During a typical day at a rural Rwandan clinic, nurses can see as many as 60 patients. Adults and children arrive with injuries, coughs and fevers after traveling long distances. With little time for diagnosis, many nurses err on the side of prescribing something — often antibiotics. “You find them giving a high number of antibiotics, just in case,” says Dr. Victor Pacifique Rwandarwacu, a physician from Rwanda. “They’d be like, ‘OK, what if I don’t give it, and then the patient comes back tonight?'”
That attitude has produced extremely high prescription rates. In new research across 32 clinics in Rwanda, 71% of pediatric visits ended with an antibiotic prescription, far higher than necessary, the authors report in PLOS Medicine. While a single unnecessary course may not seem consequential, widespread overprescribing fuels antimicrobial resistance. “In sub-Saharan Africa, the rise in antimicrobial resistance is enormous,” says Jean Claude Semuto Ngabonziza, a researcher at the Rwanda Biomedical Center and a study co-author. “We are at the edge of losing potential antibiotics.”
To address this, the researchers developed ePOCT+, a tablet-based clinical algorithm that guides clinicians step-by-step through assessment and treatment suggestions. It distills existing guidelines into an easy-to-use tool and incorporates simple diagnostics such as oxygen saturation and hemoglobin. A nurse assessing suspected pneumonia, for example, enters symptoms and test results; the tool classifies cases as simple, bacterial or viral pneumonia, and recommends antibiotics only for bacterial cases. The typical encounter using ePOCT+ takes about 10 minutes.
Training was brief: one day of practice was sufficient, according to Alexandra Kulinkina, an epidemiologist at the Swiss Tropical and Public Health Institute and study co-author. The researchers tracked nearly 60,000 visits. After implementing ePOCT+, antibiotic prescriptions fell from 71% to 25% of pediatric visits — a dramatic reduction that did not lead to worse health outcomes, the authors report. “The most important thing is they’re not compromising health outcomes,” says Dr. Sumanth Gandra, an infectious disease researcher at Washington University in St. Louis who was not involved in the study. He adds that such tools could be useful and scalable.
Kulinkina notes the 25% rate is probably still higher than ideal; under optimal conditions perhaps 10–15% of patients would need antibiotics. Some nurses did not always follow the algorithm. Still, the tool also helped nurses identify overlooked conditions like malnutrition and anemia. Patients and nurses were generally positive: one mother said she felt more thoroughly examined than in prior visits, and nurses appreciated the added guidance, though consultations took longer.
The Rwandan government has taken notice. The Ministry of Health is developing a national electronic medical record system, and officials have discussed integrating aspects of ePOCT+. The Rwanda Social Security Board, which funds public insurance, showed strong interest because fewer unnecessary prescriptions could reduce costs.
Scaling up will present challenges — training thousands of health workers, ensuring consistent use, and maintaining the system — and results may vary outside the study setting. But the researchers consider the approach promising for improving care quality and slowing the rise of antibiotic resistance. “We are at the edge of losing potential antibiotics,” says Ngabonziza. “Sometimes we do research and it’s so-so, but this is really impactful.”
