Treating antibiotic resistance: Need for diagnostics and AI

In 2016, Médecins Sans Frontières opened a microbiology laboratory in its hospital in Aden, Yemen, to test for antibiotic resistance. Intense fighting between armed groups has resulted in the arrival of hundreds of wounded patients in need of treatment, and a huge number of them suffered from antibiotic-resistant bacteria. But there was a lack of microbiologists who could interpret the test results.

So every morning, miles away at the MSF Foundation headquarters in France, Dr. Nada Malou, MSF’s microbiology and surveillance advisor, checks the test results — averaging 300 per month — and interpret them, including tests from MSF’s microbiology lab in Mali.

From that experience, the idea for a mobile application that could help laboratory technicians perform the test interpretations was born. The app, called Antibiogo, was developed by the MSF Foundation in collaboration with scientists from the University of Évry and the Bacteriology Department at the Henri Mondor Hospital, and received a grant from Google.

Malou said it allows laboratory technicians to take a photo of antibiograms — cultured bacteria taken from patients and placed in a Petri dish to see how it reacts to a set of antibiotics. With artificial intelligence, the app processes the image and helps laboratory technicians identify the antibiotics in the dish and interpret which antibiotics the bacteria is most sensitive or resistant to.

But the app, which is now being trialed in MSF and partner laboratories in Jordan, Mali, and Senegal, could also help in antibiotic stewardship. It shows a summary of the test result and a recommendation of which antibiotics clinicians can use to treat the patient following their hospital’s guideline.

This prevents clinicians from prescribing strong antibiotics when a milder but effective antibiotic is available, helping combat antibiotic overuse that could worsen antimicrobial resistance, which is the subject of a high-level dialogue taking place this Thursday. U.N. General Assembly President Volkan Bozkır will bring together stakeholders to discuss and recommit to address AMR.

Combating drug resistance is among the priorities the World Health Organization identified for 2021. Drug-resistant diseases could cause the deaths of 10 million people each year by 2050 globally, according to a 2019 U.N. report. But keeping the attention on AMR can be a challenge in the face of headline-grabbing diseases like COVID-19, and with the lack of documentation to identify just how many patients die from hospital-acquired infections caused by AMR, particularly in resource-poor settings.

Current investments in antimicrobial resistance seem to focus on finding new antibiotics to treat drug-resistant bacteria, but experts argue there is a need to invest in diagnostics too to protect existing and new antibiotics in development.

“When we reviewed the pipeline of diagnostics for AMR and treatment for AMR, you have an investment of 10% on diagnostics over 90% on antibiotics, when we know that you won't be able to use those new antibiotics if you don't have access to proper diagnostics,” Dr. Cecilia Ferreyra, head of the AMR program at the Foundation for Innovative New Diagnostics, told Devex.

Other diagnostic innovations in development focus on tests for sepsis and sexually transmitted diseases such as gonorrhea. However, experts anticipate it will take years before these innovations, if proven useful, could be made available to low- and middle-income countries. That depends on funding too.

While CARB-X, an international nonprofit, funds early development of antibiotics, diagnostics, and other products targeted at addressing drug-resistant bacteria, Erin Duffy, its chief of research and development, told Devex that except for the U.S. Biomedical Advanced Research and Development Authority, it’s unclear how many organizations fund later-stage development diagnostics.

“We worry about this a lot,” she said. “Let me tell you, the last thing we want to do is get these companies to sort of this … graduation stage, and then they fall off a cliff.”

Diagnostics to protect antibiotics

Gonorrhea is a sexually transmitted infection that has already shown resistance to a range of antibiotics.

FIND supports a number of projects, including the development of low-cost diagnostics to rationalize use of existing and new antibiotics. They include a rapid, lateral flow assay test — similar to a pregnancy test — and molecular-based point-of-care tests to identify gonorrhea. These tests do not require electricity and could be used at the primary health care level.

Ferreyra expects the technological development of these tests to conclude in July 2021, and she hopes the organization would receive donor support to continue supporting the development of these tests that WHO has recognized as a game-changer for sexually transmitted infections.

There is a critical gap in testing for gonorrhea, particularly in resource-poor settings. NAATS [Nucleic Acid Amplification Tests] platforms, which are used for HIV and TB diagnosis and coronavirus, are available in a number of countries and can be used for the diagnosis of gonorrhea. However, the cartridge used for the test can cost up to $20, which “is not at all affordable neither for patients [nor] for the health systems” in low- and middle-income countries, Ferreyra said. In addition, those tests are difficult to carry out at the primary health care level.

Current WHO recommendations to diagnose gonorrhea rely on syndromic case management, or where clinicians base their assessment on a patient’s symptoms, she said.

“WHO is not recommending any diagnostics … which means that all the patients arriving into a clinic with a vaginal discharge or urethral discharge, they won't get any kind of diagnostics. And the clinicians will follow some flow charts to determine whether the patient has an STI [sexually transmitted infection]. And [they] almost end up with 100% getting antibiotics that they might not need,” she said.

The need for faster diagnosis

Meanwhile, CARB-X is currently supporting nine diagnostic projects in its pipeline, each at different stages of development. Most of these tests appear to be targeted toward the high-income market, but there are two that could potentially be used in low- and middle-income countries. However, they are still in the very early stages of feasibility assessment, said Duffy.

One of them is a diagnostic platform to diagnose neonatal sepsis, a blood infection affecting infants that could lead to death. According to a WHO report in 2020, there were an estimated 375,000 neonatal deaths globally due to sepsis in 2018, representing 15% of all neonatal deaths. It’s a major cause of infant mortality, but it’s not easy to diagnose, according to UNICEF.

U.S.-based Baebies Inc. aims to develop a rapid diagnostic tool that first determines whether an infant has a bacterial infection through a small blood sample, and then identifies the pathogen causing the infection through a molecular test to allow clinicians to use the appropriate antibiotic treatment. Each test is targeted to be done within a 15-minute time frame.

The other diagnostic product is by another U.S.-based company, Novel Microdevices, Inc., a battery-operated molecular test that aims to diagnose sexually transmitted bacterial infections, including drug-resistant ones, through a vaginal swab or urine sample, in a span of 25 minutes.

“I think in the general sense, with our entire diagnostics portfolio, you know, there is this recognition [that] how we normally diagnose bacterial infections and also then determine what therapy to use … can take up to 48 hours, and ... frankly, you don't have a lot of time with some of these infections,” said Duffy.

This coverage, presented by the Bay Area Global Health Alliance, explores the intersection between technology, innovation, and health. How are tech, innovation, and cross-sector partnerships being leveraged to accelerate equitable access to health care?