Grantee: Macro-Eyes

Developing predictive supply-chains using machine learning for improved immunization coverage

Drew Arenth, MBA, Chief Business Officer, Macro-Eyes
A close-up of Hadija Mwasimaga’s newborn baby boy, Isaka.
UNICEF/UN0270558/van Oorsouw
17 July 2019

Bill & Melinda Gates Foundation and UNICEF are joining forces  by accelerating Innovation for strengthening Vaccination Systems -- starting with better collection and use of critical data. UNICEF’s Innovation Fund is supporting eight grant recipients of the Gates Foundation’s Grand Challenges Explorations Round 22 with its cutting edge piloting and scaling methodologies and tools.  Read more from Grant Recipient Macro-Eyes, below.



Global coverage for basic childhood vaccines has reached a record 85%, yet there has been a parallel increase in vaccine wastage.

Valuable immunization data and context collection requires engagement from the world’s foremost experts -  frontline health workers. It also requires trust in the collection and prediction process.  Frontline health workers are domain experts - they know more than anyone about the delivery of vaccines in their communities, how the catchment population perceives immunization campaigns and how they will be able to access vaccines.

Frontline health workers bear the brunt of data work, but see little reward. They deserve better.

Throughout the world, data collection has historically been a process of checking a box. That data disappears never to resurface in any meaningful way - especially for those at the frontlines of data collection. They have become tired, unrewarded, and uninspired contributors in a system that does not allow them to share the valuable context they hold. As a result, data is inaccurate, unreflective of crucial information, and incomplete.

Predictive supply chain for vaccines

Solution in Action

We observed the power of context while building the predictive supply chain for vaccines in Tanzania.  Our partner PATH and the Ministry of Health shared data, they explained their process of data collection and pinpointed features they considered important. 

They told us stories. AI is often imagined as work in some kind of sterile lab: scientists fiddling with vast fields of numbers, ordinary life nowhere to be seen. The image couldn’t be further from the truth. Refining our machine learning models, we discovered that the stories behind the data, the context of how and why data was collected, had significant impact on the accuracy of  predictions. Frontline health workers gather information valuable for understanding and predicting vaccine utilization; systems don’t benefit from this insight.

The predictive supply chain for vaccines is AI that predicts future vaccine consumption and recommends appropriate levels of supply. The technology anticipates shifts in demand and consumption, decreasing stockouts and wastage and ultimately increasing opportunities for immunization. 

Macro-Eyes demonstrated in Tanzania that the predictive supply chain for vaccines can forecast vaccine consumption, down to the level of individual facilities, with 70% greater accuracy than the best performing model on the market.  The precision supply chain is a more equitable supply chain.

 To bring the predictive supply chain to global scale – to more frequently machine learn insight in real time, be more responsive to previously invisible populations, implement in settings with limited data – we designed an expert-in-the-loop machine learning process to gather context and insight from frontline health workers.

Hadija Kuziwa (32 years old) with a prematurely born baby in front of the neonatal ward of the Mbeya Regional Referral Hospital in south-west Tanzania where Hadija works as a nurse midwife.
UNICEF/UN0270670/van Oorsouw

Team & Diversity

We are all committed to solving the last-mile problem, using AI not just to predict the future but directly shape the future, to solve health problems before they occur. We each bring different capabilities and perspectives on the world. Macro-Eyes Chief AI Officer Dr. Suvrit Sra is a professor at MIT and global leader in advancing machine learning and AI.  Dr. Sra was recently awarded National Science Foundation funding to continue his work advancing expert-in-the-loop machine learning in healthcare. 

Dr. Ramkumar Hariharan leads applied AI at Macro-Eyes. He is the architect of the Macro-Eyes predictive supply chain for vaccines in Tanzania, which proved that routinely collected data could be used. 

Chief Software Engineer Johannes Kitschke has designed and built Macro-Eyes machine learning powered software that predicts patient behaviour and uses this insight to maximize health facility utilization and increase access to care, deployed in safety-net institutions across the United States.

Chief Business Officer Drew Arenth, MBA has designed and implemented innovative supply chain and technology solutions across sectors. Drew was the Principal Investigator for the predictive supply chain for vaccines in Tanzania .

Macro-Eyes began working in global health at a time when only difficult problems remained. What remains are challenges that require unique solutions and unorthodox teams with new approaches and relentless passion.

The Way Ahead

In five years we will be expanding our impact country by country, region by region, and continent by continent to improve access to essential health care and essential health goods. Machine learning powered data use and data collection will be the pillars on which we will power our impact.