Physician, Heal Thy World

Interview by Alexander Gelfand | Photos by Audra Melton

As physician Von Nguyen ’97 explains it, public health tends to be the purview of government. Nguyen himself, for instance, spent many years at the Centers for Disease Control and Prevention and the Centers for Medicare and Medicaid Services, where he helped fight cholera in West Africa and worked to improve public health and health care delivery systems here in the U.S. This makes it all the more remarkable that he now does what amounts to public health work for Google.

At the tech giant, Nguyen is clinical lead for population health — a job that entails using his knowledge of medicine and health policy (he holds both an M.D. and a master’s in public health) to address issues ranging from health care access to extreme weather. Since joining Google in 2021, he has, for example, worked on teams that explored AI-driven methods for making mammograms and maternal ultrasounds available to women across the Global South and worked on a machine-learning platform for predicting floods and alerting at-risk populations.

Born in Vietnam and raised in the U.S., Nguyen grew up dreaming of becoming a doctor. His interest in improving the health of vulnerable populations was sparked by an experience doing HIV research for the CDC in Kenya on a Fulbright fellowship and strengthened by practicing primary care in India for Doctors Without Borders and providing medical care to underserved populations in Boston.

Nguyen talked with us about the experiences that shaped him and the work he does today.

How did your background influence your career choices?

One of the defining characteristics of who I am is that I was a political refugee from Vietnam.

You’ve seen those pictures of the helicopters taking off from the embassy in Saigon in April 1975. My family got out two days before those were taken. I was 6 months old. So I would say that everything I have now is the result of a lot of good fortune  —  a little bit of work, but a lot of good fortune. That has resulted in a need to pay it forward.

I thought medicine would be a good opportunity to do that. I thought public service in the U.S. government was a way to do that. You can complain all you want about the U.S. government — and I do my fair share — but it gave me everything, from the opportunity to get out of Vietnam to the financial aid to attend Rice. When I worked for the federal government, it was as much a calling as it was a profession.

And now that I work for a multinational tech company, I intentionally focus on projects that allow me to contribute as much as I can within the context of the job that I have.

Your role at Google involves public and population health. What does that entail?

There are two buckets of work that I do at Google. One is very much health-system focused. I know Medicare and Medicaid well, I understand the unique needs of those populations, and I help build tools for them. The other is based in the research arm of Google and involves thinking much more broadly about populations. I try to understand how to get information to people that they can use to improve health. An example of that would be our work at the intersection of health and climate change.

How are the two connected?

A changing climate impacts health in very significant ways.

Extreme weather is becoming more frequent and more severe. You’ve heard about the floods in Pakistan last year that covered a third of the country? Floods affect folks in the U.S. as well, in Australia, literally around the world. And floods have health consequences: When a flood happens, people die.

On top of that, when people are displaced because of flooding, their access to health care changes. “I moved to this new area. Well, where’s my doctor? Where are my medications? How do I continue my therapy?”

That happens in the U.S., but it happens even more in the low- and middle-income countries of the Global South.

How do you help people cope with these issues?

There are two big questions: What can we do to mitigate climate change — meaning, what can we do to prevent carbon from being released? And what can we do to help people adapt to a changing climate?

I work on the adaptation side: When there is a change in climate, what products and tools can improve health outcomes? One example is the flood prediction and notification tool that we rolled out in India and Bangladesh. At Google we have really good maps. And not just regular maps but topographical ones. If you know how much water is coming downstream and you know how much rain occurred in an area last year and the year before that, you can use machine learning to predict when a flood will hit so that people can get out of harm’s way.

I work with the engineering teams to help design those products and also to understand their impacts on health outcomes.

We’re also starting to do more work around extreme heat: working with the engineering teams to understand the tools and information that policymakers need to mitigate extreme heat and prepare their cities for the next heat wave.

You’re talking about Google’s air quality and tree canopy tools, which use sensor data and aerial imagery to help municipal officials and urban planners visualize street-level air pollution, which is worsened by extreme heat, and tree canopy coverage, which can help mitigate it.

Yes. I was a policy director at the CDC for a number of years. I understand the decisions a policymaker needs to make and how to make them.

One of the most effective ways to improve health is by making the healthy choice the easy choice. Tree canopy, for example, is really helpful for creating place-based health solutions in underserved neighborhoods. Building a walkable neighborhood increases physical activity, and if you build sidewalks with trees along them, more people use them.

There’s a growing recognition that issues outside of health care, such as transportation, housing and food insecurity, all affect health in different ways. You need to address those issues simultaneously.

You’re also developing technology to improve health outcomes outside of climate change — for example, by using AI to analyze maternal ultrasound images using cellphones.

Right. I have two girls. They are now 12 and 9. I remember going to their ultrasound appointments with my wife, who is a physician too. The technician swipes a probe around her belly and she’s like, “There’s all this snow; what are they looking for?” Eventually — magically — a face appears. It takes years to train a technician to do that.

But technology has advanced far enough that you can attach an ultrasound probe on your phone and train a computer to read it. A community health worker in the Global South can swipe left to right three times and bottom to top three times, and machine learning will interpret that snow and tell them the gestational age as well as the positioning of the baby.

That’s important because there are certain checkpoints at certain gestational ages where you do certain interventions. For example, if you’re approaching 40 weeks and the baby hasn’t turned head down, it’s important to move the mother near a hospital for a possible emergency C-section.

As clinicians here at Google, our role is to inform the decisions that our engineers are making: “What’s the right time period for a scan? What exactly happens at 36 weeks?” We have some of the best engineers in the world, but they’ve never set foot in a hospital.

Tell us about the work you’re doing to build algorithms that can interpret mammograms that are used to screen for breast cancer.

Depending on where you are, it may take a radiologist a week to get around to reading your mammogram. We basically trained a machine-learning algorithm to read a mammogram. When you compare the algorithm to a radiologist, it does pretty well.

We are not trying to replace doctors, but if a radiologist has a hundred mammograms to read, we could tell them that these are the two you need to pay attention to today. That accelerates the time to care, and that is incredibly helpful in terms of treatment. It also reduces anxiety for the women who are at risk.

So AI really can save the world — or at least improve health in parts of it?

You have to deal with the right safeguards to make sure that you’re maximizing equal outcomes and to make sure that you’re not actually doing harm, because there’s always the possibility of great harm with these tools if you’re not doing it right. One of the challenges of AI, for example, is that if you train the model using a biased dataset, it will be a biased model; it will pick up the unconscious bias that the data captures.

But there is also an opportunity to advance knowledge and improve the care that people deliver.

I remember traveling in the South Pacific in the 1990s. I had to go to the post office to make a phone call; that was the only place there was a phone.

Then I remember traveling in Africa in the 2010s, and everybody had cellphones. They skipped having landlines in people’s homes; it wasn’t necessary because of the advancement of technology. The Global South just leapfrogged that whole phase of development in tech.

I think we have the same opportunity with artificial intelligence to support decision-making and drive better outcomes.  If we play our cards right, there’s an opportunity for the Global South to catch up to, if not surpass, what we’re doing here in the West.