Chapter 6—Healthcare outside the acute care hospital

One of the remaining frontiers in infection prevention and control (IPC) in high-income countries is the outpatient world. As care shifts from inpatient to ambulatory medicine (and day surgery), so do complications. We currently have inadequate evidence on how to prevent outpatient healthcare-associated infections (HAI) that have its root in a usually brief healthcare encounter. This includes central line-associated bloodstream infection in outpatient parenteral antibiotic therapy patients, catheter-associated urinary tract infection in those with long-term catheterization, surgical site infection (SSI) in outpatient surgery, but may also involve other populations such as patients receiving joint injections for osteoarthritis pain management. Of note, it is unclear how to best capture these outpatient HAIs, as traditional surveillance systems would require a significant effort. ^{Reference Rinke, Bundy and Heo1,Reference Pivot, Hoch and Astruc2} Possibly, modern surveillance by means of patient-operated smartphone applications, ICD-10 codes of follow-up encounters or claims data, could help in identifying HAI events in these outpatient populations.

Sectors that require more dedicated research are long-term care facilities (LTCF) and nursing homes (NH). Neglected by IPC activities due to lack of personnel and financial resources (e.g., fewer cultures taken because of the costly microbiology work-up), there are many improvements to be made in these settings. The link from LTCF/NH to acute care hospitals is particularly intriguing as there is a constant transfer of patients in both directions; if one side does not identify the MDRO-colonized patient, they may inadvertently trigger an outbreak on the other side. How to reach and train the LTCF/NH workforce in infection prevention and antimicrobial stewardship is a major challenge. ^{Reference Daker-White, Panagioti and Giles3}

Chapter 7—Low- and middle-income countries

Information that stems from studies conducted in high-income countries may not necessarily translate to LMIC settings, given that the epidemiology often differs and both infrastructure and resources can be limited. Accordingly, IPC strategies have not been tailored to LMIC as much, and its benefits have not reached all areas of the world equally. ^{Reference Tomczyk, Twyman and de Kraker4} This is problematic and should lead to a push to bring cutting-edge research to LMICs. Although there are funding opportunities dedicated to resource-limited settings, such as those issued by the NIH Fogarty International Center, IPC research remains heavily underfunded in comparison to other infectious diseases research. Not all research would necessarily be costly, nor are all preventive measures expensive. Some elements rely on the local work culture, which may be amenable to low-cost interventions in terms of addressing HCP behavior (but with the potential for high yield). The same goes for areas of conflict: a recent qualitative study on how to best implement measures cited low-cost interventions such as having IPC champions and offering illustrated guidelines to healthcare workers. ^{Reference Lowe, Woodd, Lange, Janjanin, Barnet and Graham5} Knowledge sharing and knowledge sharing networks are key in disseminating information and should be cultivated in LMIC, if they do not already exist. ^{Reference Rangachari6} Regarding COVID-19, there is a good overview of research deemed necessary to reduce its influence on LMIC. ^{Reference Polasek, Wazny and Adeloye7}

LMIC are often the first victims of outbreaks as humans encroach on remaining pockets of untouched nature. A recent example of an illness spreading from an endemic area in an LMIC setting to other continents is the 2022 mpox outbreak. In the past, mpox has usually been presented as an endemic curiosity in tropical medicine textbooks. Confined to a small geographic area until recently (with very few exceptions), it stands for the neglected diseases of LMIC settings that have the potential for widespread outbreaks. The implicit theme of “equality in access to the best possible care while ensuring patient safety” may be beyond the influence of IPC experts. A noteworthy publication in the wake of the COVID-19 pandemic suggested striving for equal access to care, soliciting greater solidarity among the nations and working toward universal preparedness for challenges. ^{Reference Tomson, Causevic and Ottersen8}

More research is needed on the contribution of anthropological and socioeconomic factors to the global antimicrobial resistance and how to counter this threat best. Collignon et al used bacterial resistance and antimicrobial consumption databases and correlated the data with World Bank indicators such as governance, education, gross domestic product per capita, healthcare spending, and community infrastructure (e.g., sanitation). They conclude that improving sanitation, increasing access to clean water, and ensuring good governance, as well as increasing public healthcare expenditure and better regulating the private health sector are all necessary to reduce global antimicrobial resistance. ^{Reference Collignon, Beggs, Walsh, Gandra and Laxminarayan9}

Lastly, in terms of global travel to and from LMIC, the optimal form of screening patients following travel is unclear and should be characterized better. The risk of carrying multidrug-resistant organisms is influenced by factors such as area of travel, level of exposure to the local healthcare system, and antibiotic receipt.

Chapter 8—Networking with the “neighbors”

Innovation is known to occur more readily at the interface of different areas, and specifically, when different thoughts collide. It is therefore important to be aware of the neighboring fields and identify research questions that can be answered by the collaboration between IPC and the respective partner.

An example from our own experience is a study into the effect of operation room ventilation on SSIs, for which we worked with a team of ventilation engineers. ^{Reference Surial, Atkinson and Kulpmann10} Owing to this collaboration, we were able to identify the need for a straightforward descriptor of ventilation quality in operating rooms. Very likely, this need would not have been recognized by either collaboration partner alone. As such, we would like to reframe IPC research as one set to benefit immensely from interdisciplinary work. Stimuli may come from (but are not limited to) statistics, data science, engineering, environmental sciences, psychology, economics, behavioral science, human factors, disinfection and sterilization, quality management, patient safety, anesthesiology, microbiology, information technology, and a wide array of specialty areas such as, for example, nutritional science (e.g., the effect of malnutrition on SSI risk ^{Reference Tsantes, Papadopoulos and Lytras11} and how to correct this). Each of these areas is evolving, and new developments can trigger research ideas that were not (or could not be) addressed before. For example, the broader availability of whole genome sequencing in partnership with microbiology laboratories has made outbreak investigations much more granular and now permit tracing entire evolutionary pathways.

Although research in IPC has traditionally come from research-heavy academic medical centers, there is a need to install more inclusive, larger research networks. These should include rural sites. Although large surveillance systems can provide representative data on a condition, including pre/post data in intervention studies, they are usually hampered by the limited number of available variables. We believe that more research networks should be formed to address questions with the appropriate statistical power (as for example, the CDC Prevention Epicenters program does, https://www.cdc.gov/hai/epicenters/index.html).

This discourse touches on how to position IPC as an expert group for complex systems, which is what contemporary healthcare is. Very few other professionals are so well connected with other players inside a healthcare organization. This also predisposes IPC teams to conduct research on safe processes that are the product of multidisciplinary work (for example, IPC might evaluate the microbial contamination of stem cell products, what consequences this has on a patient receiving transfusion, and if antibiotic prophylaxis is required). In that role, IPC should encourage and help everybody across a healthcare system who does QI work related to infection prevention to disseminate their results so that others outside that institution can benefit from novel insights.

From the perspective of funding mechanisms, IPC research could theoretically be funded by nearly every National Institutes of Health (NIH) center (and others entities such as, for example, the Agency for Healthcare Research and Quality, AHRQ, and the Patient-Centered Outcomes Research Institute, PCORI), as it affects all aspects of modern medicine. This seems attractive at first sight but makes navigating the grant landscape more difficult than if there was a straightforward match (e.g., the NIDDK institute for nephrology research). We hope that funding agencies are cognizant of the fact that there is no dedicated center for infection prevention (yet) and therefore no natural “home” for IPC research.

Chapter 9—Novel research methodologies

Not all research can be pursued by conducting randomized controlled trials; we anticipate that much evidence will come from time-series analyses, before/after studies, and other designs popularized in epidemiology research in recent years such as stepped wedge studies. ^{Reference Harris, Lautenbach and Perencevich12,Reference Hussey and Hughes13} Cluster-randomized cross-over trials are an elegant way around randomization of individual patients, which often is neither feasible nor sensible in IPC. In a recent example of such a trial groups of patients in five hospitals were given either 24 or 24–48 h of postoperative antibiotics, with group assignments switching every 2–4 mo (and prolonged antibiotics did not confer additional risk reduction in developing SSI). ^{Reference Nagata, Yamada and Shinozaki14} More studies should employ this design, and it is particularly well suited for multicenter trials.

Often, however, it will be difficult to set up trials, for one because they are costly and take extended times for planning and conducting them. Second, because the outcomes may be so infrequent that a trial becomes unattractive to even begin. Target trial emulation is a novel approach that relies on available observational data and then models a comparative study. Recent examples are on vaccine effectiveness against COVID-19. ^{Reference Dickerman, Gerlovin and Madenci15,Reference Ioannou, Bohnert and O’Hare16}

There are many other trial approaches that deserve to be inspected by IPC researchers, including designs such the “personalized randomized controlled trial” ^{Reference White and Turner17} or the “durations design”. ^{Reference Quartagno, Walker, Carpenter, Phillips and Parmar18} Another innovative study design is the Sequential Multiple Assignment Randomized Trial (SMART) that is suitable for IPC work where usually multiple contextual factors are at play and a bundle of preventive measures may be needed to achieve the desired outcome. ^{Reference Kidwell and Sequential19}

Qualitative research, with its roots in sociology and psychology, has the potential to elicit mental models of providers and patients and help us understand what drives our behavior. Video-reflexive ethnography is a special form of qualitative research that involves filming healthcare practices and examining the material to detect weak points. Human factors engineering is about conceiving an environment, tools, and processes around human beings to improve our “output.”

Lastly, data science and the first steps in artificial intelligence are visible in hospital epidemiology. A recent example involves outbreak investigation. ^{Reference Atkinson, Ellenberger and Piezzi20} Beyond the local setting, applications in global scenarios are within reach. ^{Reference Patel, Surti and Adnan21} As more routine medical data become available in dedicated data science centers in healthcare, innovative approaches for using and visualizing it should be employed. These data should be supplemented with sensor-generated information (such as on whether a handrub dispenser was used or not).

An extreme form of putting data to use is the mathematical modeling of pathogen transmission and HAI acquisition. ^{Reference Stachel, Keegan and Blumberg22} These stochastic models are usually grounded in real-world data and rely on multiple assumptions to produce output. They may help with estimating the impact of public health interventions, trial planning, projections of disease burden, and more.

Chapter 10—The future state of surveillance

Surveillance activities take time and effort from the IPC team, including manual chart review, data cleaning, reporting back to the surveilled units, and communication with public health authorities. Future research should inspect ways to make this process more efficient. Automated (or semi-automated) surveillance of HAI is a result of the move to electronic medical records (EMR). The potential time freed up for IPC personnel by transitioning to electronic surveillance is considerable, with an estimated 74% reduced workload. ^{Reference Russo, Shaban, Macbeth, Carter and Mitchell23} However, dedicated time from hospital informatics is quintessential and not always offered, algorithms for different EMR platforms have to be developed, and variables relevant for surveillance need to be identified. Once surveillance becomes less time consuming, IPC personnel can dedicate more time to improvement projects. A variation of this is to identify surrogate markers for HAI; for example, ICD-10 codes for postoperative infections could be utilized to determine where detailed surveillance is needed. Yet, approaches based on nonclinical data sources need to be critically reviewed before implementation. ^{Reference Marra, Alkatheri and Edmond24}

Another aspect of surveillance that deserves more attention is the early detection of outbreaks. This can be in the form of microbial species not otherwise seen (such as emerging infections), common species becoming more frequent, or certain resistance determinants increasing. Rapid diagnostic methods may help with surveillance, but platforms are heterogeneous and availability can be an issue. ^{Reference Cury, Almeida Junior and Costa25} Also, HAIs that are not subject to mandatory surveillance and public reporting may increase in frequency and go undetected for some time. Our tools for early detection are limited, and we are likely to notice only what we have decided to measure. Algorithms to detect upticks of infections need to be developed and promise to facilitate our work. ^{Reference Baker, Yokoe and Stelling26} Again, pattern recognition tools could incorporate EMR variables such as billing codes, microbiology lab trends, and free text crawling for terms suggestive of outbreaks, and they should explore artificial intelligence.

On an international scale, there are tools that allow early reporting (such as promedmail.org); however, it is difficult to identify the most relevant threats early on, given the noise of information. HAI and AMR data need to be synthesized across regions and countries so they can serve for benchmarking and as outcomes for large public health intervention studies.

Furthermore, surveillance should not be a purpose to itself (and surveillance definitions should not be too different from what is considered a clinical diagnosis of infection) but provide data that is fed back to the providers in a digestible way and then leads to improvements; as such, the mechanics of the feedback loop should be a topic of study. Surveillance should always be actionable.