Better Surveillance for Serious HAIs

The CDC’s hospital-onset Adult Sepsis Event (HO-ASE) surveillance definition may do a better job at detecting serious hospital-associated infections (HAIs) than the criteria set by the Centers for Medicare & Medicaid Services (CMS).

Many HAIs, including central line–associated bloodstream infections, catheter-associated urinary tract infections, Clostridioides difficile infections, methicillin-resistant Staphylococcus aureus bacteremia and certain surgical site infections, are missed by CMS criteria, according to a retrospective analysis of 282,441 patients hospitalized from June 2015 to June 2018 at three hospitals. The analysis showed that almost twice as many patients met the HO-ASE criteria as had reportable HAIs (0.8% vs. 0.4%), and in-hospital mortality rates were more than twice as high for HO-ASEs (28.6% vs. 12.9%), according to Chanu Rhee, MD, MPH, a researcher at Harvard Medical School and Harvard Pilgrim Health Care Institute, and an infectious disease physician and associate hospital epidemiologist at Brigham and Women’s Hospital, in Boston (Clin Infect Dis 2021 Mar 29. doi:10.1093/cid/ciab217).

Surveillance for HO-ASEs is relatively new and voluntary for hospitals, Dr. Rhee said. It stemmed from previous work that he and his colleagues did developing and validating a surveillance definition for sepsis based on objective clinical data that could be extracted from electronic health records.

The ASE definition requires clinical indicators of treated infection (blood culture orders and antibiotics) and concurrent organ dysfunction (vasopressors, mechanical ventilation and abnormal laboratory values) and distinguishes hospital-onset from community-onset ASE based on when those criteria are met relative to a patient’s admission.

“The HAIs that hospitals are required by CMS to publicly report are certainly important to track and prevent, but we knew that our current surveillance was likely missing many other important infections that occur in the hospital,” Dr. Rhee said. “About 10% to 20% of sepsis cases arise in-hospital and these tend to be very serious infections, and so we thought it would be interesting to look at the overlap between reportable HAIs and hospital-onset sepsis events.”

The in-hospital mortality rate for HO-ASEs missed by reportable HAIs was twice as high as the mortality rates for reportable HAIs missed by HO-ASEs (28.1% vs. 6.3%). Of note, Dr. Rhee said reportable HAIs were only present in 14.5% of HO-ASEs, which indicates that the other 85.5% are caused by other serious events that are not captured routinely and reported, including hospital-acquired pneumonia, non–catheter-related bloodstream infections, intraabdominal infections, and skin and soft tissue infections.

Jonathan Baghdadi, MD, PhD, of the University of Maryland School of Medicine, in Baltimore, who was not part of the study, said surveillance of HO-ASEs has value. “One important reason to have an objective measure of sepsis events is to facilitate comparison of outcomes among hospitals,” Dr. Baghdadi said. “This analysis demonstrates that the HO-ASE measure is sensitive to underlying differences in patient populations and may favor community hospitals over teaching hospitals or hospitals with large oncology practices.”

Specifically, the study showed that incidence of HO-ASEs and reportable HAIs were higher in the academic hospital (Brigham and Women’s) compared with the two community hospitals, but mortality rates were similar.

“We don’t think that means our academic hospital does worse when it comes to infection prevention,” Dr. Rhee said. “It is likely the different patient population we care for. For example, there are a larger number of oncology patients who, even with the best of care, are prone to developing serious infections when hospitalized.”

Dr. Baghdadi said if the HO-ASE measure were to be used, it would likely require risk adjustment to account for differences in the patient populations served at different hospitals.

Despite this value, Dr. Baghdadi said HO-ASEs should only be used to complement HAI surveillance. “It cannot and should not replace HAI surveillance,” he said, especially given the fact that only a small proportion of sepsis deaths are likely preventable. “I suspect that hospitals conducting HO-ASE surveillance would find that only a small proportion of events were actionable or led to identification of gaps in care.”

Dr. Baghdadi referred to Dr. Rhee’s paper looking at sepsis-associated mortality in U.S. acute care hospitals that examined 568 randomly selected adults who died in the hospital or were discharged to hospice (JAMA Netw Open 2021 Feb 2. doi:10.1001/jamanetworkopen.2018.7571). Sepsis was the immediate cause of death in 34.9% of deaths; however, only 3.7% of sepsis-associated deaths were judged definitely or moderately likely preventable with another 8.3% possibly preventable.

Instead, Dr. Baghdadi said he viewed the purpose of monitoring for adult sepsis events to support performance improvement related to recognition and early management of sepsis—for example, shortening the time from order to administration of broad-spectrum antibiotics or providing bedside evaluation to clinically deteriorating patients.

Dr. Rhee echoed this opinion, emphasizing that the preventability of these HO-ASE events is still unknown.

“We still have to determine if they are mostly just happening in really sick patients, in whom these serious infections are sometimes inevitable,” Dr. Rhee said. “If so, that could defeat the purpose of ultimately increasing prevention and quality improvement programs to lower rates, incidence and mortality rates of these HAIs.”

However, Dr. Rhee said he believes there is likely room to make progress in preventing HO-ASEs, but more research is needed. He also noted that HO-ASE surveillance can be fully automated, which eliminates the need for time-consuming and often subjective case reviews that are needed to identify reportable HAIs.

“I don’t think sepsis surveillance would be a replacement for the current reportable HAIs, but I do think there is value in this,” he said. “They could be complementary in the sense that HO-ASEs could capture a lot of serious hospital-onset infections that are currently being missed and, potentially, identify new targets for prevention.”

This article is from the August 2021 print issue.