Buy XStore

In Silico Trials in Europe

Navigating Data Protection and Regulatory Landscapes

8-minute read!

The digital revolution has transformed every aspect of our lives, including biomedical research. A major leap forward is the rise of in silico trials—computer simulations that replicate how drugs or medical devices interact with the human body.

Imagine a lab without test tubes or microscopes, replaced by complex algorithms and virtual models of human physiology. This isn’t science fiction; it’s the new reality reshaping medical research. Yet, as these digital breakthroughs redefine healthcare, they also introduce a critical challenge: navigating the evolving regulatory landscape, particularly in the European Union, where striking the right balance between innovation and robust legal frameworks is essential to unlocking the full potential of in silico trials.

Synthetic Data: A Solution to Privacy Concerns?

Developing precise and dependable digital models relies heavily on vast amounts of data, frequently gathered from clinical trials, electronic health records, and wearable gadgets. Nevertheless, utilizing actual patient information poses significant privacy issues, particularly under the European Union’s General Data Protection Regulation (GDPR). The GDPR imposes stringent regulations on the handling of personal data, encompassing all details about a known or recognizable person.

To address these issues, researchers are turning to synthetic data – artificially created datasets that replicate the statistical characteristics of actual data while preserving anonymity. Kokosi and Harron (2022) state that synthetic data can retain essential statistical characteristics such as variable distributions and correlations, all while safeguarding personal privacy.

However, is synthetic data not covered by the GDPR? The solution is not simple. According to Recital 26 of the GDPR, data that has been anonymized, meaning it no longer allows individuals to be identified, is not covered by the regulation. However, pseudonymized data—where identifiers are replaced—still falls under the GDPR, as individuals can potentially be re-identified using additional information.

Furthermore, the European Data Protection Supervisor (EDPS) cautions that synthetic data could pose re-identification risks, especially when merged with other datasets. Therefore, although synthetic data can mitigate privacy concerns, researchers must comply with GDPR.

Representation and Bias across In Silico Models

Relying on synthetic data or in silico models raises questions about capturing the full complexity and diversity of real patient populations. Certain groups may be underrepresented, leading to biased or inaccurate results. Gemmati et al. (2020) emphasize the importance of considering sex and gender differences in biomedical research to ensure that results apply to all segments of the population.

Throughout in silico trials, adhering to the principles of data minimization and accuracy is crucial. The GDPR requires personal data to be adequate, relevant, and limited to what’s necessary (data minimization), as well as accurate and up-to-date. Building models on high-quality, representative data is essential for both regulatory compliance and scientific validity.

Navigating the Regulatory Landscape

The rapid evolution of in silico trials creates challenges for existing legal frameworks. While the GDPR provides robust data protection standards, the distinctive intricacies of virtual testing require careful attention. Furthermore, regulations for medicinal products, medical devices, and artificial intelligence (AI) are adapting to accommodate these emerging technologies (Biasin, 2023).

Medicinal Products

The development of medicinal products is being fueled by artificial intelligence and in silico techniques. AI is employed in various stages, including drug discovery through forecasting molecular interactions and pharmacokinetics, along with in clinical trials to enhance patient populations and offer evidence for regulatory evaluation (Hines et al., 2023).

The European Medicines Agency (EMA) recognizes the potential of these novel methods. Through its Innovation Task Force (ITF), it offers scientific advice and encourages dialogue with developers on emerging therapies and technologies. However, challenges persist. The lack of specific guidance on reporting, verification, and validation of in silico models makes it difficult for developers to integrate in silico evidence into their submissions (Musuamba et al., 2021).

Medical Devices 

Similarly, the Medical Devices Regulation (MDR) and the In Vitro Diagnostic Regulation (IVDR) are starting to recognize the importance of computer modeling and simulation. The MDR mentions manufacturers may include simulated use testing and computer modeling in their technical documentation (MDR Annex II). Nevertheless, the EU lacks a specific regulatory framework for in silico trials. In contrast, the U.S. Food and Drug Administration (FDA) has established regulatory pathways for in silico methodologies, providing advice on model verification and validation (Pappalardo et al., 2022).

Work is being done to harmonize EU regulations with global standards. Efforts such as In Silico World and groups like the Avicenna Alliance push for the adoption of guidelines comparable to the ASME V&V40 standard to evaluate the credibility of computational models used in medical devices.

Artificial Intelligence 

Since in silico trials often utilize AI to simulate biological processes, they fall under emerging AI regulations like the proposed European Union Artificial Intelligence Act (AI Act).

The AI Act proposes classifying AI systems based on risk levels, with high-risk systems subject to stricter requirements. AI systems utilized in medical equipment or for making healthcare decisions may be classified as high-risk, necessitating adherence to criteria for transparency, accountability, and robustness. Even though the AI Act is still under negotiation, its provisions could impact the development and deployment of AI-driven in silico models.

Nonetheless, the AI Act provides exemptions for AI systems created for scientific research and development purposes. This implies that certain requirements may not apply to in silico trials that are focused solely on research. However, once these AI systems are placed on the market or used beyond research, they must adhere to the regulations outlined in the AI Act (Biasin, 2023).

The Debate on Data Ownership

An emerging topic within in silico trials is data ownership. While the GDPR defines roles like data controller and data processor, it doesn’t explicitly address data ownership. This gap has led to confusion, especially when entities consider sharing datasets for model training or evidence generation.

Common phrases like “the hospital owns the data” or “the patient owns their data” are legally imprecise under the GDPR. Hospitals typically act as data controllers, determining how personal data is processed. Patients have rights over their data—such as access, correction, and deletion—but don’t “own” the data in a property sense.

In addition, data ownership introduces complexities related to property rights and intellectual property law. Some advocate for recognizing personal data as an economic asset, granting individuals or organizations sole rights to profit from its utilization. However, the EDPS warns against considering personal information as something that can be bought and sold, stressing that it should not be commercialized in a way that undermines individual rights (EDPS, 2017).

The Data Act and similar new legislative plans seek to control how data is shared and accessed, which could change ideas about ownership and control. These advancements underscore the importance of differentiating between data controllership, which involves determining how data is processed, and data ownership, which includes exclusive property rights and economic benefits (Ducuing, 2022).

Security and Privacy Risks

Ensuring the security and confidentiality of personal data is crucial, going beyond just meeting legal requirements. The GDPR mandates that personal data must be processed securely to prevent unauthorized access or accidental loss, ensuring integrity and privacy.

Gonzales et al. (2023) emphasize that although synthetic data can be useful, it is not completely immune to privacy violations. There are potential dangers of data leakage through re-identification attacks, especially in cases of rare conditions or small populations. Furthermore, significant dangers are posed by cybersecurity threats. A breach could compromise personal data and undermine the integrity of virtual trials, leading to potentially harmful consequences for public health.

Moving Forward: Collaboration and Innovation

Upcoming changes in pharmaceutical laws could lead to significant advancements for in silico trials, potentially aligning EU regulations more closely with global standards. The European Commission’s Pharmaceutical Strategy for Europe highlights the importance of tackling advancements in digital transformation, including in silico methods and virtual approaches in clinical trials (European Commission, 2020).

Dealing with these obstacles necessitates working together as a team. Scientists, legal experts, and policymakers must collaborate in creating structures that reconcile innovation and respect for individual rights. According to Pappalardo et al. (2022), virtual trials have great potential to advance medical research and decrease risks related to new treatments. Nevertheless, regulatory pathways need to be flexible to support these new technologies.

Incorporating concepts such as “privacy by design” and ensuring transparency can aid in establishing trust with stakeholders by integrating data protection measures into technologies early on. Progress in AI and machine learning can improve the precision and inclusiveness of models, as long as they are created ethically.

Conclusion

The healthcare field is quickly changing, and in silico trials represent a major advancement in improving the effectiveness of medical studies. Yet, reaching their maximum capability relies on effectively maneuvering the intricate legal and regulatory structures in the European Union. Collaborating and following strict data protection principles is crucial to maximizing the advantages of virtual trials while also safeguarding the rights of individuals and societal values.

Despite facing difficulties, scientists are always pushing the limits. By utilizing sophisticated simulations, they are narrowing the distance between existing medical procedures and the upcoming era of AI-powered healthcare, resulting in more precise and customized therapies.

References

  • Biasin, E. (2023). Privacy and Data Protection in In Silico Clinical Trials. In Silico World D9.2 Legal and Ethical Analysis.
  • Ducuing, C. (2022). An Analysis of IoT Data Regulation under the Data Act Proposal through Property Law Lenses. CiTiP Working Paper.
  • European Commission. (2020). Pharmaceutical Strategy for Europe.
  • European Data Protection Supervisor (2020). A Preliminary Opinion on Data Protection and Scientific Research. https://www.edps.europa.eu/dataprotection/our-work/publications/opinions/preliminary-opinion-data-protection-and-scientific_en
  • Gemmati, D., et al. (2020). COVID-19 and Individual Genetic Susceptibility/Receptivity: Role of ACE1/ACE2 Genes, Immunity, Inflammation and Coagulation. Thrombosis and Haemostasis, 120(05), 682–700.
  • Gonzales, A., Guruswamy, G., & Smith, S. R. (2023). Synthetic Data in Health Care: A Narrative Review. PLOS Digital Health, 2, e0000082.
  • Hines, P. A., et al. (2023). Artificial Intelligence in European Medicines Regulation. Nature Reviews Drug Discovery, 22, 81.
  • Kokosi, T., & Harron, K. (2022). Synthetic Data in Medical Research. BMJ Medicine, 1, e000167.
  • Liddell, K., Simon, D. A., & Lucassen, A. (2021). Patient Data Ownership: Who Owns Your Health? Journal of Law and the Biosciences, lsab023.
  • Musuamba, F. T., et al. (2021). Scientific and Regulatory Evaluation of Mechanistic In Silico Drug and Disease Models in Drug Development: Building Model Credibility. CPT: Pharmacometrics & Systems Pharmacology, 10(9), 804–825.
  • Pappalardo, F., et al. (2022). Toward a Regulatory Pathway for the Use of In Silico Trials in the CE Marking of Medical Devices. IEEE Journal of Biomedical and Health Informatics, 26(11), 5282–5290.
  • Vedder, A., & Spajić, D. (2023). Moral Autonomy of Patients and Legal Barriers to a Possible Duty of Health-Related Data Sharing. Ethics and Information Technology, 25, 23.