By Carol A. Kemper, MD, FACP

Sepsis Bundle Erodes Gains in Stewardship

SOURCE: Pakyz AL, Orndahl CM, Johns A, et al. Impact of the Centers for Medicare and Medicaid Services Sepsis Core Measure on antibiotic use. Clin Infect Dis 2021;72:556-565.

Beginning in October 2015, the Centers for Medicare and Medicaid Services (CMS) implemented a new policy for sepsis patients with initiation of a core measures sepsis “bundle,” an important part of which was time to first antibiotic dose following a diagnosis of sepsis and the initiation of broad-spectrum antibiotics within three hours of diagnosis.

These authors assessed monthly adult antibiotic usage for 111 acute care hospitals in the one year before this policy change (October 2014 to September 2015) and following this policy change (October 2015 to June 2017). Four different categories of antibiotic use were observed, including broad-spectrum antibiotics for community-acquired infection, broad-spectrum antibiotics for nosocomial infection/multidrug-resistant organisms, the use of anti-methicillin-resistant Staphylococcus aureus (MRSA) agents (such as vancomycin, daptomycin, linezolid, ceftaroline, etc.), as well as antibiotics for surgical prophylaxis.

Prior to implementation of the CMS sepsis bundle, antimicrobial usage had been diminishing gradually. Beginning in October 2015, however, an immediate 88.9% increase was observed in the overall use of antibiotics, with observed increases in all four categories. Broad-spectrum drugs for multidrug-resistant organisms and MRSA increased > 65%. Unexpectedly, antibiotic use for surgical prophylaxis significantly increased 284%. Use continued to increase incrementally every month for the next three years.

The rush to administer antibiotics within three hours of sepsis diagnosis has led to less discriminate use of antibiotics. The bundle itself presents only certain antibiotic options, and some agents are not even listed. It is estimated that at least one-third of sepsis diagnoses are due to noninfectious causes, such as acute pancreatitis, acute hypersensitivity reactions, and diabetic ketoacidosis, for which antibiotics may not be necessary. This is because sepsis is a physiological state, which may not be caused by an infection at all. And yet, as long as the hospital has satisfied their sepsis bundle goal, no one seems to care that CMS policy is directly counter to antimicrobial stewardship goals and the critical reduction of antimicrobial resistance. Notably, credit for the bundle is given in an “all or none” manner, meaning even if you sagely decide to implement everything in the bundle except the (perhaps unnecessary) antibiotics, you get zero credit for thinking.

Studies examining the use of antibacterials have observed that, once antibiotics are started in the emergency department (ED), too often they are continued for complex reasons. First, antibiotics begun in the ED often are not immediately re-evaluated by the admitting physician, who may be reluctant to disrupt the initial choice of antibiotics. Second, a gap of up to 24 hours may occur before the physician picking up the case the next day begins to tackle the process of re-evaluating the initial choice of antibiotic. Third, a gap of 24-72 hours occurs before micro data are available, and physicians may be reluctant to stop antibiotics without those data. How many times have I heard “wait at least 48 to 72 hours for negative cultures before stopping antibiotics.”

Medical teams, especially in teaching hospitals, function in a more collaborative approach (meaning that no one person “owns” the decision) and often take the path of least resistance, waiting for clinical outcomes and specific micro data before making decisions. Surgical teams, which round early in the morning when neither micro data nor antimicrobial stewardship pharmacists are available, are poorly prepared to make decisions about antibiotics during rapid-fire morning rounds. Rather, “lesser” decisions (such as antibiotics) are delegated to a junior member of the surgical team, who is more worried about negative outcomes and making the wrong decision. So, difficult decisions about antibiotics often are deferred, again leading to more prolonged use of antibiotics. 

Asymptomatic Transmission of COVID-19 in Households

SOURCE: Ng OT, Marimuthu K, Koh V, et al. SARS-CoV-2 seroprevalence and transmission risk factors among high-risk close contacts: A retrospective cohort study. Lancet 2021;21:333-343.

Early on during the COVID-19 pandemic, Singapore adopted a comprehensive approach to prevention, diagnosis, and management of COVID-19, with clear guidance for the community and recommendations for mask wearing and social distancing. A network of 800 public health and community clinics was activated to quickly perform contact tracing and quarantine of contacts in the home and in the workplace and to test contacts who developed symptoms, with a medical leave plan for those who became ill. From Jan. 23 to April 3, 2020, 13,026 close contacts were identified, including 1,863 household contacts (with 578 distinct contact groups), 2,319 work contacts (with 225 distinct contact groups), 3,588 social contacts (with 346 distinct contact groups), 2,626 transportation contacts, and 2,630 other contacts.

Of these, a total of 468 (26.3%) household contacts, 332 (14.9%) work contacts, and 458 (13.1%) social contacts were polymerase chain reaction (PCR) tested based on the presence of symptoms. A total of 188 cases were identified as secondary cases based on symptom-driven PCR screening, and another 7,582 completed quarantine without a COVID-19 diagnosis. Based on symptom-based screening, the secondary attack rate was 5.9% for household contacts, 1.3% for work contacts, and 1.3% for social contacts. Cases clearly clustered together within certain households and a few work groups; 86.3% of household contact groups and 91.6% of work contact groups had no apparent secondary cases based on symptom-based PCR screening.

Convalescent serologic testing was performed in 30% of household contacts, 9% of work contacts, and 11.8% of social contacts who completed quarantine without a COVID-19 diagnosis. An additional 5.5% of household contacts, 2.5% of work contacts, and 2.1% of social contacts were identified as secondary cases based on positive serology. Among these, two-thirds were asymptomatic and one-third had developed symptoms but had tested SARS-CoV-2 PCR negative.

Activities that increased the risk of transmission included sharing a bedroom, sharing a vehicle, or being spoken to by a COVID-19 PCR-positive person for > 30 minutes. Indirect contact, sharing objects or equipment, sharing a bathroom, and sharing a meal were not associated with SARS-CoV-2 transmission.

Secondary transmission was much more likely to individuals within households (11.4%) than in the workplace (3.8%) or social situations (3.4%), and infections clearly clustered in some households and a few workplaces. Efforts to control secondary transmission should be given to households and those contact groups where any case of secondary transmission has already been identified. Symptom-based PCR screening of contacts missed nearly half of those who developed secondary infection. At least one-third of secondary transmission cases remained asymptomatic. 

When Is Hand Hygiene Personally Protective?

SOURCE: Chang NN, Reisinger HS, Schweizer ML, et al. Hand hygiene compliance at critical points of care. Clin Infect Dis 2021;72:814-820.

Hand hygiene remains the most effective method for preventing healthcare-associated infections (HAI). Healthcare data often focus on hand hygiene on entry to or exit from a patient room and not on healthcare worker (HCW) behavior between patient care tasks. These authors used secondary data analysis from the STAR*ICU trial to examine HCW behavior during the process of patient care, breaking down the care into “care sequences” and identifying “task pairs” two consecutive tasks and the intervening hand hygiene opportunity.

Patient care tasks were categorized as either non-contaminating or contaminating, meaning more likely to contaminate HCW hands. Non-contaminating tasks included sterile tasks, device-blood (opening, connecting, injecting, etc., using an intravascular device and not in contact with mucous membrane or nonintact skin), device-other, patient (touching patient skin or a closed wound), and environment (items or surfaces in the environment). In contrast, contaminating tasks included blood or body fluid (with the potential for exposure to blood or body fluid), contaminated respiratory tract, oral, nose, or eye care; contaminated urinary catheter care, contaminated wound or wound drain care, and contaminated elimination. Tasks were defined further as critical if they were more likely to be associated with a higher risk of patient infection. Some tasks could be both contaminating and critical, depending on their position in the task pair. Multiple logistic regression analysis using repeated measures was used to examine associations between hand hygiene compliance, the type of patient care task, the order of the task, and the workload. The data were adjusted for HCW type and whether the patient was in standard, contact, or airborne isolation.

The study identified 28,826 task sequences with 42,349 hand hygiene opportunities. Critical tasks occurred significantly more often for patients in isolation than those in standard precautions. Overall, hand hygiene compliance was 43.2% before critical tasks and 38.1% before non-critical tasks. However, after adjusting for HCW type, glove use, and isolation precautions, HCWs were slightly less likely to perform hand hygiene before critical tasks compared with other tasks (adjusted odds ratio [OR], 0.97). Overall hand hygiene was 62.7% after contaminated tasks and 35% after other tasks (adjusted OR, 1.12).

HCWs tended to move from tasks that had relatively lower risk to patients to those tasks with higher risks for patients, rather than vice versa. However, they were less likely to perform hand hygiene when moving from tasks with lower risk to patients to those with higher risk, even when those subsequent tasks were considered critical. Hand hygiene decreased with increasing workload. An increase in workload was associated with increased odds of performing critical tasks. Hand hygiene also was performed more often when the patient was in isolation than during standard precautions. Nurses provided both critical and contaminating tasks more than physicians or other HCWs, and their hand hygiene generally was better than that of either physicians or other HCWs. Individuals who performed hand hygiene were more likely to practice it consistently throughout the process of patient care. In other words, for certain individuals, hand hygiene has become embedded in their routine.

The authors concluded that even though HCWs were more likely to order their work from less critical to more critical tasks, they appeared to perceive this approach as posing less of a risk to patients, rather than starting with the most critical tasks for the patient and moving to less critical tasks. This observation suggests that HCWs do not understand the rationale for hand hygiene as intended to prevent HAI, having more to do with their preferred work flow and less to do to with minimizing risks to patients. All HCWs were significantly more likely to perform hand hygiene after contaminating tasks, suggesting that they were more concerned about contaminating themselves.

The emphasis on personal protective equipment appears to have conveyed the wrong message to HCWs, who misunderstand the purpose of hand hygiene as protective to patients. HCWs choose to perform hand hygiene more often following a task than before a task, and when performing contaminating tasks than at critical moments during patient care. Hand hygiene was performed more often when the patient was in isolation, although hand hygiene and HAI are just as important for patients in standard precautions.