Privacy Enhancing Technologies pick up steam in public and health sectors

In 2020, the US Census Bureau adopted differential privacy to publicly share datasets while adding noise, or computer generated inaccuracies, to ensure no single resident could be identified. 

Experts explained to The New York Times that these inaccuracies – including data points that point to people living underwater – protect residents from attempts to reconstruct where specific individuals and demographics live. At the same time, the aggregated data still supports organisations in making decisions ranging from planning disaster responses to bus routes.

This case study can be found in a recent report published in early 2023 by the United Nations Committee of Experts on Big Data and Data Science for Official Statistics. The United Nations PET Guide compiled case studies of how national statistics offices around the world have begun to adopt privacy-enhancing technologies (PETs), such as differential privacy, to drive the accurate and secure collection of statistical data. 

PETs have emerged as a key piece of the data security puzzle, in light of emerging concerns around data sharing across borders and organisations. GovInsider speaks to privacy-enhancing technology experts within the public and health sectors to better understand how PETs have begun to drive innovation.

Supporting statistical offices in providing better data

Globally, national statistics offices are charged with collecting and publishing statistics relating to questions of demography, society, and economy at local, regional, and national levels. Once data is collected from a range of sources, from surveys to census to private sector data, official statistics can support governments and other bodies in data-driven decision-making.

When privacy guarantees are in place, national statistics offices can share much more information with the public and other stakeholders to make informed decisions, as the U.S. Census Bureau does with differential privacy.

“If we can make data available much more broadly to society in a safe and secure way, that could really help improve public services, from public transportation to green technologies,” says Ronald Jansen, Assistant Director and Chief of Data Innovation and Capacity Branch at the United Nations Statistics Division to GovInsider.

At the same time, data privacy concerns mean that getting private sector companies to share data is a challenge. Through PETs, statistical offices can better access data collected by industry players to improve official statistics.

For instance, Indonesia’s Ministry of Tourism has improved its official tourism statistics by securely cross-referencing the mobile positioning data of tourists moving across networks of multiple mobile operators.

The Ministry was able to analyse movement data without compromising the privacy of customers or the confidential business information of each mobile operator with a Trusted Execution Environment – a PET that allows for the secure analysis of data within an isolated environment.

It would otherwise be difficult to detect tourists moving around the archipelago, explains Matjaz Jug, Head of CIO Office at Statistics Netherlands, to GovInsider.

Alongside presenting such case studies, the guide serves as a playbook for how national statistics officers can incorporate PETs within their work as well as the benefits and limitations of various PETs. 

Some of these limitations include the need to manage public perception and remain transparent about the implementation of such technologies and the benefits they offer society, such that trust in the work of national statistics offices remains high, shares Jansen.  

Enabling regional health innovation

Beyond improving the work of official statisticians, PETs are driving healthcare innovation. The healthcare sector is both data-rich as well as littered with data silos, making it a sector that benefits from the appropriate use of such technologies. 

Traditionally, healthcare data is mostly descriptive or diagnostic, says Michael Andrew, Head of Data Analytics at Zuellig Pharma. As one of the largest healthcare services groups in Asia, Zuellig Pharma uses PETs as a critical foundational tool for its data analytics solutions, enabling secure and near-real-time data processing and analysis across its partners.

PETs can help pharmaceutical companies analyse and understand how diseases progress and interact with associated conditions, such as underlying health conditions, as well as how treatment plans may have to change as a result, shares Andrew. Through the secure sharing and analysis of data, as well as the application of artificial intelligence, healthcare data can move from descriptive to predictive.

For example, through the use of PETs comparing data across multiple pharmacy chains across Taiwan, one of Zuellig Pharma’s clients identified that first line treatment for patients with type 2 diabetes was insufficient, Andrew explains. 

After educating pharmaceutical players on alternative approaches, the client was able to improve patient outcomes and reduce the rate of patients switching out to other drugs by 29 per cent.

Similarly, Jug shared that the Netherlands used PETs to compare medical, insurance, and socio-economic data to analyse the success of treatment plans without disclosing any patient information.

“One hospital might just have a small sample of a specific disease. If you combine that data across 10 different hospitals, you can do better analysis… This consortium idea of combining health records while maintaining privacy is really gaining a lot of traction,” says Jansen.

Managing trust

All in all, PETs can support statistics offices in embracing a zero-trust and privacy-by-design approach, as none of the parties involved in data sharing or distribution can access the original data, Jug and Jansen explain.

They can support agencies in performing decentralised and distributed analyses across datasets while protecting sensitive data. In turn, they also enable agencies to share aggregated and anonymised datasets without fear of compromising personally identifiable data.

It is important to create a community that can support patients in understanding how their data is being used when PETs are applied, says Andrew, so that they know that their data is always encrypted and anonymised.

Singapore’s Government Technology Agency has begun exploring such privacy preserving techniques, with a recent webinar on how such technologies can be a key enabler to build trust in digital government efforts.

A report by the Royal Society United Kingdom has argued that the public sector needs to lead the way in piloting PETs, and highlighted use-cases that could benefit from the meaningful applications of PETs from healthcare to sustainability efforts.

Estonia has also begun exploring the use of PETs within their public services and has issued a procurement call for data professionals to support such efforts, wrote GovInsider previously.

Read more: Tanium partners with Microsoft to provide comprehensive endpoint visibility and management in real time