macro-eyes announced today that it has received milestone funding via a grant from the Bill and Melinda Gates Foundation to bring the Connected Health AI Network (CHAIN) to Tanzania and Mozambique. CHAIN is supply chain [artificial] intelligence that makes the delivery of healthcare predictive and efficient, enabling governments to reach more people in need with the same resources.
CHAIN builds on the successful design and testing of a core predictive engine for vaccine consumption in three regions across Tanzania. In 2018, macro-eyes machine learning technology was trained and tested using routinely collected data to predict vaccine utilization in Tanzania. Historical vaccine utilization data was pulled from the Tanzania Immunization Registry (TImR) from more than 700 sites of care. The resulting machine learning system can predict utilization with 70% greater accuracy than the best performing approach on the market today.
Increasing forecasting accuracy by 70% would mean cutting stock-outs and a dramatic reduction in the number of vaccines required to increase vaccination coverage. Cost of vaccine programs would be lowered, and resources would reach far more children.
This groundbreaking work is being done in collaboration with the Ministries of Health in Tanzania and Mozambique, PATH, and VillageReach. The consortium engages a breadth and depth of sectoral and geographic expertise to make the delivery of care predictive.
The body of work is comprised of four distinct components:
Build for interoperability: Design and build system-to-system interfaces to enable the Predictive Supply Chain for Vaccines (PSCV) to be seamlessly integrated in Tanzania.
macro-eyes will construct application program interfaces (APIs) to seamlessly integrate with existing health system and supply chain software. The PSCV will generate intelligence, designed to engage the right stakeholders and enable informed decision making for the delivery of vaccines.
The PATH Better Immunization Data (BID) Initiative team and the Ministry of Health of Tanzania will provide local expertise and system access for supply chain software and routinely collected data.
Build for scale: Design incentive structures, natural language processing models, and data aggregation models for “Human in the Loop” (HIL) machine learning to gain insight and rare information from frontline health workers: the world’s foremost experts on the populations they serve and the context for care.
In order to improve access to real time insight – macro-eyes will design and deploy HIL machine learning. The most valuable insight into the supply chain and context for care is rarely shared outside communities of frontline health workers. Frontline health workers bear the brunt of data work, but see little reward. They deserve better. They deserve to be treated as experts.
macro-eyes will work closely with the Ministry of Health in Mozambique and VillageReach to deploy cutting edge HIL machine learning to collect insights in real time from on the ground experts.
Insight from the frontlines, accessed in real-time, will complement routinely collected health data and make it possible to bring AI to make the delivery of care predictive to environments where there’s limited historical data, enabling systems to anticipate change on the ground and intervene early to avoid stockouts, increasing opportunities for immunization.
Auditability: Build machine learning models and mobile application to collect mOPV and family planning stock images and generate accurate stock levels at facilities, pharmacies, and storage sites.
Auditing stock is time-intensive and costly. Current practices often generate inaccurate results and consume valuable time from those at the front lines of care. macro-eyes is working with VillageReach in Mozambique and PATH in India to build a mobile application that can generate highly accurate counts of utilized mOPV vaccines and family planning commodity. Health workers will be able to generate stock counts by simply pointing a phone camera at where vaccines are stored. The technology will lower the time burden for data entry, help to ensure that every product is accounted for, and make it possible to measure stock accurately in real-time. Accurate, reliable counts improve forecasting, decreasing the cost of care and increasing access to health commodities.
Understand Demand: Design and build a comprehensive, predictive model for vaccine demand.
There is growing consensus that ensuring adequate supply is not enough to reach and then vaccinate every child. macro-eyes will use a rich array of data to understand what drives demand for vaccination services for every type of caregivers. Core to this work is building multidimensional phenotypes that describe each type of caregiver in detail that extends from the geographic specifics of the environment in which they live to how they access care to socio-economics. Phenotyping is a way to segment large populations into smaller, self-similar groups that can tell us much about why and how and when a family member chooses to vaccinate a child. When we understand the elements that predict whether a given group is more or less likely to vaccinate their children, partners on the ground can then much more precisely intervene to drive demand.
CHAIN – Each body of work forms part of the Connected Health AI Network (CHAIN), an overarching technology to frame the supply chain as the central nervous system of a new way to delivery healthcare, to make the delivery of care predictive. Insight into demand and the ability to predict utilization are ground-truthed with real-time information on stock. New, rare insight shared by frontline health workers – the world’s foremost experts on the context for care in their communities – brings valuable, real time insight into the health system. These components combine to build efficiency and accuracy into the supply chain, which can then become an engine for equity and efficiency.
Comments are closed.