Friday 26 October 2018

Consumers in the age of digital health and AI

The European Consumer Organisation - BEUC - recently published a new position paper on digitalisation in healthcare. The paper comes at the right time as the impact of digital technologies on health products and services indeed continues to grow. The potential benefits are recognized: better access to medical care, more effective prevention, diagnosis and treatment of diseases, support of healthy lifestyles. However, as pointed out by BEUC, the risks are also present. The issues of consumer privacy, security and safety are listed among the most salient ones.

Main insights

The position paper makes reference to the eHealth communication of the European Commission, which focused on three areas: 
  1. citizens' right to secure electronic access and share their health data (improving electronic health record systems), 
  2. improved research, disease prevention and personalised medicine (pooling data resources and using common standards),
  3. digital tools for citizen empowerment and person-centered care (shifting focus from disease treatment to health promotion and well-being, supported by digital solutions such as wearables and mHealth apps).
While endorsing these general objectives, BEUC - as was to be expected - emphasises the need for adequate consumer safeguards. This general observation is followed by a list of more specific principles and recommendations. We sketch some of them further below - for more details we invite our readers to consult the original paper.

Consumers' control over their personal health data (including the right to decide about data-sharing, to access one's own data and to report on possible errors) features prominently throughout the paper. This translates into a call for a diligent implementation of the General Data Protection Regulation with respect to this sensitive category of personal data. Importantly, BEUC argues, data protection safeguards are also relevant in the context of electronic health records and relationship between patients and physicians.

Another eminent topic addressed in the paper are the digital health tools, which, in view of BEUC, should respect the principles of privacy and security by design and by default and remain under supervision of competent authorities. The paper emphasises the connection between security and safety by providing an illustrative example of a hacked pacemaker. An argument is also made for a minimum set of security measures for all digital health connected products, including mobile health applications, as an ex ante market access requirement. Medical Devices Regulation, Radio Equipment Directive and General Product Safety Directive are listed among key instruments to be revisited in this context.

Consumers and artificial intelligence (in healthcare)

An interesting part of the BEUC paper concerns the growing deployment of artificial intelligence in the healthcare sector and in consumer markets more generally. Indeed, the advances in machine learning have made it possible to generate operable knowledge from the previously intransparent data sets. As a result, data contained in health records as well as vast amounts of data produced through our daily use of digital products and services have become an even more valuable resource. The prospect of AI transforming the way diseases are prevented, treated and diagnosed is anything but exciting. A look at the website of IBM Watson Health will give the reader a good impression. However, as pointed out by BEUC, the picture is not always so rosy.

The observations which BEUC makes in this regard largely follow its earlier position paper on automated decision-making and artificial intelligence. Similar issues were also pointed out in the European University Institute's working paper on consumer law and AI published a couple of months ago. They concern, in particular, the growing information asymmetry and power imbalance, the impact on consumers' decision-making capacities, implications for access to essential services and the risk of discrimination. These, of course, are only early contributions and both the extent of indicated problems and the possible remedies must still be investigated. Considering the growing interests in AI of both scholars and policymakers further research is certainly to be expected.