A global health hack: Gari Clifford’s 12-point path to a lifesaving maternal health program

By Mary Loftus Aug. 25, 2025

portrait of Emory University professor Gari Clifford

Gari Clifford was recognized as the 2025 Dean’s Eminent Investigator in the School of Medicine. Photo by Stephen Nowland, Emory Photo/Video.

Biomedical engineer Gari Clifford, professor and chair of biomedical informatics at Emory and Georgia Institute of Technology, was recognized as the 2025 Dean’s Eminent Investigator at a virtual event on Aug. 5. The Dean’s Eminent Investigator title recognizes one senior basic scientist each year in the School of Medicine whose contributions exemplify scholarly achievement and impact. 

Clifford’s research focuses on the design and application of signal processing and AI in medicine to classify, track and predict health and illness, particularly in low-resource settings.

From working with midwives in Guatemala on fetal cardiac monitoring to inventing low-cost health technologies to managing the effort to establish the first open-access critical care hospital database, Clifford “embodies the ideal of service to humanity,” says Emory Medical School Dean Sandra Wong.

Below are a few highlights from Clifford’s talk — 12 snippets from the story of how his work reached this point.

  1. For me, it started in the mid-1990s when I ended up in an operating theater in Venezuela after an injury. I saw the lack of resources available. I was a physicist who had few practical skill sets and zero idea of what I wanted to do, other than something with health care and resource-constrained areas.
  2. One of the first projects I worked on was building a bedside monitor from which we could collect data to build predictive models. I began investigating neural networks and machine learning. I was then lucky enough to work on a sleep staging device, adapted from a portable cardiac “Holter” monitor. We rigged up an ECG device to run a neural network inside it and commercialized it.
  3. I realized teaching is one of the best ways you can learn things. As a post-doc at MIT, I designed a course on mHealth in low-resource settings and ended up mentoring an incredible bunch of students who built a system for mobile diagnostics to help detect cervical cancer, among other diseases.
  4. I learned a lot about rapid prototyping while at MIT: How to build a team, how to take a first pass at whatever system you’re building and iterate until you have a viable tool. Using this approach, we put together a low-cost digital stethoscope. We looked at using car parts, soup ladles, egg cups, all sorts of things. We settled on an egg cup with a hole drilled in the back and glued the hands-free kit from a phone into the hole. The idea was to see if we could detect rheumatic heart disease in rural locations; eventually this system was tested in South Africa.
  5. My “day” job during this time was building and disseminating databases to supercharge data science in health care. With PhysioNet, we created the MIMIC-II database, the world’s largest (and for 10 years the only) open-access hospital database. It had about three terabytes of data, which doesn't sound like much these days, but in 2009, it was enormous.
  6. After returning to Oxford, I founded the Center for Affordable Healthcare Technologies, and that was where global health took off for me. It’s also where I got into digital mental health, co-founding the Sleep and Circadian Neuroscience Institute.
  7. I learned to adapt and pivot, and that the first idea will not be the best. There are always going to be things missing. This is what led to Safe+Natal. Thinking about how we could improve the digital stethoscope, along with [my wife] Rachel Hall-Clifford, who is my co-investigator, co-conspirator and co-faculty at Emory. At the time, we were both at Oxford; we became pregnant with our first child and started playing with a low-cost Doppler device to see if we could improve cardiac screening and fetal monitoring.
  8. Rachel was working in Guatemala, and I would often go along for the ride. We started brainstorming how we could use the device to support Indigenous midwives and mothers in Guatemala. We presented it to an NGO we’d been helping there, and they became quite excited about it. With NIH funding, we did a randomized trial to see if we could improve outcomes in the population. But the secret sauce is that it’s important to involve the community in the design and to make sure they feel ownership of the system.
  9. For over a decade, we’ve been providing decision support for these midwives through low-cost health technology, blood pressure devices, Doppler ultrasound being driven by a mobile phone. And we managed to drive maternal mortality down from 1,000 per 100,000 to less than 40 per 100,000, which is better than some U.S. states and exceeds the Millennium Development Goals.
  10. The Safe+Natal program is still running, and we’ve been developing AI on the data captured by the midwives. The AI is now being ported to the phone and used in the field: 100 midwives are using the system, and we’re being asked to implement similar systems in Colombia, Peru and Sierra Leone, where we’ve been running a trial for the last year, as well as Morocco, Kenya, Ethiopia and Tanzania. We’re currently thinking about how to scale this program up.
  11. I think the future of health care in lower-resource settings has to move toward Tiny Machine Learning (TinyML) and Edge AI. Edge AI is artificial intelligence plus edge computing, or processing data and decisions locally without relying on centralized systems or a network, e.g., using calculators, smartwatches, traffic lights, washing machines, fitness trackers, smartphones.
  12. Funding has grown from an initial NIH grant of about $250,000 to more than $6 million. In global health, $6 million can sometimes be more like $60 million when you're working in low-resource areas, so we feel incredibly fortunate to have this level of funding. Recently it was awarded MacArthur Foundation funding to scale it up and Google.org funding to get the real-time AI system into the hands of the midwives.

Gari Clifford (right) in Guatemala with his wife and co-investigator, Rachel Hall-Clifford, and their children. Photo provided.

Tags

Recent News