TL;DR
New data protection laws like GDPR and CCPA, intended to safeguard personal information, appear to be driving clinical trials, particularly early-phase and industry-funded ones, to countries with weaker data privacy regulations. This suggests a potential unintended consequence of these laws, shifting research to regions with fewer protections for participants' data.
What Problem Does This Paper Address?
The implementation of robust data protection regulations, such as the EU's GDPR (May 2018) and California's CCPA (effective January 2020), aims to enhance individual control over personal data. While these laws are designed to govern how medical data is collected and analyzed, their impact on the global landscape of clinical trials remained unclear. Specifically, the authors questioned whether these regulations would lead to a shift in where clinical trials are conducted, potentially moving them to jurisdictions with less stringent data privacy frameworks. This is a critical question because clinical trials involve sensitive personal health information, and any relocation could have implications for patient privacy, ethical oversight, and the overall integrity of research.
Methods
The researchers employed a quantitative, longitudinal analysis of clinical trial data and scientific publications. They analyzed two major clinical trial registries: ClinicalTrials.gov (maintained by the US government but encompassing global trials) and the World Health Organization's International Clinical Trials Registry Platform (ICTRP). Data from these registries were collected from January 1, 2011, to July 31, 2021, allowing for a comparison of trial trends before and after the implementation of GDPR and CCPA (defined as May 25, 2018, for analytical purposes).
To assess the impact, they used statistical methods including segmented regression and the Chow test for structural stability to identify significant changes in the number of trials per month around the data protection implementation (DPI) date. They also correlated changes in trial numbers with country-level indicators, notably the civil liberties freedom score from Freedom House, as a proxy for data protection levels. Additionally, they analyzed scientific papers published on PubMed to see if publication trends mirrored the observed shifts in trial locations.
Datasets
- ClinicalTrials.gov: A comprehensive registry of clinical trials worldwide, used to track trial registrations before and after the implementation of data protection laws.
- World Health Organization's International Clinical Trials Registry Platform (ICTRP): Another major global registry used for comparative analysis of clinical trial data.
- PubMed: A database of biomedical literature, used to analyze trends in the publication of scientific papers describing clinical trials.
The study did not introduce new benchmarks but rather analyzed existing, large-scale datasets to observe real-world trends.
Key Insights and Ideas
- Unintended Consequences of Data Protection: The most striking insight is that stricter data protection laws, while well-intentioned, may inadvertently push clinical research to countries with weaker privacy protections. This highlights the complex interplay between regulation and global research practices.
- Differential Impact by Trial Phase: The shift is not uniform across all types of trials. Early-phase trials (Phase 1 and 2), which are often less complex and require less extensive infrastructure than Phase 3, show a more pronounced move to non-data-protected countries.
- "Ethics Dumping" Reimagined: The findings resonate with the concept of "ethics dumping," where research might be moved to regions with lower ethical standards. However, the study suggests this phenomenon might extend beyond just "unethical" experiments to include a broader relocation driven by data privacy concerns, even to relatively developed nations.
- Publication Trends Lag Behind Trial Shifts: The fact that publication trends did not show a similar shift suggests that the relocation of trials might be a phenomenon where researchers in "helicopter research" models conduct studies in less regulated environments without necessarily fostering local scientific collaboration.
Key Results
- Significant Rise in Non-DP Countries: After the DPI date, the number of Phase 1 and Phase 2 clinical trials in countries not adhering to data protection laws rose significantly (P < 10⁻⁴ for ClinicalTrials.gov data). This trend was not observed in countries that did implement these laws.
- Correlation with Civil Liberties: The largest increase in trials occurred in countries with lower civil liberties freedom scores, indicated by a negative correlation (-0.48, p = 0.008) between the freedom score and the increase in trials.
- Industry-Funded Trials Lead the Shift: Industry-funded trials saw a substantial increase (18%) in non-DP countries after DPI, with no significant change in DP countries.
- Interventional Trials More Affected: Interventional trials showed a larger relative change, with a 28% increase in non-DP countries compared to a 3% increase in DP countries.
- No Significant Shift in Publications: Analysis of PubMed data showed no statistically significant change in the trend of published papers following the DPI date, suggesting a disconnect between where trials are conducted and where research is published.
- Phase 3 Trials Less Mobile: Phase 3 trials, being large-scale and long-term, did not show a similar shift, likely due to the complexity and infrastructure required.
Limitations and Future Work
The authors acknowledge several limitations:
- Causality: While the study shows an association between data protection laws and trial relocation, it cannot definitively prove a causal link. The observed changes could be influenced by other concurrent factors.
- Data Registries Imperfect: Both ClinicalTrials.gov and ICTRP have known biases and limitations, and trial registration is often voluntary. The study attempted to mitigate this by using two large datasets.
- Publication Lag: The time lag between trial completion and publication could mean that the observed publication data does not fully capture the impact of DPI.
- Definition of "Non-DP" Countries: The grouping of countries with "adequate protections" with GDPR-implementing countries might mask some nuances.
- Proxy for Data Protection: Using civil liberties as a proxy for data protection, while reasonable, is not a direct measure.
Future work could involve more granular analysis of specific data protection provisions, exploring the motivations of trial sponsors, and investigating the ethical implications of this observed shift in clinical trial locations. Further research into the "helicopter research" phenomenon and its impact on local scientific capacity would also be valuable.
Summary updated