Shahnoza Eshonkhojaeva of Sinostream AB in Sweden will use machine-learning algorithms to predict the amount of medicines and supplies needed at individual health clinics in low-resource settings, and to inform medical stores for delivery. Their approach involves obtaining daily consumption patterns that are recorded on smart paper stock cards at rural health clinics, which requires no training, internet access, or electricity. These cards will then be scanned at district health service centers, the data digitized, and algorithms used to calculate consumption patterns and waste, and automatically predict future demand. They will build a prototype system and field test it in Uganda to evaluate how well it avoids under- or overstocking products, and the cost-saving and time-saving benefits of having an automated stock management system.
More information about Health Systems Strengthening: Ensuring Effective Health Supply Chains (Round 19)