Systematic Sepsis Screening Shows Efficacy
Major Finding: Systematic, twice-daily sepsis screening of postoperative patients using the SIRS score had a 16.3% positive predictive value and a 99.9% negative predictive value.
Data Source: Single-center study of 959 patients hospitalized during May-October 2009.
Disclosures: Dr. Moore and her associates had no relevant disclosures.
LAS VEGAS — Systematic, twice-daily assessment of postsurgical patients by registered nurses using an inflammatory screening tool ruled out sepsis in some patients and was modestly effective for identifying sepsis in others, in a single-center study of 959 patients.
Regular screening of patients for sepsis using the systemic inflammatory response syndrome (SIRS) method “is a sustainable process” that has now been used at the Methodist Hospital, Houston, for 3 years, Dr. Laura J. Moore said.
The screening approach, which uses three sequential evaluation steps by three different staff members, was initially applied to patients in the surgical intensive care unit before it was rolled out to non-ICU, postoperative patients on the surgical floor. The Methodist staff now uses the screening tool in two additional wards, said Dr. Moore, a surgeon and medical directorof the wound care servicect Methodist.
The initial step of the screening scheme takes about 30 seconds. An additional 5–7 minutes is needed to evaluate patients who score positive on the screen and require further assessment. “If you can prevent needless transfer of patients to the ICU, it's a tremendous benefit to the hospital and to patients,” she said. In addition, systematic, routine screening aids in the early identification of sepsis, which is otherwise problematic because the early signs—such as oliguria, hypothermia, and altered mental status—are nonspecific. “When sepsis is missed, it delays the start of treatment,” she noted.
The three-part screening tool starts with a nurse calculating a disease severity score for each patient twice a day using a detailed version of the SIRS score. Like conventional SIRS scores, the Methodist tool rates four clinical parameters: heart rate, temperature, respiratory rate, and white blood cell count. Each criterion is scored on a scale of 0–4, which means a patient's overall score ranges from 0 to 16. (See box.) The screening cutoff used by the researchers flags patients who score 4 or higher for further sepsis assessment.
In step two, a nurse practitioner or a resident evaluates the patient for six different possible sites of infection: vascular access, pulmonary, abdominal, skin and soft tissue, urinary tract, or other. Finally, an attending physician assesses patients with indications of a site-specific infection for a definitive diagnosis of sepsis.
This scheme was first validated in a study of 920 surgical ICU patients hospitalized at Methodist during 2007. Among these patients, who had a sepsis prevalence of 12.2%, the three-step screen had a sensitivity of 96.5%, specificity of 96.7%, positive predictive value of 80.2%, and negative predictive value of 99.5%. During the period of screen use in 2007, sepsis-related mortality in the ICU patients fell to 23.3%, down from 35.1% in 2006 (J. Trauma 2009;66:1539–46). In contrast, there was no change in mortality during 2007 compared with 2006 in the emergency department, cardiac ICU, or medical ICU units, where the sepsis screen was not used.
Dr. Moore and her associates introduced the screen onto the surgical floor in May 2009, and during the first 6 months they applied it to 959 patients. The average age of the patients was 57 years, and about two-thirds were women. The sepsis prevalence was 1.7%.
A total of 55 patients had a score of 4 or higher on the initial screening tool; 16 of these had sepsis, including 2 with severe sepsis. One patient died from sepsis. The results meant a sensitivity of 99.9%, specificity of 91.3%, positive predictive value of 16.3%, and negative predictive value of 99.9%.
Vitals
Source Elsevier Global Medical News
My Take
Simple, Practical Tool Needs More Work
This is a novel, interdisciplinary screening routine that involves the nursing staff. I am seduced by the simple and practical application of this tool. It excludes sepsis extremely well, with a negative predictive value of 99.9%. But the numbers suggest there is more work to be done. In the low-incidence population assessed, the screen provided only a small improvement over guessing every time that patients had no sepsis. Guessing no for all these patients would have been correct 98.3% of the time because the actual sepsis incidence was 1.7%.
SAMIR M. FAKHRY, M.D., is professor and chief of general surgery at the Medical University of South Carolina, Charleston. Dr. Fakhry made his remarks as the designated discussant of the paper. He reported no disclosures.
