When clinical research fails to capture who is actually enrolled, the resulting evidence base skews toward populations that may not reflect real-world patients — and treatments validated in narrow groups can perform unpredictably across diverse communities. A large-scale audit of the ClinicalTrials.gov registry now quantifies exactly how pervasive this blind spot remains, even after years of regulatory pressure to improve representation.
Analyzing 58,163 completed studies with posted results — representing a mean enrollment of roughly 1,215 participants each — researchers found that 44.8% reported neither race nor ethnicity data to the registry. Progress is real but uneven: combined race-and-ethnicity reporting rose sharply from just 7.4% of studies in 2013 to 54.6% by 2024, suggesting that policy signals are having an effect. Notably, observational studies were significantly less likely to report demographic breakdowns than interventional trials (OR 0.55, 95% CI 0.49–0.61). Phase 4 trials — the post-market studies that monitor real-world drug performance in broad populations — showed the lowest odds of reporting among all trial phases (OR 0.32), a counterintuitive gap given that these studies enroll the most heterogeneous patient groups.
The findings arrive at a moment when FDA guidance and NIH diversity mandates have raised expectations for demographic transparency, yet implementation clearly lags behind intent. Phase 4 underperformance is especially concerning: these trials are precisely where differential safety signals across racial and ethnic subgroups are most likely to surface. The study's cross-sectional design cannot assess whether missing data reflects non-diverse enrollment or simply unreported diversity, a distinction with very different remedies. Still, the scale of the dataset — spanning 15 years and nearly 60,000 studies — makes this one of the most comprehensive audits of reporting practices to date. For clinicians and patients, it reinforces a caution already familiar in evidence-based medicine: population-level efficacy data may not translate uniformly, and the gaps in what gets reported make that uncertainty harder to quantify.