Improving the Early Diagnosis of Neonatal Sepsis in Malawi

Globally, infection leading to sepsis in infants mimics many conditions and hence early diagnosis is difficult. However, failure to treat early with antibiotics will uniformly lead to death or major disability. As a result, more neonates are unnecessarily treated with antibiotics. Over-treatment of neonates creates resistant superbugs and wastes scarce resources. To address this problem, we will build a diagnostic algorithm incorporated into a mobile phone app to help doctors identify babies in need off rapid treatment at the time of presentation for suspected bacterial sepsis, based on subtle variations in their vital signs. The bold innovation uses advanced molecular blood tests to detect the response to infection in a newborn with high accuracy and hence use this test to improve the diagnostic precision of mobile diagnostic devices. We will obtain clinical and vital sign data, and a small blood sample in 500 newborns presenting to the Kamuzu Central Hospital in Malawi over a one-year period. Vital sign data will be captured using a low-cost hand/foot probe connected to a mobile phone, and will be used to build diagnostic predictor models based on the molecular signature of a bacterial infection measured in blood of newborns, used as a gold standard. The predictor model will be incorporated into a low-cost mobile app and sensor device (a commercial partnership with LGT Medical) that will guide critical life-saving interventions for newborns with sepsis around the world.