Critical Care |

Analysis of an Automated Sepsis Screening Alert at a Tertiary Cancer Center FREE TO VIEW

Amit Pandit, MD; Kaye Hale, MD; Tess Pottinger; Dona Bugov, MD; Sherard Lacaille, MBBS; Sarah Rebal, ACNP; Stephen Pastores, MD; Neil Halpern, MD
Author and Funding Information

Memorial Sloan Kettering Cancer Center, New York, NY

Chest. 2015;148(4_MeetingAbstracts):342A. doi:10.1378/chest.2274431
Text Size: A A A
Published online


SESSION TITLE: Sepsis and Shock Posters

SESSION TYPE: Original Investigation Poster

PRESENTED ON: Wednesday, October 28, 2015 at 01:30 PM - 02:30 PM

PURPOSE: Practice guidelines recommend routine screening and protocolized care for potentially septic patients. Automated sepsis screening alert systems have been used to identify patients at risk for progression to severe sepsis and septic shock. The objective of our study was to examine the sepsis screening alert characteristics and their effect on escalation of care and in-hospital all-cause mortality.

METHODS: We retrospectively identified 236 cases by ICD-9 (995.92 or 785.52) who were admitted to a tertiary care cancer center from April to August 2014. The automated sepsis alert consisted of 3 or more abnormal vital signs or systemic inflammatory response syndrome (SIRS) criteria (fever, tachycardia, tachypnea, hypotension, hypoxemia, altered mental status, and rigors). Cases were reviewed for presence of an automated alert, rapid response team evaluations (RRT), critical care medicine (CCM) consults, ICU admissions and hospital discharge status (alive or dead). 570 non-septic patients that alerted during the same time period were used for specificity calculation and construction of a receiver operating characteristic curve (ROC).

RESULTS: 102/236 sepsis cases alerted for a sensitivity of 43% while 570/11022 non-septic cases alerted providing a specificity of 95% (AUC 0.69). 34% of alerted patients were escalated to RRT vs. 16% of non-alerted patients (OR 2.65, p<0.01). There was no difference between the frequency of escalation to CCM consult or ICU admission between the two groups. 43 non-alerted patients were escalated to CCM, most commonly due to isolated hypotension (51%). 59% of alerted patients died during hospitalization vs. 25% of non-alerted patients (OR 4.37, p<0.01).

CONCLUSIONS: Despite the low sensitivity, the automated sepsis alert was successful in identifying a population of patients at higher risk of death and in need of a higher level of care.

CLINICAL IMPLICATIONS: Our findings are in agreement with a recently published large retrospective study that showed limited sensitivity for using two or more SIRS criteria for the diagnosis of patients with severe sepsis. New and refined diagnostic criteria may improve the detection and early recognition of this high-risk population.

DISCLOSURE: The following authors have nothing to disclose: Amit Pandit, Kaye Hale, Tess Pottinger, Dona Bugov, Sherard Lacaille, Sarah Rebal, Stephen Pastores, Neil Halpern

No Product/Research Disclosure Information




Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Find Similar Articles
CHEST Journal Articles
PubMed Articles
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543