Pre-print of published article: Nielsen, M. W., Andersen, J. P., Schiebinger, L., & Schneider, J. W. (2017). One
and a half million medical papers reveal a link between author gender and attention to gender and sex analysis.
Nature Human Behaviour, 1, 791-796. DOI: 10.1038/s41562-017-0235-
x
One and a half million medical papers reveal a link between
author gender and attention to gender and sex analysis
Mathias Wullum Nielsen
1*
, Jens Peter Andersen
2
, Londa Schiebinger
1
, and Jesper W. Schneider
2
1
History of Science, Stanford University, 450 Serra Mall, Stanford, CA 94305, United States
2
Danish Centre for Studies in Research and Research Policy, Department of Political Science, Aarhus
University, Bartholins Alle 7, Aarhus 8000, Denmark
*To whom correspondence should be addressed. Email: mwn@ps.au.dk
Keywords: gender and sex analysis, medicine, gender diversity, women in science, social studies of science
2
Gender and sex analysis is increasingly recognized as a key factor in creating better
medical research and healthcare
1-7
. Using a sample of more than 1.5 million medical
research papers, our study examined the potential link between women’s participation
in medical science and attention to gender- and sex-related factors in disease-specific
research. Adjusting for variations across countries, disease topics and medical research
areas, we compared the participation of women authors in studies that do and do not
involve gender and sex analysis. Overall, our results show a robust positive correlation
between women’s authorship and a study’s likelihood of engaging gender and sex
analysis. These findings corroborate discussions of how women’s participation in
medical science links to research outcomes, and illustrate the mutual benefits of
promoting both women’s scientific advancement and the integration of gender and sex
analysis into medical research.
Despite a burgeoning scholarship, gender and sex differences remain unaddressed
in large parts of medical research. Consider, for instance, the paradigmatic example of
biological sex differences in cardiovascular disease (CVD) – the no. 1 ‘killer’ among adult
populations in the Western world. Since the 1980s, the annual number of CVD-related deaths
has been higher for women than for men; but women continue to be underrepresented as
participants in clinical trials
8-10
. A similar pattern is found in cancer research, where male
research subjects dominate the trials
11
, or in studies of cell tissue and laboratory animals,
where the sex of the subject often goes unrecorded
12,13
. Even in research involving both
sexes, data is often not analysed by sex, and results may therefore not be accurate for either
male or female subjects
14
. Both women and men face disadvantages under these
circumstances. Around one third of osteoporosis-related hip fractures occur in elderly men,
but osteoporosis research tends to focus on women, resulting in osteoporosis in men being
underdiagnosed, undertreated, and underreported
15
.
3
Not only biological sex but also gender (i.e., the social attitudes and behaviours
associated with being a woman or a man) is known to be a critical determinant of human
health
8,9
. Women and men are exposed to different occupational hazards (e.g., ergonomic
demands and psychosocial stressors), and differ on a wide array of health-related lifestyle
behaviours (e.g., exercise, alcohol and tobacco use)
16,17
. Further, gendered psychosocial
factors have been found to predict recurrent outcomes in patients with acute coronary
syndrome and to moderate patient-perceptions of pain
18,19
. Despite robust evidence,
associations between gender, biological sex and health outcomes remain largely neglected in
the literature
20
, with potentially life-threatening and costly consequences. Of the ten drugs
withdrawn from the U.S. market between 1997 and 2000, eight involved health risks for
women that may have been avoided if more attention had been devoted to gender- and sex-
related factors
21
.
These examples demonstrate how gender and sex analysis (henceforth GSA) has
the potential to improve medical diagnosis and treatment. Here we define GSA as scientific
approaches aimed at understanding how social and behavioural differences between women,
men and gender-diverse people (gender analysis) and biological differences between female
and male research subjects (sex analysis) relate to health outcomes.
A growing number of medical scholars and journal editors already acknowledge
the importance of GSA in fostering excellence in medical science and healthcare
1-7
. Science
agencies, too, subscribe to this idea (for an overview of national policy efforts see ref.
22
).
Through the Horizon 2020 program, the European Commission (EC) has committed itself to
“integrating the gender dimension into the research and innovation content (…) in order to
improve the quality of research and stimulate innovation”
23
. Similarly, the US National
Institutes of Health (NIH) recognizes that both gender and sex “play a role in how health and
disease processes differ among individuals,” and have implemented guidelines
4
recommending that “sex as a biological variable will be factored into research designs,
analyses, and reporting in vertebrate animal and human studies”
24
.
Both the EC and NIH prioritize policies: 1) to advance the careers of women
scientists, and 2) to encourage gender and sex analysis in research design. Yet, we know little
about how these policy objectives may be linked. Are women and men equally likely to
integrate gender and sex analysis into their research designs?
While recognizing the crucial role played by funding agencies and scholarly
journals in putting gender and sex on the scholarly agenda, our study is unique in analysing
potential gender differences in medical scholars’ involvement in GSA. If gender variations
are detected in our data, we are not proposing that they stem from innate differences in
women’s and men’s scientific styles or preferences. Instead, we follow tenets in the literature
on cultural diversity, and see gender as a cultural category shaping cognition, experience and
perspective in the workplace, with implications for research interests and focus
25-27
.
Sociological research illustrates how societal norms and expectations operate to
cultivate gender-differentiated career aspirations in higher education and the labour market
27
.
Existing research, for instance, documents clear gender differences in students’ selection of
fields and medical specialties
28,29
; but gender norms and expectations may also spur
variations in researchers’ choice of different forms of research within a given field or
specialty
27
. Indeed, this idea already finds some support in the literature. Using meta-data
from over eight million articles in JSTOR, Jevin West and colleagues demonstrate notable
differences in women’s and men’s primary areas of specialization in a range of scholarly
disciplines spanning the natural sciences, social sciences and humanities
30
. Using data from
Sociological Abstracts, Ryan Light finds that women sociologists are overrepresented in
research areas such as gender, race, family and medicine, while men dominate in political,
comparative and economic sociology
31
. Juan Dolado and colleagues demonstrate notable
5
variations in women’s and men’s primary subfields in economic research: health, education,
welfare and labour economics tend to be popular topics among women, while agricultural
economics, fluctuations/business cycles, general equilibrium, comparative systems, and
corporate finance are popular among men
32
. Finally, recent studies document a clear
overrepresentation of women authors in social science journals on gender, feminism and
sexuality
33,34
.
Given the traditional conception of gender- and sex-related research as a woman-
dominated domain in the academy, we expect women’s involvement in GSA to be more
congruent with prevailing gender norms and expectations than men’s. Broader cultural
influences about appropriate gender-typed work may, in other words, draw a disproportionate
number of women towards (and men away from) this form of research. A pioneering study
finds women medical investigators more likely to address gender and sex in successful
research proposals for the Canadian National Institutes of Health
35
. However, the results
reported in this study are descriptive, and do not adjust for potential spurious associations
resulting from differences in women’s and men’s participation across diseases and medical
subfields.
Our study has been designed to accommodate these potential biases. Using a
global sample of more than 1.5 million medical research papers and adjusting for variations
in the gender-composition of author teams across disease topics, countries and medical
research areas, we compared women’s general participation and share of first and last
authorships in studies that do and do not involve GSA. Our prior conjectures were that:
1) A study’s likelihood of involving GSA increases with the proportion of women
among its authors.
2) A study’s likelihood of involving GSA increases if the first author is a woman.