Growing up, Valencia Koomson wasn't supposed to be the one playing with the electronic set - that was her brother's Christmas present. She played with it anyway, and now she runs a lab at Tufts University focused on designing precision biosensors. She's turned that expertise toward addressing a long-overlooked problem in medical technology: pulse oximeters that work less reliably for patients with darker skin.

In this episode, she explains what it actually takes to engineer a more inclusive pulse oximeter and she shares how she's navigating the complexities of commercialising a new medical device.

You'll learn:

- Why pulse oximeters tend to be less accurate on patients with darker skin tones

- The engineering principles Koomson is using to redesign the device from the ground up

- How she thinks about bridging the gap between lab innovation and real-world impact

---

*** Help us shape the future of Made For Us! Take our 4-minute listener survey for the chance to win a $25 Amazon Gift Card and get instant access to our curated reading list of every book ever recommended on the show: https://forms.gle/pNyrCooa23oYsaVk7

---

About Valencia Koomson

Valencia Koomson holds the position of Associate Professor in the Department of Electrical and Computer Engineering at Tufts University with a secondary appointment at the Department of Computer Science and Tisch College of Civic Life. She is the founding director of the Advanced Integrated Circuits & Systems Lab at Tufts with a research focus on medical device innovation, global health technology, and health equity advocacy. Dr. Koomson completed B.S. and M.Eng. degrees in electrical engineering and computer science at MIT. She was awarded the George C. Marshall scholarship to pursue post-graduate studies at the University of Cambridge where she received the M.Phil. and Ph.D. degrees in electrical engineering. Dr. Koomson has authored over 60 publications, book chapters, and patents. She was awarded the Dr. Martin Luther King Jr. Visiting Professorship at MIT in 2021.

Follow Valencia Koomson on LinkedIn: https://www.linkedin.com/in/profkoomson/

Learn more about the Advanced Integrated Circuits and Systems Lab at Tufts: https://engineering.tufts.edu/ece/koomson

---

Other episodes you might like:

The pulse oximeter problem: a trusted medical device comes under the spotlight

How to design a fairer healthcare system

---

Connect with Made for Us

- Show notes and transcripts: https://made-for-us.captivate.fm/

- LinkedIn: https://www.linkedin.com/company/madeforuspodcast

- Instagram: https://www.instagram.com/madeforuspodcast

- Newsletter: https://madeforuspodcast.beehiiv.com/

The first thing is having inclusivity in mind. To start off thinking that we're going to design a device that works over the full range of skin tones that are encountered on our planet.

This is Made For Us, the show where we explore how intentional design can help create a world that works better for everyone. I'm your host, Tosin Sulaiman. Today, we'll meet the scientist reinventing a medical device that became a household item for millions of people during the COVID pandemic. Pulse oximeters are used to measure oxygen levels in the blood, but studies have shown that these devices tend to be less accurate in patients with darker skin. As Tufts University Professor Valencia Koomson began to dig into this research, she thought

This is fundamentally a sensing and signal processing problem that we can actually solve. We know how to solve.

And that's exactly what she's doing, designing a new type of pulse oximeter that works across diverse skin tones. It's the kind of challenge that draws on her expertise. She's a professor of electrical and computer engineering, and her lab at Tufts focuses on designing biosensors that improve people's quality of life. In this episode, she tells me how she's reimagining a device whose core design has barely changed in four decades and what it takes to get it into the hands of patients who need it.

My name is Valencia Koomson. I am a professor of electrical and computer engineering at Tufts University. And I'm also the founding director of the Advanced Integrated Circuits and Systems Lab at Tufts.

You have a really fascinating background, engineering and computer science degrees from MIT, a PhD from Cambridge. You're a senior member of the National Academy of Inventors and you hold a patent, you have another one pending. I wanted to start though with before you became a professor, what's the earliest memory that you have of being fascinated by how things work?

Well, when I was very young, at Christmas time, for example, my brother would get a lot of fun electrical circuit sets to play around with, whereas I would get the dolls. So I used to play around with my brother's electronic set. He had a electronic switchboard where you could play around with resistors and capacitors and antennas. I didn't know what a lot of that meant, but I was always fascinated by how complex systems work.

 I was always asking questions about, especially our TV right? How does this box sit in our living room and we get all these beautiful images and videos, where does it all come from? And I remember, you know, as an elementary school kid in grades one, two, and three, asking those questions. So I was also fascinated by living organisms as well. So I felt engineering was a really perfect way to bring together all of my fascinations and curiosities.

That's a great story. And when you decided to become an engineer, were you following in the footsteps of anyone in your family? So I don't know if your brother's older than you or, if you're the older one.

No, my brother is older, but my parents didn't attend college and I didn't have any engineers in my family at all or scientists. So my parents grew up in the South, they moved to the North in Washington for better job opportunities in the nineteen sixties in the midst of the civil rights movement. So I didn't have very many people in my family who had gone into a postgraduate degree level or had pursued degrees in science or STEM or as academic faculty as well.

Let's talk some more about your work. Can you tell us more about your lab at Tufts and what are the big questions that you and your team are preoccupied with right now?

And one of the core areas of focus for my lab is, you know, how can we build technology that can directly translate into improving people's health and quality of life? And so that's a central theme. And that's a question I'm always asking myself as we conduct research, as we form partnerships with collaborators, as we write proposals for grant funding. And so I started working in the area of biosensors very early in my career. And we wanted to find ways in which we could build systems that could really help to track biometric data and to do it non-invasively.

And we have made such advances over the past six or seven decades in terms of how we can build systems that can communicate over very long distances, over very short distances, very low power, wearable, wireless, you know, these types of technologies are commonplace now. You can buy them off the shelf and they become very affordable. And we typically see them in the domain of entertainment and for fitness tracking. And so I wanted to see, we have quite a bit we can do just in our handheld device, like a mobile phone. You can build a mobile phone that can look at the electrical activity of your heart. It can track your heart rate. It can track blood pressure. It can track so many various vitals. It can track where you are in terms of physical location as well. And so we do so much with these handheld devices. And I wanted to see, how can we take it to the next level and build medical devices that could really be of clinical use.

So my focus on biosensors really just directly translated into our work with pulse oximeters. I mean, when we started to study and learn about some of the longstanding accuracy problems in pulse oximeters, especially for people with dark skin pigmentation. My first instinct was, you know, this is fundamentally a sensing and signal processing problem that we can actually solve, we know how to solve.

During the COVID-19 pandemic, you know, pulse oximeters really came to light and really it just became a household product really, as people were encouraged to buy them, to use them at home, to track their blood oxygen saturation level. We started to look at the device more closely and kind of say, well, it's pretty much stayed the same in terms of its core design for the past four decades. How can we improve upon it? How can we modernize it? And so that was my focus even before the pandemic, but it was, we kind of shifted gears and placed more focus on that device during the pandemic.

So many of us will have seen a pulse oximeter maybe on our finger at the hospital, but we don't necessarily know what's happening under the hood. Can you walk us through the basics? What's the purpose of these devices? How do they work?

So these devices really rely on the fact that a certain protein hemoglobin that carries oxygen through the blood, it absorbs light differently than when oxygen is not attached to that protein. So oxygenated and deoxygenated hemoglobin absorb light very differently. So by modeling those absorption characteristics, the device can infer the blood oxygen level, right? We use red and infrared light. So we use light at two different wavelengths. And we shine it through the skin and we measure how much light is absorbed. And we use that information to try to estimate the level of blood oxygen saturation. It's completely safe and non-invasive. You know, we use the same light that's found something like a TV remote control, right? Some TV remote controls use radio frequency signals. Some of the older ones still use the infrared light. And so we're using the same sort of principles.

So you talked about how pulse oximeters have stayed the same in terms of their core design for about four decades. What's causing the issue with regards to the discrepancy in their performance?

There are many different types of pulse oximeters on the market. There are hundreds of manufacturers of pulse oximeters, all with different levels of accuracy. Some follow what we call the FDA, the Food and Drug Administration standards, which are set by the United States federal government, has standards for clinical use and the amount of error that can be accepted, right? Because it's not gonna be one hundred percent accurate all the time. So there is some error that is acceptable with these devices and they set those standards.

And some of the issues can come about when the devices are not tested on a wide range of users. And we know that melanin is an absorber of light and this range of wavelengths that we use, that we operate pulse oximeters at, so melanin concentration can be one confounding factor in affecting the accuracy. The presence of finger polish on the finger could be another confounding factor. Actually, the blood perfusion level can also be another confounding factor. So all these factors need to be accounted for when testing these devices and evaluating them before they're used in clinical settings.

I always say the accuracy depends on the situation in which it's being used in, right? So if you have a device and you don't test it on a wide range of skin tones, for example, you know, it could lead to a device that does not work on all types of users that will encounter the device. That's the main issue. Also, if you are very sick and your blood oxygen is very low, you know, some devices can have higher levels of error at low oxygen saturation. What some of the studies have shown is that pulse oximeters tend to overestimate the blood oxygen saturation for people with dark skin pigmentation. So what that means is that if you have a certain pigment level of melanin and your blood oxygen saturation is low, maybe eighty-five, which is dangerously low, it can overestimate your blood oxygen to be in the ninety percentage range. Several studies have shown this.

And was it during COVID that the discrepancy in the performance across different skin tones first came on your radar? Was it something that you were already aware of? Like, I'm just curious, like, when did you first learn about this?

Yes, I was aware of the studies before COVID. But again, typically those studies were done using consumer grade devices. In the United States, there are over two hundred or so pulse oximeters that are FDA approved. So there isn't a comprehensive study over every pulse oximeter that you can encounter on the market. We were aware of some studies, some limited studies that had been done with maybe fifty to one hundred patients, very small sample sizes.

I also noticed that as an educator, when teaching biophotonics in the classroom. And I noticed most textbooks when they talk about the absorption spectra of different biological media like water and lipids, hemoglobin, I noticed that melanin was not mentioned in those textbooks. So to me, that kind of raised a little bit of a red flag too, because that is an important biochemical to be studied.

It's an engineering challenge to design a device that, first of all, is based on an empirical formula. We're taking measurements using the device, but then actually the calculation is based on an empirical formula. And it's hard to fit that formula to every person. There are eight billion people on the planet. It’s very different from something like a blood pressure cuff. Even with a blood pressure cuff, if your arm is fairly big, you have to use a different cuff. They have different cuff sizes, for example, right? So there isn't one blood pressure cuff that can fit every person on the planet.

So it's kind of this, there's no one size fit all, I would say, for a pulse oximeter at the moment. For us, it's kind of really thinking, going back and thinking more deeply about how these devices actually work. See, skin pigmentation is a major confounding factor. How can we design a device that can compensate for skin pigmentation? So that's one challenge that we're taking on with our patent, is skin tone compensation.

So let's talk about that. So you've decided to take on this challenge of designing a new type of pulse oximeter. You founded a medical device company to commercialize the technology that you've developed in your lab. What was the tipping point that made you say, I need to build a new device?

Well, I think it was a couple of different factors. One is that, you know, I love a challenge. It's a great way to bring together the expertise in my lab. We have built our expertise in designing very high precision micro and nano scale electrical systems. And if we make things smaller, we can reduce noise and reducing noise also always kind of helps to improve accuracy when designing any kind of electrical system. I mean, so we wanted to use that expertise, our expertise in AI and machine learning techniques to help us with this problem. And also in basically how we actually process the signals that we actually receive from the device. And so we have just sort of applied a number of different techniques, a lot of my skills in radio frequency system design, right?

You know, we use cell phones and cell phones pick up radio frequency signals that travel from very, very long distances and have to travel through clouds and across buildings and through various, various structures to get from one point to another. And it has to do it in a very noisy environment.

There's a lot of our radio frequency activity going on in our environment at many different frequencies, right? And those systems work very well without colliding with each other, right? You can talk on your cell phone, sitting right next on the train with someone else, and your phone calls don't collide, right?

So we know how to process and build very sensitive systems that can pick up very low amplitude, weak radio frequency signals in noisy environments. And so I use a lot of the core principles in electrical and computer engineering that I've studied over time and I teach in my classes, applying that to solve this particular problem and thinking about it in a new and a fresh way.

We'll be back to the conversation in a moment, but first I wanted to let you know about a podcast that I think you'll enjoy checking out. Overlooked is all about the women's health gap and the real ways that impacts women's lives. It's hosted by Golda Arthur and it features immersive and powerful stories that explain the science of our bodies in a way that everyone can understand. Golda is a fantastic storyteller and I think you'll really enjoy discovering the show. You can search for Overlooked wherever you get your podcasts. Now back to the show.

So tell us what is different about your approach. Tell us a bit more about the device that you and your team are developing.

What's significant about our approach really is that we're trying to think about signal integrity. We're trying to think about signal integrity before we even try to compute oxygen saturation. We're trying to ask ourselves, am I picking up the cleanest high fidelity signal that I can from the body? And so we try to kind of think about from a signal to noise point of view, like how much signal are we processing versus how much noise is there?

So we did lot of advanced signal processing. We have a patent that focuses on the ability to compensate for skin tone, where we know that melanin is an absorber in the same range of wavelengths that we're trying to transmit through the finger. So how do we compensate for the absorption of light as the light passes through the skin, which can encounter melanin?

I think the first thing though is having inclusivity in mind, to start off thinking that we're gonna design a device that works over the full range of skin tones that are encountered on our planet. And I think that's the first core principle, you know, and ensuring that we design a device that can do that.

So pulse oximeters used in hospitals have to be approved by regulators. In the US, what's needed for the device to be approved?

The Food and Drug Administration, or also known as FDA, approves medical devices for use in hospitals and other healthcare settings. And the current standard, which has been in place for a number of years, is that you test on at least ten individuals, and you ensure that out of those ten individuals, or the sample sizes used for testing, you ensure that a certain percentage of the sample size has dark skin pigmentation, which is quite subjective.

And they try to ensure that within the sample size, you have a range of different ages, different genders. And the goal is to try to have a test sample size that reflects the population in a way that will give you a sense of the error performance of the device.

So essentially the pulse oximeters on the market are meeting those standards.

Yes, based on the way the standard is currently written, they do go through this approval process before they're used in healthcare settings. And part of the discussion that happened after the release of Dr. Sjoding’s paper and other studies that came after that, there was a push for more discussion about these regulations. And there were a number of open forums that were hosted by the FDA, which allowed scientists, researchers, medical device manufacturers and just the public at large to come together and to provide comments and suggestions and recommendations about how those standards should change.

And how are you testing the device to ensure that it works on a diverse population?

So we used a industry standard lab that's based in California at the University of California, San Francisco. There is a lab called the Hypoxia Lab and they actually perform testing, what we call a controlled desaturation study. They take healthy individuals and they have them inhale a certain amount of gas that can change their blood oxygen from seventy percent and modulate it all the way up to one hundred percent. So it’s a controlled hypoxia study.

And at that lab, they have instrumentation and they recruit participants to conduct measurements on. So many medical device manufacturers will use this lab in order to generate data that they submit to the FDA for their device approval. They ensure that you actually can meet the FDA standard of having at least 15 % of your test population with dark skin pigmentation.

And I'm curious what your device looks like. Does it look like a standard pulse oximeter that you can clip on the end of your finger? Does it look bit different?

Yes, it does. And I do have a picture that I can provide for you to show, but it's actually a two-part device. There's a part that sits on the wrist, which looks like more like an Apple Watch, for example. And then there's a wire connection to a finger clip. It's a two-part device at the moment, just because we're using off-the-shelf components, which are kind of bulky. But the ultimate goal is to miniaturize it into a single finger clip.

And we can do that by using semiconductor device manufacturing, which is another area of expertise in my lab is building a nano and micro scale electrical systems on a small microchip. So we can take all of the functionality from the wrist watch part of the device and actually integrate it into something that is miniaturized and more wearable and a small compact form factor.

As you approach this challenge of developing any device, what's been the most daunting aspect?

I wouldn’t say it was daunting. I mean, you know, we love a challenge, particularly with designing electrical systems and systems that can be used to capture biometric data. I think it's extraordinary what the capabilities that we have at our fingertips, right, to do this type of work. And so what we wanted to do though is just ensure that, you know, integrity, right, we want to maintain the integrity of the device and making sure that we're thinking about inclusivity throughout.

So we try to meet with sort of wide range of stakeholders, just kind of understanding how to conduct market research, how to understand how hospitals actually purchase medical equipment. And so I think it helped us appreciate the complexity of the whole medical device landscape as an industry, because hospitals don't buy just one component. They buy suites of components. You know, they buy bedside monitors that monitor many different biometrics, not just blood oxygen. And so it's not that simple.

So we kind of, we try to think about a business model for marketing our device that would help us to get to market quickly. We're pursuing a couple of different routes. One is working with a large medical device manufacturer and licensing our technology to that manufacturer, just so that we could have a wider reach in terms of the hospitals that we can reach. Also working with large distributors that have catalogs of medical devices that hospitals use to purchase equipment, getting our product in that medical device catalog. How do we go about that process? And then as another route, going directly to consumers. I think we have a niche market that we can really fill a gap, and that is designing an affordable medical device that can be sold directly to consumers.

I read a stat that just five percent of American engineers are black and less than two percent are black women. You're one of the two percent. Just curious how that shaped the way you approach your work.

I think it gives a certain level of urgency to try to tackle engineering challenges that directly affect people of color. We hear this all the time that we need to be at the table, we need to be in the room. I think this is a clear example of that, being able to bring voice in an engineering design team to say, let's think about it deeply and let's think about a solution that can be inclusive for a wider range of the population.

I understand that researchers at other universities are also developing alternative versions of the pulse oximeter. Is there a lot more funding for this type of research?

I'm privileged to live in a state like Massachusetts, for example, where at the state level, our governor is putting funding towards health equity research. And so I have applied for two or three grants just over the past three or four years. We're seeing more grant funding at the state level, which is focused on women's health and are really trying to tackle health disparities. You know, at the national level, so I've noticed particularly since 2020 and the COVID-19 pandemic, I think there was a renewed interest in putting more funding towards tackling these issues. So yes, I'm really glad to see that we're starting to make some inroads, but there's more that can be done.

For any aspiring engineers and founders listening, what's one key takeaway that you'd want to leave them with?

I think that this is a wonderful time to be an engineer, to be a scientist, because we do have a number of global challenges facing humanity, and we have the ability to solve them. Your talents and skills can be used to really make a difference in the world, and always look for those opportunities and make a difference.

How can people learn more and follow your work?

So I'm on LinkedIn or you can also find me at Tufts on the Tufts website at engineering.tufts.edu/ece/koomson.

Thank you so much for joining me Valencia. It's been a pleasure having you on the show.

Thank you so much, Tosin.

That was Tufts University Professor Valencia Koomson. I've included links in the show notes if you'd like to learn more about her work. For more on this topic, you should also check out my interview with Dr. Michael Sjoding from the University of Michigan, who Valencia referred to in this episode. He led a study examining racial bias in pulse oximeters that was published in the New England Journal of Medicine in 2020. I've linked to that episode in the show notes.

Thank you for listening to this episode of Made For Us. If you liked it, let us know by leaving a rating or review on Apple Podcasts or Spotify. And don't forget to text such a friend or colleague who hasn't discovered this show yet. I'm Tosin Sulaiman, see you next time.