Can Sepsis Be Detected in the Nursing Home Prior to the Need for Hospital Transfer?

OBJECTIVES: To determine whether and to what extent simple screening tools might identify nursing home (NH) residents who are at high risk of becoming septic.

DESIGN: Retrospective chart audit of all residents who had been hospitalized and returned to participating NHs during the study period.

SETTING AND PARTICIPANTS: A total of 236 NH residents, 59 of whom returned from hospitals with a diagnosis of sepsis and 177 who had nonsepsis discharge diagnoses, from 31 community NHs that are typical of US nursing homes overall.

MEASURES: NH documentation of vital signs, mental status change, and medical provider visits 0-12 and 13-72 hours prior to the hospitalization. The specificity and sensitivity of 5 screening tools were evaluated for their ability to detect residents with incipient sepsis during 0-12 and 13-72 hours prior to hospitalization: The Systemic Inflammatory Response Syndrome criteria, the quick Sequential Organ Failure Assessment (SOFA), the 100-100-100 Early Detection Tool, and temperature thresholds of 99.0°F and 100.2°F. In addition, to validate the hospital diagnosis of sepsis, hospital discharge records in the NHs were audited to calculate SOFA scores.

RESULTS: Documentation of 1 or more vital signs was absent in 26%-34% of cases. Among persons with complete vital sign documentation, during the 12 hours prior to hospitalization, the most sensitive screening tools were the 100-100-100 Criteria (79%) and an oral temperature >99.0°F (51%); and the most specific tools being a temperature >100.2°F (93%), the quick SOFA (88%), the Systemic Inflammatory Response Syndrome criteria (86%), and a temperature >99.0°F (85%). Many SOFA data points were missing from the record; in spite of this, 65% of cases met criteria for sepsis.

CONCLUSIONS: NHs need better systems to monitor NH residents whose status is changing, and to present that information to medical providers in real time, either through rapid medical response programs or telemetry.