What rumours is your fitness tracker spreading about you? In its latest Internet of Things themed sweep, the Office of the Privacy Commissioner of Canada reviews what personal information is being collected about Canadians by “smart” health and fitness devices.
Many of us will remember Time Magazine’s audaciously titled September 2013 issue, which splashed the following headline across its cover page: “Can Google Solve Death?”
At the time, there were more than a few skeptics who might have dismissed Google’s investment in Calico, a biotech subsidiary, as another moonshot investment by the tech giant or as part of a long-term expansion strategy.
Fast-forward less than three years. Regulators continue to play catch-up with the burgeoning industry at the intersection of data analytics and user-generated personal health data. The ballooning number of connected devices that make up the so-called internet of things (“IoT”) has accelerated in scale at a heart-clutching rate. The Office of the Privacy Commissioner of Canada (“OPC”) quoting estimates that, by 2020, there will be between 20 and 30 billion connected devices. While devices that generate data specific to the function and use of the human body represent a subset of these devices, it is hard to deny the growth in the sophistication and potential use (and misuse) of the datasets generated from users’ health and biometric data.
Connected health technology has come a long way since the days of telephonic medical alert systems infamously portrayed in infomercials featuring “help, I’ve fallen” pushbutton necklaces. While application driven smart-phones, watches and fitness wearables are top of mind, the healthcare industry has adopted a range of smart devices that quietly gather and amass a steady stream of data about their users: baby monitors, respiratory and glucose meters, scales, pillboxes, thermometers, contact lenses, heart-monitors, and even band-aids are but a few of the previously inert devices that have become IoT-enabled. For individual consumers, health practitioners, and public health officials, there are extremely compelling use cases to prevent regulatory authorities from stifling the innovation in this sector. For individual patients and clinicians, the devices open what was previously a black-box allowing insight into the lives of individuals outside a clinical setting. The data gathered will enable the healthcare industry to open new service lines focusing on early detection and intervention as well as ongoing health monitoring. Similarly, public health authorities can benefit from large-N data-mining that could potentially offer new insights into determinants of disease, healthy aging processes, and general population wellness.
However, without adequate regulation the (mis)use cases for health data are equally compelling. Digital archives of health data represent new targets for data-breaches and fraud. While the OPC singles out harm to dignity and the integrity of the human body as coordinate reasons for the special protection it affords to leaks/exposures of health data, it has also identified two particular concerns that arise from the proliferation of health data: threats to individual users’ future insurability and employability.
IoT Health Devices Raise Multiple Concerns
The cross-border fluidity of data, the proliferation of health data start-ups, the lack of consumer awareness, and the dangers of misuse of health data have not gone unnoticed by the OPC. Last week, the OPC announced that as part of the Global Privacy Enforcement Network (“GPEN”) review of IoT devices, the focus of the OPC’s 2016 “sweep” would be health devices.
In order to build a clearer picture of the practises of Canadian businesses the OPC declared that between April 11th and 15th, 2016 it would be testing products, examining privacy information published on businesses’ websites, and directly petitioning businesses for responses to specific privacy-related questions.
As in previous years, the 2016 sweep is part of a coordinated effort by the OPC as a member of GPEN to increase public and business awareness of privacy rights and responsibilities, encourage compliance with privacy legislation, identify concerns that may be addressed through targeted education or enforcement, and enhance cooperation among privacy enforcement authorities. GPEN is an OECD organization composed of local data protection authorities of certain participating member states. Though members of GPEN do not rule out either further follow-up with, or enforcement action against noncompliant businesses, the OPC has stated that it does not consider the sweep an investigation, nor does it conclusively identify compliance issues or possible violations of privacy legislation through a sweep alone.
Those readers who have been monitoring communication from the OPC will not be surprised by the focus of the sweep. In its June 2015 report, “The OPC Strategic Privacy Priorities 2015-2020”, the OPC identified “The body as information” as one of its four main priority areas and pledged to both “conduct an environmental scan of new health applications and digital health technologies being offered on the market and research their privacy implications”. The OPC has stated that it believes the human body to be the “vessel of our most intimate personal information”, and, as such, will strive to promote respect for its privacy and integrity. In its strategic plan the OPC drew particular attention to biometric data associated with wearables as well as data gathered from direct-to-consumer genetic testing products and services.
Misuse of Health Information
Legislative activity reflects the increasing concern over the use and misuse of health information. Bill S-201, as adopted by the Canadian Senate on April 14th, is expected to provide a measure of much-needed protection against discrimination on the basis of genetic characteristics. In addition to amending the Canadian Labour Code, the Canadian Human Rights Act, the Privacy Act, and the Personal Information Protection and Electronic Documents Act (“PIPEDA”) to protect against genetic discrimination, the bill introduces a Genetic Non-Discrimination Act, which makes it an offence for a service provider to collect or use the results of a genetic test of an individual without that individual’s written consent.
However, even in countries where genetic discrimination protections are already enshrined in law, insurance firms continue to intensify investment in IoT analytics. For instance, as regards health-related services, between 2014 and 2015, there has been an acceleration in the use of data from health and fitness monitors by insurance companies with the percentage of firms having launched or piloted health and fitness IoT-connected insurance initiatives rising from 10% in 2014 to 39% by 2015.
Businesses engaged in either the collection or use of such information should be aware that the results of the sweep will, in the medium term, likely be incorporated in new OPC guidance that will identify standards for privacy protection in products and services as well as new “no-go” zones for data collection. The OPC has identified several horizontal, cross-cutting strategic concerns which it has stated it will apply to IoT devices and services:
- Exploring innovative and technological ways of protecting privacy;
- Strengthening accountability and promoting good privacy governance;
- Protecting Canadians’ privacy in a borderless world;
- Enhancing the OPC’s public education role; and
- Enhancing privacy protection for vulnerable groups.
As regulators continue to fine-tune their approach through information gathered in exercises like the 2016 sweep, businesses that gather, trade-in, or use health data should be monitoring both changes in the regulatory landscape as well as the wider technological environment. For instance, today’s means of de-identifying health data by scrubbing personally identifying information or pseudo-anonymizing individual users through the use of standard cryptographic methods like a hash may not be enough to protect the sensitive data that is at the heart of the OPC’s interest in the “body as information”.
Businesses cannot think of the information that they gather and warehouse as having a ‘static’ risk profile. Over time, as use cases for health data and the analytical tools available to businesses become increasingly mature, we expect to see a commensurate growth in the capability of third parties (or poorly governed business units) to ‘link’ today’s privacy compliant data with offline and online datasets to recreate an identifiable profile of a de-identified or pseudo-anonymous person.
In the more mature retail environment, the Massachusetts Institute of Technology was recently able to use de-identified credit card purchase information from a 1.1 million person dataset to match 90% of the cases to specific publicly available information on social media sites such as LinkedIn, Facebook, Twitter, and Foursquare.
Similarly, businesses should continue to appraise on an ongoing basis both the validity of end-user consents and the capability of users to access and correct health data as new use cases for the data are piloted.