Chapter 7: The Data Problem

Bridging the Gap: The Urgent Need for Data-Driven Solutions in Disparities Research

This article is brought to you by:

The Labora Collective 🌱

Where innovation meets advocacy. Where your voice shapes the future of women’s health.

---

So what does this have to do with disparities research? Well, not much—at least not at first. The only connection is to say that I’ve become increasingly interested in the very research I once tried so hard to avoid.

How can we even begin to offer patients equitable care when someone like me—who has played by all the rules, studied hard, gotten good grades, worked myself to exhaustion—isn’t protected?

If I’m not safe, patients have no chance.

At the risk of being pigeonholed, I’ve become genuinely excited about disparities research. The more I learn about how medicine is actually practiced, the more curious the Harvard-MIT-trained scientist in me becomes. For instance, how is it that we’ve been studying the staggering disparities in preterm labor rates for over 15 years and have made NO meaningful progress in closing the gap? No one can even clearly quantify the major drivers of the disparity. After two decades, the best we have is an inactionable list of risk factors—while Black babies keep dying, and their mothers are left devastated. That feels like a data problem. A design flaw.

The scientist in me was offended by the methodological sloppiness that often passed for disparities research. We had decades of studies documenting the problem—African American women dying at three to four times the rate of white women during childbirth, Black infants dying at twice the rate of white infants, persistent gaps that couldn’t be explained away by socioeconomic status or access to care. But when it came to actionable solutions, the research often grew vague and unfocused.

Traditional approaches to disparities research felt like describing the symptoms of a disease without ever trying to identify the underlying pathogen. We would catalog disparities, control for the obvious confounders, and then throw up our hands when the gaps persisted. The studies were descriptive rather than interventional, observational rather than experimental.

They documented suffering without offering tools to alleviate it.

As people prepare to march on Washington to demand more focus on maternal health, I hope some of that energy is used to bring the rigor of science into an arena that has historically drawn little scientific interest—despite its staggering impact on real lives. In the past half-century, we’ve put a man on the moon, brought internet access to half the planet, cured or eradicated deadly diseases, and decoded the human genome.

I’m confident that this is a problem science can solve—with the right will and motivation. The tools already exist. First, let’s measure the problem in a meaningful way. Let’s quantify it. The fact that I can’t walk into my own hospital and determine how much disparity exists in the care we provide is, frankly, unacceptable. I don’t want some abstract, national-level metric—I want to know what’s happening right here, for my patients. How equitable is our care? I’ve asked multiple people. No one knows. No one is even sure if it’s measured. In an era of unprecedented technological advancement, that simply cannot be right. At a Harvard hospital, we don’t track disparities at the local level.

This was perhaps the most shocking discovery of all. Here I was, working at one of the most prestigious medical institutions in the world, surrounded by brilliant minds and cutting-edge technology—and yet we had no real-time data on whether the care we were providing was equitable. We could tell you, within minutes, the exact blood pressure, heart rate, and oxygen saturation of every patient in the building. But we couldn’t tell you whether Black mothers were more likely to experience complications than white mothers.

We had elaborate quality assurance programs that tracked infection rates, readmission rates, and patient satisfaction scores. We had dashboards that monitored everything from hand hygiene compliance to medication errors. But disparities? Those were someone else’s problem, measured at the population level by researchers using data that was often years old by the time it was published.

As a Black patient, I don’t know if I have a higher risk of a bad outcome here, across the street, or in a hospital in another state. But that feels measurable to me.

Imagine you bought a plane ticket and because of the color of your skin, your chance of crashing was higher. Wouldn’t you want to know if American Airlines was safer than Delta when flying your loved ones across the country? At the very least, if I’m spending the same amount (and last I checked, I don’t get a discount on my health insurance even though the care I receive may not be as good as what my white colleagues receive), shouldn’t I have access to the information needed to make an informed choice?

The airline analogy lays bare the absurdity of our system. Airlines track safety data obsessively—not just nationally, but for individual routes, aircraft, and even pilots. If systematic differences in outcomes based on passenger characteristics existed, it would be headline news and trigger immediate federal investigation. In healthcare, we accept disparities as an unfortunate but immovable fact of life.

The problem wasn’t just the absence of data—it was the absence of systems designed to collect and act on that data. Our electronic health records can track thousands of variables, yet race-stratified outcome measures aren’t built into the standard dashboards. Our quality improvement programs are quick to respond when problems affect all patients, but they lack the sensitivity to detect problems affecting only some.

This wasn’t an accident. It was a design choice.

We had built systems that were optimized to maintain the status quo rather than identify and correct inequities. The absence of disparities data allowed everyone to assume that their care was equitable without having to confront evidence to the contrary.

Yet the tools to change this already exist. The same EHRs (Electronic Health Records) that track everything else can be programmed to flag disparities in real time. The same quality improvement strategies that revolutionized patient safety can be applied to equity. The same data visualization tools that clarify complex clinical data for providers can be used to make disparities visible—and actionable.

What’s missing isn’t technical capability—it’s will. The same institutions that mobilize armies of data scientists to optimize billing or reduce length of stay somehow can’t allocate the resources to measure whether their care is literally killing people based on the color of their skin.

I’m not claiming to be an expert in disparities research. I’m not. I’ve spent my career training to be an engineer and scientist. But I’m excited to bring those skills into this new arena. There’s a lot of work to do. I don’t know exactly how to do it yet, but I’m confident I’ll figure it out. I’m also not sure I’ll be taken seriously—but honestly, who cares? Some things are more important than being a Dean at a medical school.

The transition from nanochannels to disparities research felt like coming home to work I was meant to do. The same analytical skills that served me in the lab—the ability to design rigorous experiments, control for confounders, spot patterns in complex data—are desperately needed in equity research.

But this shift is also personal. Every time I walk past one of those “Human First” signs in the hospital hallways, every time I experience or witness discrimination, every time I see maternal mortality stats showing women who look like me dying at unconscionable rates, I feel the pull toward work that matters not just intellectually—but morally.

The science of disparities isn’t “soft science” at all. It might be some of the hardest science there is—demanding not just technical skill, but the courage to ask uncomfortable questions and the persistence to pursue uncomfortable answers. It requires the willingness to see what others refuse to see, and to insist on solutions others find inconvenient.

Who knows? Maybe the next Dean of Harvard Medical School will be a disparities researcher.