Where Doctors Go: Using Data to Map Where Physicians Start Their Careers

by Julia Cohen

doctors in hallway
Photo Credit: JazzIRT/iStock

If a patient needed a medical specialist and lived in a socioeconomically disadvantaged neighborhood, the odds are low that one trained at a top medical school would be nearby. A new study from researchers at USC’s Viterbi Information Sciences Institute (ISI) takes a novel, data-driven look at geographic healthcare inequities. Their paper, “Medical School Ranking and Neighborhood Characteristics of Initial Practice Location among Physicians” was published on May 28, 2025 in the JAMA Network Open, a monthly open access journal from the American Medical Association. 

Researchers Mayank Kejriwal, a principal scientist at ISI, and ISE PhD student at USC Viterbi School of Engineering Navapat Nananukul analyzed where newly-licensed physicians choose to begin their careers, and found a striking trend: doctors from elite medical schools are significantly less likely to work in areas with high levels of socioeconomic deprivation, even though those communities often have the fewest medical resources. 

A Question of Access

“We’re data geeks,” said Kejriwal, a Research Assistant Professor of Industrial and Systems Engineering at USC Viterbi. “We’re always looking at interesting government datasets.” It was when he noticed that a Medicare dataset included information on doctors—including where they went to school and where they work—he began exploring whether physicians were ending up in the communities that need them most, or clustering in places that already have access to care.

To find out, the team combined the Medicare data with a dataset from the U.S. Census Bureau that ranks neighborhoods using the Area Deprivation Index (ADI). The higher the ADI score, the more socioeconomically disadvantaged the area.

They looked at two cohorts of newly licensed physicians, from 2015 and 2020. In both groups, graduates of top 20 medical schools were about 50% less likely to begin their careers in high-ADI areas than those from lower-ranked institutions.

Who Goes Where

The team also looked at specialty and gender. Specialists were consistently less likely than primary care doctors to work in underserved areas. Radiologists, anesthesiologists, ophthalmologists, and surgeons all showed particularly low odds of practicing in high-ADI communities. “That’s concerning,” Kejriwal said. “These are really important medical specialties.”

There were some exceptions. Emergency medicine physicians from top-rated schools were more likely to end up in higher-ADI areas—a trend the team suspects is tied to demand. “Maybe there are simply more emergency services needed in those areas, so there are more job openings,” Kejriwal said.

As for gender, there was a small difference in 2015, with men slightly more likely to work in high-ADI areas. But by 2020, that gap had essentially disappeared.

Why Location Matters

Kejriwal emphasized that the data doesn’t suggest malicious intent: “It’s not that it’s someone’s fault that it’s not equitable,” he said. “It’s just the way it turned out, based on individual decisions.”

Still, the pattern has real implications: doctors from top medical schools often have the most advanced training, and are disproportionately choosing to work in more affluent areas, leaving behind communities that could benefit most from their expertise.

The study didn’t explore why physicians make the location decisions they do, but Kejriwal suspects the reasons are layered. Higher salaries in wealthier areas may help with student debt, which is higher for higher-ranking schools. Prestige likely also plays a role. So do practical considerations like access to amenities, quality of schools, and where people want to raise families. The researchers weren’t able to control for any of that, as information like student debt levels or personal background isn’t available in the data. But those are exactly the kinds of questions Kejriwal hopes future research will be able to address.

What’s Next?

While this particular paper focused on the patterns, Kejriwa and Nananukul are now working to make their tools and datasets available to others who want to dig deeper into questions of healthcare access. “Right now we’re building a knowledge graph based on this data,” Nananukul explained. “Eventually, we want researchers to be able to query it, asking where physicians from a certain school tend to go, or what factors correlate with working in underserved areas. The next step is to make this accessible even to people without programming skills.” 

“This is just the beginning,” added Kejriwal.

Published on May 28th, 2025

Last updated on May 28th, 2025

Want to write about this story?