It’s November 2018. A senior quantitative scientist walks into a meeting she’s not invited to. “I’m Blythe,” she says. “I’m new to Flatiron.” It’s my meeting. I have no idea who she is. 

I’d been trying to convince a group of engineers that our next company-wide Hackathon, a three-day stint every quarter where Flatiron employees can work on anything they want, should be centered on patient advocacy work. This effort included buying out a charity raffle to win breakfast with Flatiron’s Chief People Officer, Carol Jensen, who told me both that she supported the advocacy idea (and said that I could’ve just put a meeting on her calendar). 

Once I’d rallied enough support for the idea, I went about “seeding” the patient advocacy theme with established projects that folks could sign up to work on. That’s what I was doing in the meeting that Blythe Adamson, animated intruder, so politely crashed — picking the brains of research oncologists to see if anyone had brilliant research study ideas that could elucidate an unlit corner of medicine for systematically underserved patients. 

We talked about BMI as a risk modifier, disparities in time to treatment based on age, and financial toxicity. We discussed the kinds of research we could do to highlight inequity.

“What if, instead, we could demonstrate what actually makes things better?” There was Blythe, interloper, uninvited guest. She gave an impassioned speech about how, using de-identified data, we could conduct a natural experiment to determine what actually makes things more equitable. Did Medicaid expansion through the Affordable Care Act (ACA)  really improve the racial disparities in healthcare as it promised? “Let’s look at diagnoses with and without Medicaid expansion and test for differences. We’ll do a time series, with a time varying covariate based on when the policy was enacted. We’ll control using states that didn’t expand after the ACA.” 

A month later, my best friend, Blythe, was on stage pitching the project. A few people signed up to collaborate. We sat in a room and got started. We had only three days.

Aaron Cohen, a research oncologist at Flatiron, and Cary Gross, a research oncologist from Yale, refined the research study design with Blythe. Kelly Magee, nurse practitioner, pulled data on expansion vs. non-expansion states. Melissa Estévez, a scientist from the machine learning team, started coding. I dug into the data. We found that for states that chose to expand Medicaid through the ACA, disparities between black and white patients nearly disappeared. We presented our results at the end of the three days. Company-wide votes were cast for the best Hackathon project. We didn’t win.

But did a hackathon loss stop Blythe Adamson, calendar stalker and marvel of unsolicited meeting attendance?  Of course not. She gathered us again to submit our research for consideration at the world’s biggest oncology conference: the ASCO Annual Meeting. To strengthen our work, we recruited another expert, Amy Davidoff from Yale. We further enhanced  our methods. Our conclusions remained unchanged. We submitted the abstract on February 12, 2019.

We felt honored when the work was initially selected for an oral symposium presentation, but our collective jaws dropped when we later received the official invite to present during the plenary, a distinction granted to just four research projects (generally clinical trials) out of 7,000 submissions. Our preparations during the subsequent few weeks involved in-depth discussions, tighter code, and additional sensitivity analyses. I’m confident that Blythe did not actually sleep during those weeks, but coded through the night, drawing mathematical figures on her windows like Matt Damon in Good Will Hunting. In fact, I have the 3 a.m. texts and photographs to prove it. On June 2, 2019, the research was presented in front of tens-of thousands of oncology professionals. 

The response was extraordinary: 50+ stories generated 440,000+ views, and 22,000+ social media shares. Andy Slavitt, who ran the Center for Medicare and Medicaid Services (CMS) from 2015 to 2017, tweeted that our research proved that expanding Medicaid through the ACA was the “right direction.” Senator Dianne Feinstein called it “further proof that the ACA is saving lives and improving care for millions of Americans.” Former FDA Commissioner Robert Califf said “Not surprising. Hopefully, more states will catch on & act on behalf of their people.”

Even more importantly, our study serves as what Dr. Yousuf Zafar, from the Duke Cancer Institute, called a “proof-of-concept that improved access to cancer care is associated with timeliness and reduced disparities,” and an indication that we “can use real-world evidence to identify policy solutions that improve care quality and outcomes.” 

It also serves as proof that good science, fed and nurtured by a remarkable, dogged team, and given the time and space to grow, can change policy. And policy, as we’ve demonstrated, changes lives.

My advice: Buy raffle tickets. Crash meetings. Find your people. Lose hackathons. Ask for help. Keep going. Give people the freedom to do remarkable things.

Author(s)
Lead Clinical Data Analyst
Erin Williams is a lead clinical data analyst at Flatiron Health, where she organizes oncology data into actionable information that can be used to improve patient care and accelerate research. She is also a writer and illustrator.
Back