Abstract

During pregnancy, some mothers get infections, and these infections can cause labor to start too early. The baby is also at risk of getting infected from the mother and being born with an infection, especially for early or small babies, and this can lead to a deadly condition called sepsis. Infants are at highest risk of developing sepsis and dying in the first days after birth, but doctors are not available to assess all newborns in low-resource settings. Approximately 20-36 times more infants get sepsis in low- and middleincome (LMICs) countries compared to high-income countries, therefore, preventing sepsis could make a huge difference in an infant’s health. When babies are sick but not assessed, they are more likely to die or have worse health outcomes. There is a special organ, called the placenta, that forms only during a pregnancy. Doctors can examine the placenta for signs of infection, but this process takes at least a few days and requires very special training. We have developed a quick tool to assess placentas using artificial intelligence (AI) on digital photographs. The goal of this project is to develop and test a new, AI-based tool that puts together data from the mother, newborn, and placenta into a low, medium, or high category of risk for sepsis. This tool would be in the form of a smartphone app and could lead to a baby getting more care from healthcare providers. The RAISE (Risk Assessment for Infant Sepsis) tool will be user-friendly, not require specific placenta or clinical expertise, and be made specifically for use in lowresource settings. We will conduct a study at a large, regional referral hospital in Uganda that has ~10,000 births per year and many cases of sepsis in newborns; we have been conducting research here for 10 years. We will first develop the RAISE tool (aim 1) using our placenta software combined with data we have from other studies. Then we will evaluate how easy or difficult it is to use the RAISE tool (aim 2) by having midwives test it out. Finally, we will validate the RAISE tool (aim 3) by comparing the result of the tool to the result from a doctor examining a baby at birth. In this project, experts in pregnancy, newborns, placentas, and computer science will all work together to develop and test the tool. We expect the RAISE tool will improve treatment of newborns with infections that may seem healthy right at birth, which in turn may reduce the number of bad outcomes and deaths in these vulnerable infants.