SAN DIEGO – More than 16% of emergency department sepsis patients are not admitted to the hospital, preliminary results from a large retrospective cohort study found.
“Nothing is really known about this topic,” lead study author, said in an interview at an international conference of the American Thoracic Society. “In previous research, we’ve been focused on patients with sepsis who are admitted to the hospital. We have never thoroughly recognized that a fair number of patients who meet clinical criteria for sepsis in the emergency department are actually triaged to outpatient management. We don’t really know anything about these patients. What are their clinical characteristics and what are their outcomes like? And what are the factors that are leading them to be discharged from the ED rather than be admitted to the hospital?”
To find out, he and his associates retrospectively reviewed the medical records of 12,002 adult ED patients who met criteria for sepsis at two tertiary hospitals and two community hospitals in Utah between July 2013 and December 2016. They excluded trauma patients, those who left the ED against medical advice, those who were discharged to hospice or who died in the ED, and eligible patients’ repeat ED encounters. Patients transferred to another acute care facility were considered admitted, while transfers to non-acute care such as skilled nursing or psychiatric facilities were classified as discharges. Next, Dr. Peltan and his associates employed inverse probability weights using a propensity score for ED discharge based on age, sex, Charlson score, ED acuity score, initial ED vital signs, white blood cell count, lactate, sequential organ failure assessment ()score, busyness of the ED, and study hospital to compare 30-day mortality between patients admitted to the hospital versus those discharged from the ED.
Of the 12,002 patients included in the analysis, 10,032 (83.6%) were admitted, while 1,970 (16.4%) were discharged. Compared with admitted patients, discharged patients were younger (a mean of 53 vs. 60 years, respectively; P less than .001); more likely to be female (65% vs. 55%; P less than .001); more likely to be nonwhite or Hispanic (21% vs 17%; P less than .001), and had fewer comorbidities and physiologic derangements. In addition, crude mortality at 30 days was lower in discharged versus admitted patients (1.0% vs. 6.2%, respectively; P less than .001). After the propensity-adjusted analysis, there was no significant difference in 30-day mortality for discharged vs. admitted sepsis patients (adjusted odds ratio 1.0).
“We were worried that discharged ED sepsis patients were being mismanaged and weren’t going to do well as similar patients who were admitted to the hospital,” Dr. Peltan said. “This analysis is still a work in progress, but with that caveat, our findings so far suggest that physicians are making pretty good decisions overall.”
The researchers also found that, among 89 ED physicians who cared for 20 or more eligible patients, some did not discharge any of their sepsis patients, while others discharged 39% of their sepsis patients. “That was surprising,” Dr. Peltan said. “This could mean that some hospital sepsis admissions depend on physician practice style more than the patient’s condition or treatment needs.”