August/September 2005). The 95% confidence interval (95% CI) for the difference of proportions and Fisher's exact test were used to analyse categorical data; P < 0.05 was considered statistically significant.

Results

During this period, 2643 patients were included in the study and 606 had CAS (23%) – of those, 520 (20%) had severe sepsis/septic shock. Over time more patients with septic shock received inotropes (45% vs 88%, 95% CI 0.149–0.701, P 49), had SvcO₂ measured (8% vs 32%, 95% CI 0.073–0.412, P < 0.01) and low-dose corticoids administered (38% vs 70%, 95% CI 0.123–0.510, P < 0.01). No significant difference was seen for CVP measurement (57% vs 47%, 95% CI –0.299 to –0.105, P = 0.408) or administration of vasopressors (78% vs 95%, 95% CI 0.035–0.307, P = 0.102).

In the last 2 months more patients had specimens collected for microbiological studies before antibiotic administration (72% vs 85%, 95% CI 0.010–0.250, P = 0.067) and more blood cultures done (89% vs 97%, 95% CI 0.010–0.142, P = 0.1). No significant differences were found in the first 6 hours of hospital admission for: measurement of serum lactate (61% vs 64%, 95% CI –0.121 to 0.186, P = 0.74), administration of fluids (67% vs 67%, 95% CI –0.145 to 0.158, P = 1.0) and antimicrobial administration (47% vs 49%, 95% CI –0.140 to 0.188, P = 0.867). No significant difference was seen in mortality in both periods.

Discussion

During the study period our performance improved in the compliance of the majority of the bundles. We need to feedback to the participating ICUs the weak points of their practice on the SSC recommendations implementation as part of the audit process.