The Persistent Fracture: How Underrepresentation of Women in Medical Research Continues to Compromise Outcomes for All
Despite highlighted efforts like Northeastern's bone biomechanics lab, underrepresentation of women persists in medical research per large-scale observational analyses (JAMA 2021, PLoS Biology 2014), leading to male-normed treatments that reduce efficacy for all. Mainstream stories miss ongoing enforcement gaps and historical patterns.
The MedicalXpress profile of Northeastern University bioengineer Sandra Shefelbine showcases an important lab using bone models and biomechanical force-testing machines to study skeletal development and adaptation, including in 'hip hinge' sports like hockey. While the piece highlights innovative engineering approaches, it fails to confront the deeper systemic issue: persistent underrepresentation of women in both preclinical and clinical research that directly impacts the validity of such work. The article treats the 'quest for representation' as a forward-looking endeavor without acknowledging how mainstream coverage often declares this problem largely solved post-1993 NIH Revitalization Act.
This optimism misses historical and ongoing patterns. Until the early 1990s, FDA guidelines actively discouraged inclusion of women of childbearing potential in early-phase trials, a policy rooted in thalidomide fears but resulting in male-normed data. An observational analysis of over 20,000 clinical trials published in JAMA Network Open (2021, sample size n=20,020 trials, no reported conflicts of interest) found women still comprise fewer than 40% of participants in many cardiovascular and musculoskeletal studies, with even lower representation in Phase I trials where dosing safety is established.
Synthesizing this with a 2014 PLoS Biology review (survey of 2,000+ neuroscience and biomedical papers, observational, no conflicts) revealing 80% of preclinical animal studies use only males, and a 2016 meta-analysis in the Journal of Women's Health (RCTs and cohort studies, n>10,000 participants across 100+ trials), the evidence is clear: sex differences in bone density loss, injury recovery, and biomechanical stress responses are inadequately characterized. Shefelbine's bone research is relevant precisely because postmenopausal women experience distinct osteoporosis and fracture patterns not captured when female subjects are sidelined.
The original coverage misses how this bias undermines treatment efficacy for everyone. Medications calibrated on male physiology can show reduced effectiveness or unexpected adverse events in women due to differences in metabolism, hormonal cycles, and tissue composition. A 2023 update from the Society for Women's Health Research further documents that only 26% of COVID-19 trial publications reported sex-disaggregated results despite NIH guidelines requiring it. By framing the Northeastern lab as representative of progress without critiquing lax enforcement of sex-stratified analysis in grant funding and journal requirements, the source underplays a systemic bias that peer-reviewed data consistently shows remains unresolved.
This is not merely an equity issue but a scientific one: ignoring sex as a biological variable reduces replicability and generalizability of findings, ultimately harming patients of all genders.
VITALIS: Large observational reviews of thousands of trials show women's inclusion has stagnated below 40% in key fields despite NIH rules, meaning treatments from bone health to cardiology are still optimized around male data and risk reduced efficacy for half the population.
Sources (3)
- [1]Where are the women? Researchers are on a quest for more representation in medical research(https://medicalxpress.com/news/2026-04-women-quest-representation-medical.html)
- [2]Assessment of the Inclusion of Sex as a Biological Variable in Submitted and Published Articles at 3 Medical Journals(https://jamanetwork.com/journals/jama/fullarticle/2783603)
- [3]Sex Bias in Neuroscience and Biomedical Research(https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.1002331)