A new analysis of two of the nation’s most prestigious medical journals shows that women and people of color rarely served as lead or senior authors of research articles published during the past three decades, and that gains in authorship during that time have lagged behind the increasing presence of these groups within medicine.
In fact, when looking just at the main authors of high-profile journal articles — a metric closely tied to career success — representation of women and people of color appears to have stagnated in recent years, despite an increased focus on racial, ethnic, and gender equity in medicine.
Fewer than 1 in 4 lead or senior authors of articles in the Journal of the American Medical Association and the New England Journal of Medicine were women, while Black researchers likely accounted for between 2% to 7% of these main authors, and Hispanic researchers between 2% to 4%, the study found.
“It wasn’t surprising the numbers were low, what’s more disturbing is the slow rate of change,” said Moustafa Abdalla, a fourth-year student at Harvard Medical School who was the lead author of the study, published last week in the Journal of Racial and Ethnic Health Disparities. “It appears that even in my lifetime, we won’t see change.”
While many leaders pushing for diversity in medicine have criticized the lack of authors from groups underrepresented in medicine who publish in elite medical journals, actual numbers have not been available because author demographics have not been collected or made public by the journals. Authoring research papers is critical for success in the highly hierarchical system of academic medicine: Being a “senior” author connotes a level of status and leadership, while being a “first author” — a designation that often goes to graduate students or junior faculty who do much of a research paper’s heavy lifting — is crucial for getting tenure or job offers.
In statements emailed to STAT in response to questions about the study, the top editors of both journals said diversity in medicine — and in medical publishing — was critical and they were making changes to include more authors from underrepresented groups. Both journals have taken steps in the past year to diversify their editing ranks and accept more papers on issues of health equity.
“Diversity in medicine is important, and we are taking steps where we can to encourage change or highlight inequities. For example, in September 2021 we published an editorial that said that having better diversity among researchers is one way to help make trials more representative,” Eric Rubin, NEJM’s editor-in-chief said.
The paper, Abdalla said, was rejected at both JAMA and NEJM. (Both journals said they could not confirm or deny this due to the confidentiality of the publishing process.)
Abdalla’s project was not initially focused on race and gender, but began in 2019 as a history of science project. Abdalla, who earned a Ph.D. in computational systems and machine learning before entering medical school, used those skills to create a novel dataset of nearly half a million research papers published in JAMA and the NEJM over the past 200 years as part of a project to create a searchable database that would allow scholars to track trends in the use of techniques and terms in medicine over time.
Then in 2020, as issues of equity roiled the nation and the medical establishment, it occurred to Abdalla and his co-authors that they could use the database to track the gender and race of article authors. “We realized we had the dataset to add a voice to this important conversation,” he said.
But the work showed how difficult it is to analyze both race and gender, because these data have not been historically collected by journals. Since authors do not self-identify their race or gender, the team used various methods to arrive at estimates of authors’ race, ethnicity, and gender.
For the gender analysis, the researchers determined gender by using first names, excluding authors with ambiguous names, such as Jamie or Pat. Abdalla said the technique was admittedly limited because it focused only on male and female and disregarded other gender identities, but he thought it was an important first step in collecting gender data.
While women make up 36% of the nation’s physicians (and 54% of current medical students), the average percentage of women who were senior authors in JAMA or NEJM from 1990 to 2020 was just 19%, while the average percentage of lead female authors was 24%. The numbers have not increased substantially since 2000 despite a large increase in women in the physician workforce.
Determining the race of authors was more difficult. “It’s an incredible challenge,” said Abdalla. “Any approach is an imperfect approach unless people self-identify and say, ‘Here’s my racial group, here’s my ethnic group.’”
The authors used two methods to provide a plausible range for authors’ racial identity. In the first, they used Census data that gives the odds someone with a certain last name is of a certain race. For example, of people with the last name Abdalla, 24% self-identify as Black, 55% self-identify as white, and 16% identify as mixed-race. They used these data to assign racial “points” to each last name to determine overall percentages of the races of authors. Abdalla called this method a “best-case scenario” that likely overestimated non-white authors, since medical journal authors are mostly white.
The second method, or “worst-case scenario,” used the Census data to assign a single racial identity to each last name. Abdalla said this approach could have missed some non-white individuals with predominantly white last names but was likely to overestimate the total number of non-white authors. To be conservative, the authors presented data using the best-case scenario, but noted that there were far fewer authors of color when the worst-case scenario was used. “It doesn’t matter how you chop it,” Abdalla said. “The results are concerning.”
The analysis found that the percentage of Black lead authors in both journals averaged 7% and the percentage of Black senior authors averaged 7.3%. Of concern to the researchers, the analysis showed the number of Black authors remained virtually unchanged since 1990 at NEJM and decreased over time at JAMA. (Under the worst-case scenario, the number of Black lead or senior authors was below 2%.)
Numbers were stark as well for Hispanic authors. While 6.8% of the nation’s physicians are Hispanic, the average percentage of Hispanic lead authors at both journals was just 4% and the average percentage of senior authors was 3.7%, a number that has remained basically unchanged over the past three decades at NEJM. (Under the worst-case scenario, the percentage of Hispanic authors was roughly one percentage point lower.) The authors called the number of authors of American Indian or Alaska Native descent “extraordinarily low,” hovering at less than 1% with no increase in the past three decades.
Authorship rates were lowest for women of color: The fraction of articles with Black women as lead or senior authors in both journals ranged between 0.3% and 3.6%, and for Latina researchers, it was between 0.4% and 2.5%.
For all racial and ethnic groups, the rate of increase in authorship between 1990 and 2020 was nearly zero.
The numbers come as no surprise to Melissa Simon, a Latina physician and health equity researcher who serves as vice chair for research in the department of obstetrics and gynecology at Northwestern University’s Feinberg School of Medicine and has been outspoken about bias in medical publishing.
While the methodologies to assess both gender and race had flaws, she said the overall finding of low authorship among women and people from underrepresented minority groups seemed valid. The low numbers, she said, likely reflect the dearth of women and people of color not only in academic medicine, but especially among researchers who lead major studies, such as large, multisite clinical trials that top journals are eager to publish.
“It’s really rare for a female or someone from an underrepresented group to lead these kinds of studies,” she said.
A second issue, she said, is that women and people of color are often shut out from the inner circles of editors who reach out to researchers they know to solicit submissions. “JAMA and NEJM are not just passive acceptors of content,” Simon said. “They actively reach out to some investigators.”
Ray Givens, a cardiologist at Emory University who has done his own painstaking analysis of the race of journal editors, said he had some concerns about the accuracy of the numbers in the new study given the need to estimate racial identities, but he understood how difficult such work can be. “I’m glad that more researchers are taking up this kind of meta-research, turning a lens on which subjects are represented in studies, who gets to report the findings, and who decides which studies are greenlighted for publication,” Givens told STAT.
Givens said he found the dearth of Black authors in JAMA particularly painful, given the American Medical Association’s history of excluding Black doctors from membership for a century and refusing to insist on hospital desegregation. JAMA is owned by the AMA, but is editorially independent. “AMA and the JAMA Network have a particular debt to pay, which they haven’t even begun to address with decades-overdue apologies, declarations of racism as a public health crisis, or bland equity statements,” he said.
What’s needed at both JAMA and NEJM, he said, are more diverse leaders, but also more diverse authors and reviewers who can then “become a more diverse slate of editors who disrupt the old networks that have been responsible for poisoning the literature.”
Last year, the handling of race in JAMA’s network of journals came under fire after the host of a JAMA podcast dismissed the issue of racism in medicine because it was illegal and a related tweet said, “No physician is racist.” The issue led several editors to step down, including JAMA editor-in-chief Howard Bauchner.
In a statement provided to STAT, Phil B. Fontanarosa, JAMA’s interim editor-in-chief, called the new study an ambitious undertaking and said “JAMA is committed to promoting diversity, equity, and inclusion (DEI) in scientific publishing, and as reported in a March 8, 2022 editorial, has multiple ongoing initiatives and activities.” Those include: creating a special collection of more than 1,000 articles the network of JAMA journals has published on issues of race, racism, and racial and ethnic health disparities; actively working with 52 other publishing organizations to determine how best to collect and incorporate self-reported diversity data about authors and reviewers in journal manuscript and peer-review systems; and appointing 11 diversity, equity, and inclusion editors at JAMA journals.
NEJM does not ask authors to self-identify their race, ethnicity, or gender, said Rubin, but like JAMA, the journal is also working with a group of publishers, led by the Royal Society of Chemistry, to develop best practices to encourage author diversity.
While the new research highlights the lack of diversity in academic medical publishing, both Givens and Simon called out the authors for contributing to the problem by not making their team more inclusive. The papers’ first three authors are siblings who identify as Egyptian or North African; medical student Moustafa Abdalla recruited his brother Mohamed, a Ph.D. candidate in computer science at the University of Toronto, and his sister Salwa, an undergraduate studying computer science, to work on the project. Other authors included David S. Jones and Scott Podolsky, white physicians and researchers from Harvard’s History of Science department and medical school.
“Intentionality is required in all aspects of everything we do to improve DEI,” Simon said. “If you are leading work in this area, you’d like to see that reflected in the team. You’d like to see them walk the walk.”
Moustafa Abdalla said that he identified as being part of a group underrepresented in medicine, but that he agreed with Simon’s view, and said a more detailed follow-up paper now underway includes more women and people of color as co-authors. He said he hopes the work will continue to hold JAMA, NEJM, and other medical journals accountable, despite the potential risk to his publishing career.
“Here I have put two prominent journals on the spot. It’s the same as redlining. It’s the promulgation of a system that’s unfair,” he said. “Am I worried about my career? Maybe. But I think this is important. I want to make the path for those following me easier.”
This is part of a series of articles exploring racism in health and medicine that is funded by a grant from the Commonwealth Fund.