Artificial Intelligence Tool Helps Combat Antibiotic Resistance, Misuse

A new study published in JAMA Pediatrics found that an artificial intelligence (AI)-based clinical decision-making tool can help improve antibiotic stewardship for diarrheal disease in settings with poor sanitation and hygiene.

According to the World Health Organization (WHO), diarrheal disease is the second leading cause of death in children under 5 years old. The disease caused the deaths of 370,000 children in 2019. Viruses cause many cases of diarrhea, but cases of the condition resulting from septic bacterial infections are on the rise.

Some diarrhea cases facilitate the need for antibiotics, but the varied causes of diarrheal disease require other treatment options. However, most pediatric diarrhea cases in developing countries are treated with antibiotics, regardless of the condition’s cause, according to the study authors.

“That means the vast majority of cases are treated inappropriately,” said Eric J. Nelson, MD, PhD, a University of Florida (UF) Health physician, a tenured associate professor in the UF College of Public Health and Health Professions’ environmental and global health department, and one of the study’s lead authors, in the press release.

Not only are antibiotics ineffective against viral diarrhea, but the inappropriate use of these medications is a major contributor to antibiotic resistance worldwide, the authors noted.

To combat this, the researchers developed an AI tool, known as the Diarrheal Etiology Prediction (DEP) algorithm, to estimate the probability that a case of diarrhea is caused by a virus alone to aid antibiotic stewardship. The tool works by gathering data, including clinical history, patient-specific symptoms, and local weather information, and using that data to generate its estimates. The algorithm was integrated into a smartphone-based clinical decision-support tool.

To test whether the tool could help doctors make better decisions about antibiotic use, the researchers collaborated with scientists in Bangladesh and Mali to enroll 30 doctors and 1,000 patients between the ages of 2 months and 5 years in 2020 and 2021.

Overall, the researchers found that using the DEP was not associated with a reduction in antibiotic use when analyzing antibiotic use independent of the likelihood that a patient had viral disease alone. However, when the predicted probability of viral-only diarrhea increased by 10 percent, antibiotic prescriptions fell by 14 percent.

These findings show the potential for using these clinical decision-making tools to help combat global health burdens, especially diseases with seasonality and geographic localization, such as malaria and different respiratory infections, the study authors stated.

“We’re trying to figure out how to build clinical decision-support software that is fast, easy to use, accurate and meets the needs of health care providers in global health settings,” concluded Nelson.

This is the latest research effort focused on using AI to address global health concerns.

Last week, Google announced that it had developed a deep-learning model that can detect tuberculosis as accurately as radiologists using chest radiographs.

Earlier this month, Mayo Clinic shared a newly developed AI-based risk prediction model that successfully informed pregnant patients of labor risks and improved clinical decision-making.

In addition, a study published last month found that a prediction model could accurately forecast health-related quality of life (HRQOL) among adult survivors of childhood cancer using sociodemographic, lifestyle, and health state factors.