Assessing the Secondary Use and Scientific Impact of Shared Clinical Trial Data: A Cross-Sectional Study of Clinical Trials Shared on the YODA Project Platform

This cross-sectional study of Johnson & Johnson-sponsored trials shared via the YODA Project demonstrates that data sharing generates substantial, sustained secondary research, with external investigators increasingly contributing over time through diverse analytical approaches that achieve comparable scientific impact to internal studies.

The Gist

Imagine you and your neighbors build a massive, state-of-the-art community garden. You spend years planting seeds, watering, weeding, and harvesting the best vegetables. Once the harvest is done, you write a detailed report about what grew, how much, and what the weather was like. This is the original clinical trial.

For a long time, once that report was published, the garden would sit quiet. But recently, a new idea emerged: "Why not let other people visit the garden?"

This study is about a specific garden (sponsored by Johnson & Johnson) that opened its gates to the public through a special key system called the YODA Project. The researchers wanted to know: If we let strangers into the garden, will they just walk around, or will they actually grow something new? And how does their new growth compare to what the original gardeners grow?

Here is the story of what they found, broken down into simple parts:

1. The Garden Opens Its Gates

The researchers looked at 336 different "gardens" (clinical trials) that were opened up for sharing. They found that these gardens were incredibly fertile.

• The Result: These shared gardens produced 1,167 new research papers (secondary publications).
• Who grew them? About 82% were grown by the original gardeners (the scientists who ran the trial). But about 18% were grown by strangers (external scientists) who got the keys through the YODA Project.

2. The Timing: Who Plants When?

Think of the timeline like a day at the garden.

• The First Few Years (The Original Gardeners): Right after the main harvest, the original team is still there. They are looking at the leftovers, checking for pests, and writing more reports. In the first few years, almost all the new papers come from them.
• The Long Game (The Strangers Arrive): As time goes on, the original team moves on to new projects. But the "strangers" keep arriving! By year 11, the strangers are actually writing more papers than the original team. The garden keeps producing fruit long after the original owners have left.

3. What Did They Grow? (Different Tools, Different Crops)

The original gardeners and the visiting strangers used the same soil (data), but they planted different seeds.

• The Original Gardeners focused on the basics: "Did this specific plant work better than that one?" and "Did anyone get sick?" They were checking the safety and effectiveness of the specific medicine they tested.
• The Strangers were like master chefs and inventors. Because they could combine soil from many different gardens (pooled data), they did things the original team couldn't:
  • They built predictive models (like a weather forecast for diseases).
  • They invented new statistical recipes (new ways to analyze data).
  • They validated tools to see who is at risk for future problems.

The Analogy: If the original team built a car and tested its speed, the strangers took that car, combined it with 50 other cars, and built a whole new highway system or a GPS navigation app.

4. How Famous Was the Harvest? (Scientific Impact)

The researchers checked how famous these new papers became.

• The Strangers' Papers: These were published in more prestigious magazines (higher impact journals) and got a lot of online buzz (social media shares, news mentions). This makes sense because combining data from many trials creates bigger, more powerful studies that sound exciting to the world.
• The Original Gardeners' Papers: These got cited (referenced) slightly more often in official rulebooks (clinical guidelines and policy documents). This is likely because the original team published first, so doctors and policymakers had more time to read and trust their work before making rules.

The Verdict: Both groups did great work. The strangers brought fresh, innovative ideas that got a lot of attention, while the original team provided the steady, trusted foundation that shaped medical rules.

5. Why Does This Matter?

This study proves that sharing data is like lending a library book.

• If you keep the book on your shelf, only you can read it.
• If you lend it out, someone else might read it, write a review, and inspire a whole new story.

The study shows that when companies share their clinical trial data through structured, safe platforms (like YODA), it doesn't just help the original company. It creates a snowball effect. It allows scientists to ask questions the original team never thought of, leading to new discoveries, better predictions, and faster medical progress—all without needing to run expensive new trials from scratch.

In a nutshell: Opening the garden gates didn't just let people in; it allowed the garden to keep growing, changing, and feeding the world long after the original harvest was over.

Technical Summary

1. Problem Statement

While the sharing of individual participant-level data (IPD) from clinical trials is widely recognized as a mechanism to promote transparency, reproducibility, and innovation, there is limited empirical evidence regarding its long-term scientific impact. Key concerns persist regarding:

• Sustainability: Whether data sharing initiatives justify the substantial investment in infrastructure and personnel.
• Utilization Patterns: Whether external investigators (those not affiliated with the original trial) actually utilize the data, and if so, how their research objectives differ from original investigators.
• Scientific Value: Whether secondary research generated by external investigators achieves comparable scientific impact (citations, journal prestige, policy influence) to research generated by the original trial teams.

2. Methodology

The study employed a cross-sectional design to analyze a portfolio of Johnson & Johnson (J&J)-sponsored clinical trials shared via the Yale University Open Data Access (YODA) Project.

• Data Source & Cohort:
  • Trials: 336 eligible J&J-sponsored trials with IPD available on the YODA platform as of December 31, 2021.
  • Inclusion Criteria: Trials must have a full-length, peer-reviewed primary publication (reporting primary endpoints) by the original investigators.
  • Exclusion: Trials without English primary publications or where the primary publication could not be identified.
• Secondary Publication Identification:
  • Search Strategy: Citation searches of primary publications in Web of Science All Databases (conducted Sept–Dec 2024; updated June 2025).
  • Screening: A team of five reviewers screened 135,011 records. Inclusion required the use of IPD from the included trials. Exclusions included commentaries, non-peer-reviewed papers, and studies using only summary-level data.
  • Classification:
    • Internal: Authored by at least one original trial investigator or an employee of J&J.
    • External: No affiliation with the original trial team or J&J.
• Data Extraction & Metrics:
  • Variables: Research objectives (e.g., predictive modeling, comparative effectiveness), analysis type (single trial vs. pooled), and data access mechanism.
  • Scientific Impact Metrics: Journal Impact Factor (JIF), annual citation count, annual Altmetric Attention Score (media/social attention), annual Mendeley reader count, and citation in clinical guidelines/policy documents.
  • Normalization: Metrics were standardized to "annual" rates to account for publication age.
• Statistical Analysis:
  • Used Mann-Whitney U tests for continuous impact metrics and Chi-square tests (with Monte Carlo simulation for small cell counts) for categorical variables.
  • Significance level: $P < 0.05$.

3. Key Contributions

• Comprehensive Scope: This is one of the largest evaluations of secondary research generated from a single industry-sponsored data sharing platform, covering 336 trials and 1,167 secondary publications.
• Temporal Dynamics: It provides a longitudinal view of how the proportion of external vs. internal research shifts over time (up to 27 years post-primary publication).
• Differentiation of Research Objectives: It systematically categorizes the distinct research agendas of internal vs. external investigators, moving beyond simple publication counts to analyze what is being studied.
• Impact Benchmarking: It offers a nuanced comparison of scientific impact, revealing that while external research achieves high journal prestige and media attention, it lags slightly in long-term citation accumulation and guideline integration compared to internal research.

4. Key Results

Publication Volume and Timing:

• Total Output: 265 of 336 trials (78.9%) generated at least one secondary publication, totaling 1,167 secondary publications.
• Internal vs. External: 958 (82.1%) were internal; 209 (17.9%) were external.
• Temporal Shift:
  • In the first 5 years post-primary publication, 100% of secondary publications were internal.
  • The proportion of external publications increased steadily over time, exceeding 50% of all secondary publications by Year 11.
  • Internal publications appeared significantly sooner (median 3.3 years) than external ones (median 9.9 years).

Research Characteristics:

• Analysis Type: External publications were significantly more likely to be pooled analyses (72.2% vs. 55.7% for internal).
• Research Objectives:
  • External: Focused heavily on predictive/prognostic modeling (51.7% vs. 33.6%), development of statistical models/algorithms (28.7% vs. 11.9%), and validation of existing methods.
  • Internal: Focused more on comparative effectiveness (58.8% vs. 39.7%), adverse event characterization, and non-primary endpoint assessments.
• Data Access: Among external publications reporting access mechanisms, 84.7% obtained data via formal sharing platforms (mostly YODA), rather than direct requests to investigators.

Scientific Impact Metrics:

• Journal Impact: External publications were published in journals with higher median Impact Factors (6.7 vs. 4.6; $P=0.002$).
• Altmetric Attention: External publications had higher annual Altmetric scores (2.1 vs. 0.6; $P<0.001$), likely due to the novelty of pooled analyses and larger sample sizes.
• Citations: External publications had lower annual citation counts (2.7 vs. 3.4; $P<0.001$).
• Guideline/Policy Influence: External publications were less likely to be cited in clinical guidelines (11.4% vs. 29.2%) or policy documents (7.6% vs. 25.6%).
• Readership: No significant difference in annual Mendeley reader counts.

5. Significance and Implications

• Extended Scientific Lifespan: The study demonstrates that clinical trial data remains a valuable resource for decades. While original investigators dominate early post-trial research, external investigators drive a sustained, growing volume of research over the long term.
• Complementary Value: Internal and external investigators pursue complementary goals. Internal teams refine efficacy and safety profiles, while external teams leverage pooled data for methodological innovation, risk prediction, and generalizability.
• Validation of Data Sharing Models: The findings support the efficacy of structured data sharing platforms (like YODA) in facilitating independent research. The fact that most external access is mediated through these platforms rather than direct negotiation suggests they are critical infrastructure for scientific progress.
• Policy Recommendations: The authors argue that despite some lag in guideline citations (potentially due to publication timing), the high journal impact and methodological innovation driven by external researchers justify continued investment in data sharing infrastructure.
• Limitations: The study is limited to J&J trials (though diverse in therapeutic area) and relies on citation-based identification, which may miss non-citing secondary studies. Additionally, the "annual" citation metric may not fully capture the cumulative impact of older internal papers.

Conclusion: Structured clinical trial data sharing generates substantial, sustained, and scientifically impactful secondary research. External investigators play a crucial role in extending the utility of trial data, particularly in methodological development and predictive modeling, achieving a scientific impact that is comparable to, and in some metrics (journal prestige, media attention) superior to, that of the original trial investigators.