Insights From the Stakeholder Day

This blog post covers our recap and insights of the stakeholder day held on the 25th of March 2024.

Purpose of the Day

We had the great pleasure to welcome around 20 stakeholders from various governmental institutions, private companies, and universities on March 25, 2024, in Olten.

The purpose of the gathering was fourfold:

a) to present our ongoing SNF Bridge Project,

b) to address their practical issues regarding anonymization,

c) to explore potential collaborations, and

d) to introduce them to the Swiss Data Anonymization Competence Center (SwissAnon) while seeking their expectations regarding the center’s capabilities.

This blog post will cover exactly these parts. The slides can be found here.

Presentation of the SNF Bridge Project

Analyzing data and using modern AI techniques is crucial, but access to data is often limited due to data privacy concerns regarding applications that gather and handle person-related data.

While there is a lot of research on the anonymization and synthetization of cross-sectional data, research lacks to provide solutions for data with longitudinal information, observing the same individuals over time. Longitudinal information allows for compliant, in-depth profiling and adds significant value. Simply removing direct identifiers like names and addresses isn’t enough. We need advanced methods for anonymizing and generating synthetic data, as well as assessing the risk of identification, to safeguard individuals’ privacy.

This project aims to develop innovative solutions for anonymizing longitudinal data. By conducting complex simulations, we’ll evaluate the risks and benefits of these methods, advancing towards a more secure data-driven future.

Practical Issues Regarding Anonymization

Following central topics were discussed during the day:

The stakeholders have a general interest in publishing data like microdata to the general public or to researchers. For instance, credibility of scientific research is increased when findings are reproducible, which necessitates to publish the related data in an anonymized form. The anonymization itself often relies on methods from statistical disclosure control that are applied on the original data. The signing of a data usage agreement (DUA) is not a substitute for data anonymization, as it only relaxes the grade of anonymization. The concern was expressed that anonymized data may lack the utility needed, i.e., to have high fidelity data to work with.

Soundness and utility of data

In the realm of data management, both soundness and utility play pivotal roles. Soundness refers to the quality and accuracy of the data, ensuring that it faithfully represents the original information. This is particularly crucial when producing synthetic data, as stakeholders rely on its fidelity for their analyses. On the other hand, utility pertains to the usefulness and effectiveness of the data for various stakeholders. In the context of anonymized datasets, ensuring utility involves meeting constraints and user needs, which requires a deep understanding of the subject matter and specific data objectives. While soundness focuses on the accuracy of the data, utility emphasizes its practical value and relevance to stakeholders’ requirements.

There is a knowledge gap in anonymization practices on the side of data users as well as data providers, for instance, within research ethics committees. Stakeholders highlighted challenges faced by these committees in discerning between pseudonymization and anonymization, leading to ambiguity in their application. Addressing this issue necessitates comprehensive knowledge-sharing initiatives tailored to equip research ethics committees with a nuanced understanding of these concepts and their operational implementations. Additionally, stakeholders expressed frustration over the lack of effective platforms for discussing legal compliance, training opportunities, and consultancy services pertaining to anonymization practices. A proposed solution involves the establishment of a centralized competence center specializing in anonymization to facilitate productive and timely discussions on these matters. Furthermore, stakeholders emphasized the need for homogeneous group interactions, such as focus groups or dedicated forums, to address specific challenges collectively. Overall, stakeholders advocate for diverse forms of knowledge transfer mechanisms, including webinars and blogs, to foster continuous learning and collaboration in the field.

Best practices

In navigating the complexities of handling intricate datasets and diverse use cases, an exclusive reliance on conventional methods may prove inadequate. It is imperative to meticulously assess the subtleties of disclosure risk and data utility. Even ostensibly straightforward scenarios necessitate a meticulous determination of the appropriate risk assessment metrics. Establishing best practices in this domain requires leveraging insights from the collective expertise of the community and drawing upon successful precedents. Sole reliance on data providers, legal professionals, or anonymization specialists for the validation of applied methods is insufficient; instead, collaborative evaluation is essential. It is essential to acknowledge that the challenge transcends technical dimensions, encompassing legal considerations as well. Hence, fostering open communication with legal experts is paramount. The provision of illustrative datasets, real-world use cases, and best practices serves as a valuable aid in this endeavor. Given that legal professionals often lack a scientific grounding in anonymization techniques, tangible examples from practical contexts can significantly enhance their comprehension. Furthermore, highlighting instances of information misuse or underutilization underscores the limitations of current approaches. Moreover, comprehensive consideration of edge cases is indispensable in informing the development of robust legislative frameworks.

Potential Collaborations and Expectations Regarding the Centers Capabilities

Resource hub

Stakeholders anticipate the development of a comprehensive resource hub specializing in anonymization methodologies, providing them with the necessary tools and methods to navigate anonymization challenges and regulatory requirements effectively. A key approach involves offering personalized consulting services designed to cater to the specific needs of individual stakeholders. Through close collaboration and tailored support, the competence center can deliver targeted assistance, facilitating the successful deployment of anonymization techniques while ensuring strict compliance with applicable legal frameworks and regulations.

Training

An additional service involves the provision of training and educational programs by the center to enhance stakeholders’ proficiency in implementing anonymization techniques. This encompasses comprehensive instruction on theoretical frameworks, established methodologies, and cutting-edge tools. By facilitating the dissemination of this expertise, the center empowers key personnel to adeptly utilize anonymization methods, thereby bolstering their effectiveness and ensuring the integrity of data anonymization processes.

Multidisciplinary

Involving various perspectives is essential in tackling the complexities of anonymization effectively. While the core team at SwissAnon possesses expertise in this area, the intricate nature of anonymization necessitates input from diverse fields. Involving experts from legal domains is crucial for navigating regulatory frameworks and addressing privacy-related concerns. Similarly, data scientists can provide valuable insights into the technical intricacies of anonymization techniques, while collaboration with cybersecurity specialists ensures robust protection against potential breaches. Through the assembly of a multidisciplinary team, SwissAnon aims to adopt a comprehensive approach to anonymization, thereby enhancing its ability to tackle the multifaceted challenges inherent in this process.