3

Lessons Learned from the COVID-19 Pandemic

In this session of the workshop, speakers and participants considered the unique challenges for the development of rapid point-of-care diagnostics to address antibiotic resistance. Speakers shared lessons learned from other disease areas, including COVID-19, discussed the development and use of rapid diagnostics to address drug-resistant bacterial infections, and considered generalizable applications and practical approaches to overcome barriers to innovation.

LESSONS LEARNED FROM THE COVID-19 PANDEMIC

From a U.S. perspective, there are three key lessons learned during COVID-19 that are applicable to combating AMR, said Joseph Lutgring (Medical Officer, Division of Healthcare Quality Promotion, National Center for Emerging and Zoonotic Infectious Diseases, Centers for Disease Control and Prevention):

• Testing is not just for symptomatic patients;
• There is a need for more than one type of test for a given disease; and
• Testing is not just for people.

Relatively early in the pandemic, said Lutgring, it became clear that asymptomatic and presymptomatic people could transmit severe acute respiratory syndrome coronavirus 2 (SARS CoV-2), which had a major impact on testing approaches and recommendations. However, screening of asymptomatic people is only beneficial if there are interventions that can protect that person and/or prevent transmission to others. As an example, Lutgring noted if someone receives a positive COVID-19 test result, they could reduce the risk of transmission by self-isolating, wearing a face covering, or avoiding contact with high-risk patients. Over the course of the pandemic, testing of asymptomatic people was recommended for a number of populations, including nursing home residents, K-12 students and staff, hospitalized patients, and workers and residents in correctional facilities. Screening of asymptomatic individuals has also had applications in the context of drug-resistant bacterial infections, particularly for hospital-acquired infections, said Lutgring. For example, the Department of Veterans Affairs (VA) implemented a methicillin-resistant Staphylococcus

aureus (MRSA) nasal swab screening program for hospitals; hospitals in Israel have implemented active screening of CRE to reduce infections; and the Netherlands has a policy for active MRSA screening, which may account for the low rates of MRSA in that country (Health Council of the Netherlands, 2007; Schwaber and Carmeli, 2014; VA, 2007). Screening for other organisms—including Vancomycin-resistant enterococci (VRE) and bacteria that produce extended spectrum beta-lactamase (ESBL)—has been attempted, but Lutgring said the utility of such screening depends on whether there are interventions that can be implemented when a person tests positive. For multi-drug resistant organisms, the main interventions are contact precautions (e.g., providers wear gowns/gloves) or isolation. However, in the future there may be decolonization interventions that could reduce the organism burden of the individual and prevent transmission to others. Screening and interruption of transmission have the potential to reduce infection even more than the development of a new treatment. He pointed to a recent workshop hosted by the Centers for Disease Control and Prevention (CDC) and FDA to discuss the state of evidence supporting decolonization and pathogen reduction in colonized patients as a strategy to prevent infection.¹

According to Lutgring, the second key lesson learned from the COVID-19 pandemic is that more than one type of test is needed for a given disease. As of October 2022, the FDA had issued EUAs for a variety of SARS-CoV-2 tests, including almost 300 molecular tests, 51 antigen tests, and 85 serology tests (FDA, 2022). These tests, said Lutgring, vary in terms of performance characteristics, costs, turn-around times, specimen types, and can be utilized in different ways. For example, patients admitted to the VA hospital in Atlanta are tested using both an antigen and a molecular test; antigen results are faster while molecular tests are more sensitive. A combination of test results can be helpful to inform decisions about bed placement or whether a patient should be quarantined in an airborne isolation room. Clinicians have become savvier about using both types of tests for clinical decision-making, he said. For example, if a patient tests positive on both antigen and molecular tests, that patient likely has an active COVID-19 infection and should be treated. However, if a patient has a negative result on an antigen test but a positive result on a molecular test, more consideration may be given to whether or not that patient has an active COVID-19 infection requiring treatment. In the area of AMR, said Lutgring, a similar testing strategy can be used for Clostridioides difficile. There is debate about the best strategy for detecting C. diff, with some arguing for nucleic acid amplification tests (more sensitive) and others arguing

___________________

¹ See https://info.rescueagency.com/en-us/drug-development-consideration-virtual-public-workshop-cdc-fda (accessed January 25, 2023).

for toxin enzyme immunoassays (more specific). There are facilities that perform a combination of these tests, e.g., screening with the more sensitive molecular test and then performing the more sensitive test to give indications about disease severity, likelihood, or true current infection. Lutgring cautioned that there can be unintended consequences of using multiple diagnostics, particularly if clinicians do not understand what a positive or negative result means. For example, a patient with a positive nucleic acid amplification test and a negative toxin enzyme immunoassay test might be treated with antibiotics even if they do not have an active infection. Lutgring emphasized that as testing strategies become more complex and as more diagnostics become available, it will be important for clinicians to understand how results should be interpreted to best care for their patients.

Finally, Lutgring said, the COVID-19 pandemic has indicated that testing is not just for people. Early detection and containment of pathogens are crucial, but individual patient testing is time- and resource-intensive. The National Wastewater Surveillance System (NWSS)² was launched by the CDC in September 2020, to coordinate and build the nation’s capacity to track the presence of SARS-CoV-2 in wastewater samples. Information about the presence and level of SARS-CoV-2 in the wastewater combined with other community-level information can help track the impact of COVID-19 on a community. There is ongoing research about how to use this type of surveillance system for antimicrobial resistance detection, particularly for rare pathogens like carbapenemase-producing organisms. For example, a new CDC initiative, the Healthcare Wastewater Antimicrobial Resistance Network, will be conducting research on how to best use wastewater testing, with an initial focus on carbapenemase-producing organisms and Candida aureus.

Lessons Learned from the “Access to COVID-19 Tools Accelerator” Program

AMR, like COVID-19, is a global crisis, said William “Bill” Rodriguez (Chief Executive Officer, FIND, The Global Alliance for Diagnostics). FIND is aimed at supporting equitable access to diagnostics everywhere in the world, with a focus on diseases of poverty such as tuberculosis, malaria, hepatitis, and schistosomiasis. When SARS-CoV-2 emerged, FIND’s mandate and activities expanded dramatically, he said. The Access to COVID-19 Tools Accelerator (ACT Accelerator)³ initiative was designed by WHO to speed up efforts for development and deployment of tests, treatments, and vaccines; FIND co-led the diagnostics pillar of the initiative.

___________________

² See https://www.cdc.gov/nwss/index.html (accessed March 9, 2023).

³ See https://www.act-a.org/ (accessed January 25, 2023).

Before delving into the lessons learned from COVID-19, Rodriguez gave workshop participants a brief overview of the landscape of diagnostics around the globe. Almost half of the world’s population, he said, lacks access to essential diagnostics for diseases such as tuberculosis, HIV, hepatitis, diabetes, and hypertension (Fleming et al., 2021). Basic diagnostic capacity is available in only 1 percent of primary care clinics and 14 percent of hospitals in some low- and middle-income countries (Leslie et al., 2017). Further, 47 percent of the world’s population cannot receive a diagnosis when sick, and diagnostic gaps for diseases like diabetes, Hepatitis C, and HIV are significantly larger than treatment gaps, meaning treatments are available but not able to be deployed due to a lack of diagnoses (Fleming et al., 2021). The diagnostic gap is greatest at the primary care level, Rodriguez said, but not much better at most district hospitals and referral hospitals; for example, basic technologies such as tests for pregnancy and syphilis, microscopes, and X-rays are not readily available at most hospitals in some low- and middle-income countries (Leslie et al., 2017). Rodriguez urged workshop participants to keep this “stark environment” in mind when thinking about the global capacity to test for AMR. FIND divides diagnostics into three levels (Figure 3-1). Self-tests are those administered at home or in the community and do not require power, water, or lab equipment. True point-of-care tests are delivered in a primary care facility and do not require lab equipment or reliable power to be stored and administered. Near-patient point-of-care tests are conducted in district hospitals and may

require basic lab procedures such as centrifuges and biosafety containment measures. Rodriguez emphasized that when considering diagnostics—whether for COVID-19 or AMR pathogens—it is essential to meet patients where they are and where they present for care.

The COVID-19 pandemic put testing in the spotlight as it has never been before, said Rodriguez. Before the pandemic, most non-expert stakeholders he talked with were unfamiliar with tools such as PCR tests or rapid tests, but “everyone understands testing now everywhere in the world.” Rodriguez noted that this change put a new emphasis on what testing can accomplish and set a new global priority on testing for all diseases. The COVID-19 pandemic also demonstrated what a large investment of resources can do. Over the first nine months of the pandemic, testing capacity went from nothing to dramatically high capacity in most places in the world, including most of Africa and parts of Southeast Asia. It took about five years to develop a molecular test for tuberculosis that could be used globally and a few years for hepatitis C. In contrast, the first commercially available molecular tests for COVID-19 took only 64 days, and noncommercial tests were developed a mere 72 hours after the genome of the virus was identified (Molecular Devices, 2022). The first rapid test for COVID-19 available at scale globally took 236 days, while rapid tests for tuberculosis and hepatitis C are still “pipe dreams.” The lesson, said Rodriguez, is that when resources are applied and the prioritization is high enough, a lot can be accomplished. He noted, however, that there remain barriers to the development and deployment of tests around the globe, including regional manufacturing capacity and regulations. Rodriguez asserted that expansion and diversification of local production is needed to meet the global needs for testing, and regulators need to be more responsive. He noted that more complex tests, such as multiplex molecular tests, will likely have an even greater challenge getting through regulatory hurdles.

In addition to advancing the development and deployment of molecular and rapid tests, Rodriguez noted that the ACT Accelerator also dramatically improved the global capacity to perform genomic sequencing tests. Nearly every country can now do sequencing at a reasonable level, and many countries can do sequencing at high levels and have national genomic programs. At the time of the workshop, nearly 13 million sequences of SARS-CoV-2 had been contributed from 215 countries and territories (GISAID, 2023). Global capacity to perform sequencing is important for COVID-19 but is also critical to the future of efforts on AMR.

Rodriguez said that issues like COVID-19 or AMR are like climate change; they are complex, political, health-related, and fragmented. Similarly, the response requires close coordination and a shared agenda across many agencies that are all trying to respond to the crisis. One of the

major successes of the ACT Accelerator was bringing everyone together on a weekly basis to share successes, challenges, and priorities for testing, research and development, commercialization, and policy. This model, he said, needs to be applied to other diseases, especially AMR. However, despite collaboration and investment of resources, gaps persisted in procurement and uptake among countries. To control the spread of SARSCoV-2, stakeholders developed a consensus goal of one test per 1000 people per day. Over the first two years of testing, low-income countries were conducting about 0.02 tests per 1000 people, middle-income countries were conducting 0.15 tests per 1000 people, and high-income countries were vastly exceeding the target with a high mark of 7 tests per 1000 people per day (Figure 3-2). Rodriguez said that while stakeholders attempted to provide equitable access to COVID-19 testing, they “didn’t come close” to meeting their target to address the pandemic.

With lessons from the COVID-19 pandemic in mind, Rodriguez stressed that “the patient journey needs to be at the center” of a diagnostics strategy. Many of the emerging platforms for AMR—including POC platforms for sepsis and triage, and near-POC multiplex molecular platforms—would not be fit-for-purpose in many settings around the world. Rodriguez laid out a few challenges during the COVID-19 pandemic that may be relevant for AMR, including

• complexity of the global problem requiring collaboration among clinicians, epidemiologists, and politicians;
• limited evidence to inform policy;

• underprepared market;
• lack of clear product requirements;
• insufficient regulatory harmonization.

Rodriguez said that collaboration between academia and industry resulted in the rapid development and approval of COVID-19 tests, and government investment helped reduce costs for diagnostics, thus increasing manufacturing capacity and capability and improving access globally. The question, said Rodriguez, is whether these lessons learned can be applied to “rise to the challenge of AMR” over the years to come.

PROGRAMS TO ACCELERATE THE DEVELOPMENT OF DIAGNOSTICS

Paul Eder (Senior Scientific Officer, Concept Acceleration Program—Diagnostics, National Institute of Allergy and Infectious Diseases) began by describing what he means by “AMR” and “AST,” noting that these terms are used in different ways by different people. For the purposes of his talk, AMR means antimicrobial resistance prediction based on genotypic data, while AST refers to antibacterial susceptibility characterization through empirical data.

At NIAID, said Eder, there are a number of programs and resources aimed at accelerating the development of diagnostics for AMR. NIAID supports Combating Antibiotic Resistant Bacteria Biopharmaceutical Accelerator (CARB-X) Diagnostics, Therapeutics, and Preventives, and offers pre-clinical services for diagnostic development. Additionally, the Concept Acceleration Program (CAP)⁴—which helps support innovative technologies, platforms, and provides strategic advice to help shepherd candidate products through the research and development (R&D) and regulatory processes—at NIAID is well-established in the areas of vaccines and therapeutics, while the focus on diagnostics is newer. CAP has worked closely with the CARB-X and FIND to gather expertise and fill gaps in knowledge of the R&D and regulatory processes. He noted that CARB-X is currently in its second round of funding, which includes a focus on developing cost-effective and quick testing for gonorrhea. Bringing new technologies and assays to market quickly, and ensuring their success, said Eder, will improve the nation’s response to future emerging infectious diseases.

___________________

⁴ See https://www.niaid.nih.gov/about/biodefense-research-resources-contacts (accessed March 6, 2023).

NIAID’s Omnibus Broad Agency Announcement (BAA)⁵ is an annual solicitation for proposals to combat antimicrobial resistance. The areas of focus for 2023 include

• Faster bacterial identification direct from whole blood in under 4 hours;
• AST from positive blood culture or bacterial isolate in less than 2 hours;
• Improved performance in nucleic acid sequencing, mass spec sensitivity, protein sequencing, antigen or toxin capture and detection.

The awards for BAA are based on milestones and successful deliverables, based on a negotiated statement of work. This process—in which success must be made at each phase to proceed—is different from the traditional grant process. The budget for 2023 is up to $12.8 million for contracts across multiple research areas (NIH, 2022). Eder noted that contract awards are based on agency priorities, and currently they will not support development of diagnostic tests for pathogen ID from culture or isolate (bacterial plate), diagnostics that rely solely on the detection of host-response proteins, basic research and discovery of new host-based diagnostic targets, or diagnostic efforts that will require significant hardware development. Eder explained that other opportunities exist for these areas of development, but the hardware development is often an “albatross” that can significantly delay the process.

Pre-Clinical Services (PCS) is another way in which NIAID supports the development of diagnostics, said Eder. Helping product developers with PCS lowers their risk and encourages commitment to product development to reduce the burden on infectious disease. The program provides expertise and capability in product development to accelerate promising discoveries and fill gaps along the product development pathway. The agreements between NIAID and the developer, said Eder, assure confidentiality, maintain intellectual property of the developer, and encourage publication. Services provided by NIAID include specimen acquisition, reagents and assays, and product development support (e.g., design control, risk management). Services that are not provided include testing in animal samples and instrument and consumables development. Applying for support is a simple process, and most groups are eligible, including non-U.S. entities, academics, start-ups, NGOs, companies, and government entities. NIAID makes decisions based on priority, significance, innovation, preliminary data, value, and product development plan.

___________________

⁵ See https://sam.gov/opp/70720b778a894ce2ae311cc844f2a410/view (accessed March 6, 2023).

Finally, said Eder, NIAID offers research resources for diagnostics developers. These resources include organisms and reagents; resources for basic, bioinformatics, and ‘omics research; structure determination of proteins; and biocontainment facilities.

PANEL PERSPECTIVES

The panel brought together individuals with different perspectives to discuss the challenges in development and use of rapid diagnostics in healthcare settings. Eder moderated the panel, as well as a question-and-answer session that followed.

Clinical Pathologist Perspective

The COVID-19 pandemic led to advancements in medicine that were accomplished in a short amount of time, said Alex Greninger (Assistant Professor and Assistant Director of the Clinical Virology Laboratories, University of Washington Medical Center). Greninger said it is possible to capture this energy again for an issue like AMR, but the volition is lacking. There have been significant advances in rapid diagnostics in the healthcare space in recent years. Multiplex testing platforms, such as the BioFire FilmArray Pneumonia plus Panel, have demonstrated rapid automated testing of more than 30 targets with turnaround time of about an hour and hands-on time expected to be about a minute (Buchan et al., 2020). As panels grow to include ever-increasing numbers of analytes, reimbursement becomes a major issue. Greninger shared the 2022 Clinical Laboratory Fee Schedule (CLFS) rates for testing for respiratory viruses: with 3–5 analytes, the National Limitation Amounts (NLA) is $143; for 6–11 analytes, it is $218; and for 12–25 analytes, it is $417 (Centers for Medicare and Medicaid Services, 2023). The question, said Greninger, is how to value scale. There are efforts within the federal government to reduce reimbursement for laboratory testing, but with the government looking to cut costs while laboratory tests increase in complexity, said Greninger, this may not be a sustainable path forward.

Greninger shared a few thoughts on the development and use of rapid diagnostics in the healthcare setting from his perspective as a clinical pathologist. First, he said, “flexibility matters.” Multiplex tests and other rapid tests may end up in a wide variety of settings, from a quaternary hospital with AMR experts to a community hospital that doesn’t have the relevant antibiotics on formulary. Second, interpretation of tests should be “baked in” to their design. Greninger said that many people, including clinical pathologists, do not have an expert-level understanding of AMR and the tests used. Third, he said that reimbursement will be a major issue,

particularly for tests with a high number of targets. There is no way to pass along costs for highly complex inpatient testing with expensive reagents, and this is money “leaving the system.” Instead of using a complex machine to run tests automatically, Greninger said he would rather hire a technician who can run tests manually as well as answer the phone and fill out paperwork. Fourth, it is a challenge to get specimens for validation of new markers. As there are greater numbers of esoteric antimicrobial resistance markers, labs need to be able to order bacterial panels to validate targets. Greninger said there is an opportunity for a government agency to make these panels available so that tests can be brought online more quickly; he suggested that the Biological Exposure Indices (BEI) Resources Repository could play a role in distribution.⁶ Finally, he said that there is a need for research on whether separating AMR detection from a given species is appropriate or useful.

Regulatory Perspective

Kristian Roth (Deputy Director, Division of Microbiology Devices, FDA) shared his perspective as part of the review organization that authorizes tests for infectious disease. He said that companies are sometimes hesitant to bring new technology and new approaches to the FDA for review, particularly when there are no established clinical practice or clinical guidelines. Tests can sometimes also be well-established in clinical practice but not yet reviewed by the FDA. He gave several examples of how the FDA deals with these types of situations. When there is a low burden pathway to catch up with clinical practice, it is fairly simple to grant a claim (e.g., reviewing a quantitative test for cytomegalovirus when there are a number of qualitative cytomegalovirus tests on the market). The FDA has also used master protocols to grant new claims, for example, extragenital testing for gonorrhea. When multiple sponsors are following a master protocol, the FDA can grant a claim that would be difficult to validate to several sponsors in parallel. The FDA has also used a combination of existing literature and smaller clinical studies; for example, procalcitonin was used for years and was finally authorized after a process that included a panel meeting, outside opinion, evidence from the literature, and a small clinical study that validated each test’s performance. The final approach, said Roth, is the “brick-by-brick approach” in which data comes in and allows the granting of new claims for different intended uses.

Roth turned to lessons learned from the COVID-19 pandemic. He said that accelerated FDA review times are “good for everybody,” including

___________________

⁶ See https://www.niaid.nih.gov/research/bei-resources-repository (accessed January 25, 2023).

developers and the public. At-home testing for COVID-19 paved the way for at-home testing for other infectious diseases and more of these advancements are on the horizon. However, Roth noted a major negative impact of the pandemic has been the reduced level of clinical data that is expected. Manufacturers are in favor of the lower burden, but Roth questioned if it is good for public health. The EUA process greatly reduces the quality requirements that test manufacturers are responsible for; this has resulted in a tremendous number of recalls. According to Roth, Class I recalls, the most serious risk recall, nearly tripled in the last two and a half years. Further, the FDA has had considerable integrity concerns with the clinical data that are submitted, such as clinical studies that are too good to be true or biased clinical studies. These issues, said Roth, take “tremendous resources to address in a manner where all developers get an equal chance to get an EUA.”

Industry/Economic Perspective

There are several challenges in moving a new AST test from development to clearance, said David Persing (Head of Research and Development, Cepheid). There is growing complexity of resistance mechanisms for each antimicrobial class; complex issues including the numbers of sample types, different transport media, and CLIA waivers; and accessing the required specimens for clinical trials is difficult given the decreasing number of labs willing to participate. Given these complexities, said Persing, public-private partnerships may be a useful mechanism for moving tests into clinical use. Persing shared his experience of developing an assay called “MTB/RIF” that simultaneously and quickly tests for Mycobacterium tuberculosis (MTB) and resistance to rifampin (RIF). The test uses species-specific, nested PCR amplification of the MTB drug resistance target, and detection of rifampin resistance-conferring mutations via molecular beacons (Boehme et al., 2010; Hunt et al., 1994). MTB/RIF was the first demonstration of scalable implementation of direct-from-specimen detection of drug resistance, which started in 2005 with support from FIND. It was launched globally in 2010 with WHO endorsement and received 510(K)-clearance in the United States in 2015 with data from a National Institutes of Health (NIH) supported clinical trial. Persing shared that since 2015, 112 million test cartridges have been produced.

Since launching MTB/RIF, Persing has been asked a number of questions from people around the world. Investigators and clinicians want to know if they can get the equivalent of MTB/RIF for NG ceftriaxone resistance or a cartridge for NG fluoroquinolone susceptibility. They want a direct-from-urine test for ESBL-producing organisms, or a cartridge for direct detection of ESBLs, and carbapenemases for severe lower respiratory

infections and complicated UTI/urosepsis cases. However, said Persing, it is important to consider whether it is acceptable to test for a drug-resistance allele while not also identifying the organism at the species level. For example, does it matter if a UTI is caused by Escherichia coli (E. coli) or by klebsiella, or is it enough to detect the resistance allele to predict the likelihood of being resistant? The connection between predicting phenotypic resistance from genotypic information is “strong and getting stronger,” said Persing, but it is not as strong for predicting susceptibility. For now, conventional susceptibility testing is still required to fine tune drug selection.

Persing shared his experience working on the Xpert® Carba-R Test to detect carbapenem-resistance genes. The cartridge detects five families of resistance genes and can use specimens from rectal swabs, peri-rectal swabs, or carbapenem non-susceptible colonies (Jin et al., 2020). Persing emphasized that one important feature of this test is that it differentiates between metallo-beta-lactamases and non-metallo-beta-lactamases for the purpose of determining appropriate therapies. It is becoming “increasingly challenging” to keep up with the sheer number of targets that are emerging; because it is unknown which will become dominant, there is a need to maintain vigilance and modify tests as necessary to keep up with the variants. To obtain regulatory clearance, over 3,000 samples were tested, cultured, and sequenced as gold standards, and supplemental strains were used to evaluate the tests for contrived specimens. Government support was critical in validating the assays, said Persing. For example, the Antimicrobial Resistance Isolate Bank has bacterial isolates with emerging AMR genes that have been confirmed by DNA sequence analysis.⁷ Persing said that this continually updated resource is very valuable due to its comprehensiveness. The Independent Test Assessment Program (ITAP) at NIH is another government resource that could help with the regulatory process through the FDA.⁸ ITAP is aimed at accelerating regulatory review and availability of over-the-counter tests with an initial focus on COVID-19 tests. ITAP has recently expanded its purview to include mpox,⁹ and it could be a potential mechanism to accelerate the availability of AMR tests.

In summary, said Persing, direct detection of drug resistance is possible and effective when results are delivered in an actionable timeframe. There is significant potential for high-impact rapid AMR testing in patients with

___________________

⁷ See https://www.cdc.gov/drugresistance/resistance-bank/index.html (accessed January 25, 2023).

⁸ See https://www.nibib.nih.gov/covid-19/radx-tech-program/ITAP (accessed January 25, 2023).

⁹ Shortly after this workshop was held, the World Health Organization recommended changing the name of “monkeypox” to “mpox.” These proceedings have been edited to use the updated language. https://www.who.int/news/item/28-11-2022-who-recommends-new-name-for-monkeypox-disease (accessed January 30, 2023).

sepsis and “sepsis-adjacent” conditions. In an environment where technologies and resistance mechanisms are increasingly complex, the pathway to FDA clearance will become more complicated; public-private partnerships may be a critical approach to facilitate timely FDA clearance of novel products.

DISCUSSION

Advantages and disadvantages of separating bug and drug testing

Eder pointed to the potential for separating “bug and drug” testing—for example, conducting a test to detect the presence of antimicrobial resistant genes rather than the pathogen responsible for the infection. Eder asked panelists to comment on the advantages and disadvantages of this approach.

Persing used sepsis as an example of a case in which testing for AMR but not the specific organism could be appropriate. About 20 to 40 percent of sepsis cases originate in the urinary tract and about 20 to 40 percent in the respiratory tract (Chou et al., 2020). The question, he said, is whether testing for resistance markers at these sites would be sufficient to direct therapy in a patient with sepsis, and what impact would this testing have on patient outcomes? Persing noted that studies on this approach would be complicated but worth doing because they could lead to dramatic improvement in antibiotic stewardship and antibiotic selection. R. Patel said that she has grappled with this issue in the area of UTI diagnostics and shared two key considerations. First, it is important to ensure the patient being diagnosed with an infection actually has an infection; for example, there are patients with asymptomatic bacteria who may be diagnosed with UTI. Second, there is debate over whether it is necessary to know the name of the responsible organism or whether knowing how to treat it is sufficient. This issue created a lot of controversy among R. Patel and her colleagues, and she observed that the field of medicine may not be quite ready to accept treating an infection without knowing what it is.

Carroll said that an important consideration is where the infection is located and what this means for interpretation of a test for resistance markers. For example, if a patient has signs and symptoms of a UTI and there are resistance markers in the otherwise sterile urine sample, Carroll would feel comfortable using this information to direct treatment without information about the specific pathogen. However, a respiratory infection is a “whole different quagmire.” Resistance markers can be linked to normal respiratory flora or to the pathogens responsible for the infection. Because of this difference, she said, there is no universal approach for separating bug and drug testing, but there could be potential in areas that are otherwise sterile.

J. Patel agreed that there is likely no universal approach but pointed out that there are some resistance markers that lead to clearer decisions. For example, identifying a carbapenemases encoding gene in a urine specimen is very different than identifying ESBL in a urine specimen. There are more treatment options for an ESBL-producing organism, and J. Patel expressed worry that this type of testing might unnecessarily drive treatment escalation.

Rodriguez suggested that different types of clinicians might be amenable to using rapid diagnostics to inform decision-making. For example, emergency department physicians are accustomed to making decisions with minimal data to triage patients, while an infectious disease specialist may want a more comprehensive dataset. Ultimately, it may come down to regulators and payers and how they view the utility of these tests. In the global context, most clinicians would likely be willing to make decisions based on minimal data, but the future of these types of tests may depend on whether the WHO believes that it is sufficient to test for antimicrobial resistance without identification of the organism.

Following up on the importance of the regulatory perspective, Eder asked Persing and Roth to comment on the considerations for moving this type of approach through the regulatory process. Persing said that Cepheid does not have a particular strategy at this point but is interested in moving into the UTI space because of the potential for a significant impact. Roth said that in making a regulatory decision on this approach, the FDA would consider many issues, including sterile vs. non-sterile sites, potential for co-infections, and the availability of phenotypic data to validate claims.

Challenges of at-home testing

During the COVID-19 pandemic, at-home testing in the United States became more widespread and more user-friendly than it had ever been, said Eder. However, he noted that at-home testing requires self-collection of specimens, which may be prone to error. Roth pointed out that self-collection and interpretation improved as time went on during the COVID-19 pandemic. While there was a learning curve, people may now be as capable of collecting a sample as a health provider. However, this may not be the case for all types of samples. For example, it would be challenging for an mpox patient to collect a quality sample. Persing and Eder agreed that there is a need to design diagnostic tests in a way that enables users to collection specimens of adequate quality.

Another challenge of at-home testing, said Eder, is in the interpretation of results. He shared a story about a relative who called him after an at-home COVID-19 test and reported that his results were positive because there was a “huge line at the C, and C must mean COVID.” Daniel Bausch (Director of Emerging Threats & Global Health Security, FIND, The Global Alliance

for Diagnostics) suggested that as the technology improves for at-home tests and it becomes possible to run multiplex tests, it may be beneficial to simplify the results. For example, instead of having seven lines on a lateral flow assay that tests for seven pathogens, there could be one line for “viral disease” and one line for “non-viral disease” Based on this simplified result, the patient could follow up with their provider as necessary and potentially forestall some AMR by reducing the use of unnecessary antibiotics.

While not the primary focus of this workshop, global applications should also be taken into account. Millions of people around the world do not have routine access to health facilities and clinical testing. When considering the value of self-tests on a global level, Rodriguez said, the comparison should not be between self-testing and clinic testing, but between self-testing and no testing. Some self-tests may not perform quite as well as clinical tests, but the choice for many people is between a test with a slight drop-off in performance or no test at all. Rodriguez said that the performance of a self-test depends in part on the ease of specimen self-collection and that some specimens are easier to collect than others; the utility of a self-test needs to be considered on disease-by-disease, test-by-test, and sample-by-sample bases. However, he emphasized that the ability to access a test usually “dramatically outweighs the small drop-off in performance from self-testing.”

Leveraging pandemic collaborations for AMR

Eder shared a list of resources that NIAID offers to developers and stated that public-private partnerships bring together the resources of the government with the ingenuity of the private sector. Roth shared a project in which NIH collaborated with the University of Massachusetts to collect data on COVID-19 testing; these data were collected during an emergency situation and are now being used to make regulatory decisions. Roth said that they are hoping these sorts of partnerships can continue in more traditional settings and in other areas outside of COVID-19. Trainor asked Roth whether there are plans to leverage ITAP for other disease areas; Roth said he didn’t know. ITAP was tremendously important for the FDA during COVID-19, he said, because it removed questions about data quality and integrity due to the master protocol style. Data were submitted on an ongoing basis, which allowed the FDA to make input or ask for corrections or interpretation as the program went on. It was a tremendous success, resulting in six over the counter COVID-19 tests and is now being used for mpox.

Leveraging pandemic data management processes and infrastructure for AMR

During the COVID-19 pandemic, data management processes and infrastructure were established to help manage the sheer number of tests

being conducted, said Trainor. When considering how these resources could be leveraged for AMR, Rodriguez said that FIND is actively working to apply COVID-19 tools and policies to the issue of AMR. Currently, there remains a significant gap in AMR data for many parts of the world. A new WHO hub for pandemic preparedness will consider global surveillance mechanisms and may yield insights that could apply to AMR (WHO, 2022b). The challenge, said Bausch, is figuring out how to integrate AMR surveillance into routine healthcare systems; parallel systems are infeasible and unsustainable.

A workshop participant observed that due to COVID-19, there has never been a more acute awareness of infectious disease. “Amidst chaos lies opportunity,” he said, and this is the time to get out the message that antimicrobial resistance is a major threat to health.

This page intentionally left blank.