How major healthcare systems are aligning to unleash the power of data sharing in EMEA
It is widely acknowledged that healthcare data has tremendous potential to improve public health and to unlock significant innovation in both the public and private healthcare sectors. Use cases based on sharing and processing healthcare data, such as machine learning and AI, are numerous.
Recent examples include data-driven algorithms that improve cancer diagnosis, reduce dangerous drug interactions and predict the best treatment option based on a patient’s symptoms, history and demographics.
Because healthcare data is spread across many actors, a key enabler is healthcare data sharing, which is the focus of this paper.
Sharing requires overcoming distinct challenges, including patient privacy, anonymisation of data without loss of relevance, and the harmonisation of data across disparate sources and IT systems.
These barriers are gradually being broken down, driven by a few major national initiatives in Europe (Health Data Hub in France, NHS in the UK, MII in Germany) and in Israel, which we review in this paper.
While regulation of healthcare data sharing is still at an early stage, what opportunities do start-ups and corporations have to launch data sharing projects?
When should they enter the game, and what is missing to seize these opportunities alongside the national initiatives?
The European directive on data sharing leaves healthcare out of its scope
In 2019, the European Commission adopted a new directive on open data and the re-use of public sector information (PSI). It introduces the concept of high-value datasets, held by public sector bodies and public undertakings, which must be made available for re-use [1] by a wider audience. To date, these priority datasets cover geospatial, earth observation and environment, meteorological, company and mobility data.
The current directive notably excludes both healthcare data (avoiding the complexities of personal data) and privately owned data.
All these high-value datasets share an important characteristic: they do not require extensive anonymisation before release, and they are mostly owned by public entities. These are two good reasons why the European Commission did not include healthcare data in the scope of the first directive.
Sharing the precious public healthcare data of millions of citizens at European level takes time. The data is also fragmented, redundant and incomplete, and partly owned by private parties that may not be keen to share it for free. Nevertheless, major recent national initiatives are paving the way for broader data sharing at European level.
The private sector has already faced its share of data management constraints since the introduction of the GDPR (General Data Protection Regulation), and this will not remain the status quo. Mandatory data sharing for private actors, such as PSD2 (Payment Services Directive 2), is a vivid example of how sharing can be imposed on companies (banks, in this case), and the next iteration of the GDPR is also in development.
Public sharing of healthcare data is gaining momentum, and European companies (pharmaceutical companies, medical facilities, insurers, etc.) are not immune to regulated sharing. Healthcare actors may also have good business reasons to start sharing their data proactively.
The asymmetry of access to data in healthcare supports the case for more sharing
Health data falls into two main categories: individual data (patients’ health status, medical diagnoses, payments and reimbursements) and aggregated data (relating to facilities or practitioners, including overall medical payments and reimbursements).
These datasets are generated by different entities, each with different levels of access to the same data. Exhibit 1 summarises data ownership, although each country has its own specificities.
Although healthcare data is often sensitive, use cases show that sharing it can bring many benefits, including better public health outcomes, greater cost efficiency and a better ability to simulate and model healthcare economics.
Selected prospective use cases include:
- Sharing data between public and private insurers, as Malakoff Médéric did with the SNDS datasets as part of the Health Data Hub, would help each party better understand the amount left for patients to pay and improve market access and coverage (e.g. pricing and reimbursement).
- Using health status and test results to conduct retrospective epidemiological studies and comparative outcome assessments [2] would improve understanding of the effectiveness of medical acts and drugs, to the benefit of all three types of actors.
- Diagnoses and prescriptions could help insurers identify inefficient or fraudulent patterns and reduce the cost of claims. Similarly, the operational performance of medical facilities could help insurers steer clients towards more efficient facilities or those with a higher quality of service, strengthening their client relationships. Health practitioners then have an incentive to share these datasets to showcase their strengths and be selected.
- Medical facilities could benchmark one another (as the NHS aims to do), while pharmaceutical companies could use prescription data held by practitioners to assess and support market surveillance.
- Combining clients’ health status with medical payments could give pharmaceutical companies a better picture of the health journey of their drug users, leading to a better understanding of health habits and new market segmentation.
Exhibit 1: the asymmetry of data ownership in the health sector.
Sharing healthcare data is complex because of its personal and sensitive nature
The main challenge of healthcare data sharing lies in the balance between aggregation (to protect personal data) and informational value (many use cases require individual-level information).
Anonymisation is a critical pain point for unlocking healthcare data sharing, if only to comply with the GDPR. Its goal is to prevent any direct or indirect [3] attempt to identify the individual behind the data. It should not be confused with pseudonymisation, which only prevents direct identification of the patient. Pseudonymisation is the approach chosen by the French Health Data Hub, to preserve high informational value.
Full anonymisation systematically reduces the value of a dataset: the more it is anonymised or aggregated, the narrower the range of applications. Finding the right balance is a challenge, and the data sharing platform must handle anonymisation workflows carefully.
Data harmonisation, as illustrated by the German initiative to form consortia of health actors (described in the next section), is another top-priority challenge. Making sure the data has the same meaning and granularity for both sharing parties can be extremely time-consuming, especially for unstructured data, which represents a sizeable share of healthcare data.
Many other challenges stand in the way, technology choices among them. Many voices were raised against the French Health Data Hub for using Microsoft’s Azure cloud, out of fear that the health data of millions of French citizens could be exposed to the United States under the CLOUD Act [4].
This shows how sensitive the subject is, and that clear communication and transparency are essential to avoid a public backlash.
At the same time, innovative technology solutions are emerging, such as the Substra Foundation, recently released by the French start-up Owkin. Substra is an open-source platform based on blockchain, relying on advanced cryptography and machine learning. It offers healthcare actors a unique way to share individual-level data while protecting confidentiality. Apricity, a start-up focused on fertility, is starting to use Substra to secure the processing of sensitive fertility data under the very strict UK regulations.
Embleema, another start-up partnering with large pharmaceutical groups, is also pushing for more healthcare data sharing. It uses the distributed ledger and smart contract features of blockchain to give patients control over how, and with whom, they share their data.
Overview of major healthcare data sharing initiatives in EMEA
Exhibit 2: overview of major national initiatives in EMEA.
Several initiatives have been launched recently, with Israel the pioneer in data sharing and access (see Exhibit 2).
Israel is a powerful example of what centralised data management can achieve: only four HMOs [5] cover all Israeli citizens, and IT systems are interoperable across hospitals. With a genetically diverse population and 20 years of digitised medical records, the country is sitting on a gold mine. In 2018, a $300 million state initiative was launched to further centralise the data and make it available to companies and start-ups. This initiative helped attract Kamet, the AXA-backed start-up studio focused on healthcare, to open offices in Israel. The ties between the four national publicly owned HMOs and Israeli start-ups have no equivalent anywhere else in the world. Among many public-private partnerships, Maccabi (one of the four HMOs) gave the start-up Ibex access to 6 million pathology exams to train its algorithm to provide a second-opinion diagnosis of prostate cancer biopsies.
In Europe, France, the UK, Germany and the Scandinavian countries are leading the way.
In France, the national SNDS database combines national health insurance reimbursements, hospital medical activity and medical causes of death. A year ago, the French government launched an ambitious programme and platform, the Health Data Hub, to open access to this database to researchers and companies. In each call for proposals, projects are selected on their potential to enrich the database and their contribution to the general interest. Although access is still limited to occasional calls for projects, Health Data Hub officials say their near-term objective is a more agile and inclusive process. Many of the ten projects selected in the first call involve private companies or start-ups. Besides the Malakoff Médéric initiative mentioned above, the Hydro project stands out: the start-up Implicity is using SNDS data to build a model that predicts heart attacks.
The Health Data Hub has just launched a second call for projects aimed at AI start-ups.
The UK has opened to the public an anonymised, heavily aggregated version of data (on the population and on facility performance) from both NHS healthcare and some private hospitals, but no other private parties. The NHS has thus taken on the role of centralising public and private data for the sake of transparency, in a top-down approach. This is especially useful for comparing the performance of private and public hospitals.
Germany has also committed to sharing healthcare data through the Medical Informatics Initiative (MII). By investing €160 million by 2021, the German government is supporting IT interoperability and data integration within consortia [6] formed by local hospitals, universities, businesses and non-university hospitals. Beyond interoperability, each consortium works on specific use cases based on its data. The HiGHmed consortium, for instance, is developing an automated early-warning system to detect pathogen clusters in hospitals algorithmically, with many public and research institutions as well as private companies such as Siemens.
The Scandinavian countries have a long tradition of using their health data. In Sweden, some data points go back to before the 1960s, as government healthcare data is systematically collected and linked to education and income data. These datasets are accessible to research organisations and, to some extent, to companies, making it possible to link the care received to long-term socio-economic outcomes at individual level, for example connecting successful early intervention with the ability to find a job or obtain a degree.
The launchpad for health data sharing is set
Sharing healthcare data can create value by bridging gaps between a wide range of medical stakeholders, but this value can only be unlocked with the right sharing framework. As healthcare data exchange grows, companies should seize the opportunity to build robust exchange methods, in both data management and technology.
The remaining barriers, from anonymising personal and sensitive data while preserving its informational value to data harmonisation, are being broken down by new technologies designed to secure healthcare data sharing. The Substra Foundation, recently released by Owkin, is a perfect example.
Based on blockchain technology, it uses advanced cryptography and machine learning to let hospitals share personal data securely. As an open-source project, anyone can use the platform.
Governments have noticed this value, and momentum around healthcare data sharing is building in Europe, driven by ambitious initiatives. In the short term, we expect national initiatives to keep growing in scope and volume. Joining ongoing programmes is a way for companies to explore how to make the most of their data, while strengthening ties with tomorrow’s lawmakers.
The launchpad for private actors to start sharing healthcare data is set, and healthcare players should keep developing their skills in data technology and regulation.
[1] Free of charge, in machine-readable formats and through application programming interfaces (APIs).
[2] Evaluation of the extent to which the desired outcome occurs, and of what influences it.
[3] For example, by merging several datasets to re-identify personal data.
[4] US federal law that can compel US-based technology companies to provide data to the US government on request.
[5] Health maintenance organisations are public health service providers and insurers. Every Israeli citizen must join one of the four HMOs.
[6] Four consortia have been created to date (DIFUTURE, HiGHmed, MIRACUM and SMITH), and others may follow.
