Rodriguez-Nava et al. present a proof-of-concept study evaluating the use of a secure large language model (LLM) approved for healthcare data for retrospective identification of a specific healthcare-associated infection (HAI)—central line-associated bloodstream infections—from real patient data for the purposes of surveillance.1 This study illustrates a promising direction for how LLMs can, at a minimum, semi-automate or streamline HAI surveillance activities.

Background: Most healthcare facilities in the US apply contact precautions (CP) for patients with methicillin-resistant Staphylococcus aureus (MRSA) or vancomycin-resistant enterococci (VRE) infection and/or colonization. Most individuals with MRSA or VRE colonization will clear over time; however, frontline clinicians rarely evaluate for discontinuation of CP, resulting in increased burden on infection preventionists (IPs). Automation of time- and test-based evaluation using clinical decision support systems (CDSS) embedded in electronic health records (EHR) may increase evaluation and discontinuation of CP when appropriate, while preserving IP resources. Methods: This quality improvement initiative was implemented at Mass General Brigham (MGB), an integrated healthcare system, where patients with MRSA or VRE infection/colonization are identified in the EHR with a corresponding “infection status” and CP applied. Following MGB policy (Figure 1), CDSS features included: 1) automated time-based resolution from 2/15/2023-11/13/2023 and 2) automated ordering of screening assays for patients eligible for test-based evaluation from 6/20/2023-11/14/2023 (Figure 2). Counts of CP discontinuation and automated ordering were performed. IPs at one MGB facility performing manual review of patients self-recorded the time spent evaluating for CP discontinuation. Using these time reports, the average time to complete these tasks and the projected time savings were calculated over the implementation period. Results: Four IPs recorded the time to review patients for CP discontinuation, including reviewing recent antimicrobial administration, microbiology results, ordering screening test(s), and contacting the primary team. Twenty-five patient encounters were timed with a mean of 4.7 minutes documented per encounter. Over a 9-month period after initiation of the automated time-based resolution, the monthly mean number of patients with CP for MRSA and VRE which were automatically discontinued was 247 and 100, respectively. Projected IP time savings over the same 9-month period for MRSA and VRE were 174.1 and 70.5 hours, respectively. Over a 5-month period after initiation of automated ordering of MRSA polymerase chain reaction (PCR)/culture, as well as VRE culture for test-based evaluation, the monthly mean number of MRSA culture, MRSA PCR, and VRE culture automatically ordered for patients on CP for MRSA and VRE were 176, 24, and 145, respectively. Projected IP time savings over the same 5-month period for MRSA and VRE were 78.3 and 56.8 hours, respectively.

Conclusion: Healthcare systems that enhance their EHR with CDSS to automate CP evaluations may improve frontline clinician workflow, patient flow and bed capacity, while optimizing use of IP resources.

The COVID-19 has had major direct (e.g., deaths) and indirect (e.g., social inequities) effects in the United States. While the public health response to the epidemic featured some important successes (e.g., universal masking ,and rapid development and approval of vaccines and therapeutics), there were systemic failures (e.g., inadequate public health infrastructure) that overshadowed these successes. Key deficiency in the U.S. response were shortages of personal protective equipment (PPE) and supply chain deficiencies. Recommendations are provided for mitigating supply shortages and supply chain failures in healthcare settings in future pandemics. Some key recommendations for preventing shortages of essential components of infection control and prevention include increasing the stockpile of PPE in the U.S. National Strategic Stockpile, increased transparency of the Stockpile, invoking the Defense Production Act at an early stage, and rapid review and authorization by FDA/EPA/OSHA of non-U.S. approved products. Recommendations are also provided for mitigating shortages of diagnostic testing, medications and medical equipment.

Throughout the COVID-19 pandemic, many areas in the United States experienced healthcare personnel (HCP) shortages tied to a variety of factors. Infection prevention programs, in particular, faced increasing workload demands with little opportunity to delegate tasks to others without specific infectious diseases or infection control expertise. Shortages of clinicians providing inpatient care to critically ill patients during the early phase of the pandemic were multifactorial, largely attributed to increasing demands on hospitals to provide care to patients hospitalized with COVID-19 and furloughs.1 HCP shortages and challenges during later surges, including the Omicron variant-associated surges, were largely attributed to HCP infections and associated work restrictions during isolation periods and the need to care for family members, particularly children, with COVID-19. Additionally, the detrimental physical and mental health impact of COVID-19 on HCP has led to attrition, which further exacerbates shortages.2 Demands increased in post-acute and long-term care (PALTC) settings, which already faced critical staffing challenges difficulty with recruitment, and high rates of turnover. Although individual healthcare organizations and state and federal governments have taken actions to mitigate recurring shortages, additional work and innovation are needed to develop longer-term solutions to improve healthcare workforce resiliency. The critical role of those with specialized training in infection prevention, including healthcare epidemiologists, was well-demonstrated in pandemic preparedness and response. The COVID-19 pandemic underscored the need to support growth in these fields.3 This commentary outlines the need to develop the US healthcare workforce in preparation for future pandemics.

Throughout history, pandemics and their aftereffects have spurred society to make substantial improvements in healthcare. After the Black Death in 14th century Europe, changes were made to elevate standards of care and nutrition that resulted in improved life expectancy.1 The 1918 influenza pandemic spurred a movement that emphasized public health surveillance and detection of future outbreaks and eventually led to the creation of the World Health Organization Global Influenza Surveillance Network.2 In the present, the COVID-19 pandemic exposed many of the pre-existing problems within the US healthcare system, which included (1) a lack of capacity to manage a large influx of contagious patients while simultaneously maintaining routine and emergency care to non-COVID patients; (2) a “just in time” supply network that led to shortages and competition among hospitals, nursing homes, and other care sites for essential supplies; and (3) longstanding inequities in the distribution of healthcare and the healthcare workforce. The decades-long shift from domestic manufacturing to a reliance on global supply chains has compounded ongoing gaps in preparedness for supplies such as personal protective equipment and ventilators. Inequities in racial and socioeconomic outcomes highlighted during the pandemic have accelerated the call to focus on diversity, equity, and inclusion (DEI) within our communities. The pandemic accelerated cooperation between government entities and the healthcare system, resulting in swift implementation of mitigation measures, new therapies and vaccinations at unprecedented speeds, despite our fragmented healthcare delivery system and political divisions. Still, widespread misinformation or disinformation and political divisions contributed to eroded trust in the public health system and prevented an even uptake of mitigation measures, vaccines and therapeutics, impeding our ability to contain the spread of the virus in this country.3 Ultimately, the lessons of COVID-19 illustrate the need to better prepare for the next pandemic. Rising microbial resistance, emerging and re-emerging pathogens, increased globalization, an aging population, and climate change are all factors that increase the likelihood of another pandemic.4

The Society for Healthcare Epidemiology in America (SHEA) strongly supports modernization of data collection processes and the creation of publicly available data repositories that include a wide variety of data elements and mechanisms for securely storing both cleaned and uncleaned data sets that can be curated as clinical and research needs arise. These elements can be used for clinical research and quality monitoring and to evaluate the impacts of different policies on different outcomes. Achieving these goals will require dedicated, sustained and long-term funding to support data science teams and the creation of central data repositories that include data sets that can be “linked” via a variety of different mechanisms and also data sets that include institutional and state and local policies and procedures. A team-based approach to data science is strongly encouraged and supported to achieve the goal of a sustainable, adaptable national shared data resource.

A clinical decision support system, EvalMpox, was developed to apply person under investigation (PUI) criteria for patients presenting with rash and to recommend testing for PUIs. Of 668 patients evaluated, an EvalMpox recommendation for testing had a positive predictive value of 35% and a negative predictive value of 99% for a positive mpox test.

The coronavirus disease 2019 (COVID-19) pandemic highlighted the lack of agreement regarding the definition of aerosol-generating procedures and potential risk to healthcare personnel. We convened a group of Massachusetts healthcare epidemiologists to develop consensus through expert opinion in an area where broader guidance was lacking at the time.

Current methods of emergency-room–based syndromic surveillance were insufficient to detect early community spread of severe acute respiratory coronavirus virus 2 (SARS-CoV-2) in the United States, which slowed the infection prevention and control response to the novel pathogen. Emerging technologies and automated infection surveillance have the potential to improve upon current practice standards and to revolutionize the practice of infection detection, prevention and control both inside and outside of healthcare settings. Genomics, natural language processing, and machine learning can be leveraged to improve identification of transmission events and aid and evaluate outbreak response. In the near future, automated infection detection strategies can be used to advance a true “Learning Healthcare System” that will support near–real-time quality improvement efforts and advance the scientific basis for the practice of infection control.

Infection surveillance is one of the cornerstones of infection prevention and control. Measurement of process metrics and clinical outcomes, such as detection of healthcare-associated infections (HAIs), can be used to support continuous quality improvement. HAI metrics are reported as part of the CMS Hospital-Acquired Conditions Program, and they influence facility reputation and financial outcomes.

The rich and complex electronic health record presents promise for expanding infection detection beyond currently covered settings of care. Here, we review the “how to” of leveraging electronic data sources to expand surveillance to settings of care and infections that have not been the traditional purview of the National Healthcare Safety Network (NHSN), including a discussion of creation of objective and reproducible infection surveillance definitions. In pursuit of a ‘fully automated’ system, we also examine the promises and pitfalls of leveraging unstructured, free-text data to support infection prevention activities and emerging technological advances that will likely affect the practice of automated infection surveillance. Finally, barriers to achieving a completely ‘automated’ infection detection system and challenges with intra- and interfacility reliability and missing data are discussed.