Digital health, big tech and your privacy
- Private companies, including Big Tech, are responsible for a large number of digital health tools
- In this piece, we outline some of these tools and give some examples of Big Tech involvement in digital health
- While there may be benefits to health, there are also potential threats to your privacy from these initiatives
- Finally, this piece considers what sort of future a Big Tech dominated vision of digital health might hold for your rights, your autonomy and our society
Introduction
Harnessing new digital technology to improve people’s health is now commonplace across the world. Countries and international organisations alike are devising digital health strategies and looking to emerging technology to help solve tricky problems within healthcare. At the same time, more and more start-ups and established tech companies are bringing out new, and at times innovative, digital tools aimed at health and wellbeing.
Wellness apps, consumer wearables and medical AI are almost certainly here to stay. But at PI we’re concerned that these digital tools may not always have been designed with people’s privacy, autonomy, and rights in mind. As a result, people could be being asked (even if inadvertently) to make unfair sacrifices when seeking to improve how they manage and understand their health.
We’ve written elsewhere about the risks and benefits of digital health in general and why health and privacy must not be traded off against each other. In this piece, we take a closer look at some of the specific digital tools that can infringe people’s privacy, and in particular at how four Big Tech companies (Google, Apple, Microsoft and Amazon) are now involved in healthcare. We then discuss how a tech-first approach to healthcare, as potentially prioritised by the tech industry, could have long-lasting and potentially negative consequences for society as a whole.
Types of digital health tools
In this section, we outline types of digital tools that are deployed in healthcare and identify the effects they can have on both health and privacy. The examples listed below are not meant as an exhaustive list of digital health tools (the number of them is growing all the time!) but are examples of tools where there are potential risks of significant impacts on privacy.
1. Remote consultations
It is now common for patients to see their doctor remotely, over the phone or on videocall, rather than in person. This is sometimes called ‘telemedicine’ or ‘telehealth’. Video calls can also be used between health professionals, for example to link providers in remote areas to specialists in city clinics.
Examples
- Phonecalls between patients and their regular doctor
- Apps that offer video consultations with specialist doctors
- Satellite phonelink between remote clinic and specialist professional
Potential benefits to health
- Increased access to services
- Greater availability of specialist advice
- More convenient scheduling of appointments
Potential risks to privacy
- Data breaches
2. Apps and online platforms
Many people use apps and websites to understand and manage their symptoms and health conditions. Apps are also used by healthcare providers for formal consultations and prescription management. Social media platforms such as Facebook and YouTube are even becoming increasingly relied on as sources for health information.
Examples
- Period tracker apps
- Covid contact tracing apps
- Mental health apps
Potential benefits to health
- Can provide targeted and easily accessible information
- Help you stay on top of managing your health
Potential risks to privacy
- Sensitive data being shared with third parties (eg advertisers) without explicit consent
- Law enforcement agencies accessing sensitive data
3. Drone delivery
Remote consultation may result in a need for remote treatment. Drones are now being developed to deliver medicines, blood and other medical supplies.
Examples
- Delivery of drugs
- Delivery of blood
Potential benefits to health
- Drugs and other medical supplies can be provided to inaccessible locations
- More efficient logistics
Potential risks to privacy
- Video surveillance
- GPS tracking
4. Electronic medical records databases
Electronic medical records help to monitor and track healthcare provided to a patient. They can be used to share information about a patient’s condition, support communication, and check effectiveness of treatment. They are also collected and compiled by private companies to support research.
Potential benefits to health
- Healthcare professionals can easily access your medical history
- Supporting researchers to understand patterns in disease prevention and treatment
Potential risks to privacy
- Secondary use of data for research (public or commercial) without consent
- Breaches or leaks of data that allows people to be identified
5. Digital identity (ID) systems
Some health systems require a national digital ID (or other unique identifier) to access care. Such systems may include biometric markers such as fingerprints or iris scans. They may also result in peoples’ medical records being combined with other digital ID datasets, such as ones relating to residency, education or tax status.
Potential benefits to health
- Improve access to health programmes
- Easier to access and combine health information from different places to track health history
- Ability to track update of medical interventions (eg vaccination) at population level
Potential risks to privacy
- People being forced to provide biometric data to access health services
- Use of data for purposes other than those they were originally consented to
6. Healthcare Information Systems
In addition to medical records, electronic healthcare information systems (HIS) are used to manage healthcare data for wider purposes such as assessing drug effectiveness or population health dynamics. They can support both clinical decisions and hospital business management.
Examples
- Identifying patterns and trends in health conditions and programs
- Research and laboratory tests
- Day-to-day business needs of a healthcare institution.
Potential benefits to health
- Standardised and interoperable data
- Improvements in understanding, prediction, prevention and evaluation of health care interventions.
Potential risks to privacy
- Data leaks
- Use of data for purposes other than those they were originally consented to
7. Biometric, genetic and genomic data
HIS and other medical datasets may also seek to include data about a person’s physical characteristics and genetic makeup. This can be distinguished from other medical data because it is about a person’s inherent characteristics rather than particular traits, symptoms or conditions that they have.
Potential benefits to health
- Better understanding of genetic causes of health conditions
- Ability to identify and treat people with genetic conditions
Potential risks to privacy
- Highly sensitive data being leaked or used for illegitimate purposes
8. Wearables
A number of devices now exist that can be worn on the body and track people’s behaviour, symptoms and vital signs. Watches like FitBit and Apple Watch are perhaps most well-known, but you can also get rings, glasses, prosthetics and more that monitor your health and activity.
Examples
- Smart watches
- Blood glucose monitoring devices
- Sleep monitoring headbands
Potential benefits to health
- Increased collection of ‘real-time’ data
- Can detect medical emergencies and call for help
Potential risks to privacy
- Constant surveillance
- Insecure data collection and/or transmission practices
9. Artificial Intelligence
There are many ways that AI is used in the health sector to augment human judgment. This area looks set for rapid growth as governments and businesses alike seek to harness the power of AI to drive efficiency and innovation in healthcare.
Examples
- Analysis of x-rays and other medical images to support diagnosis
- Algorithms supporting decision-making about eligibility for health insurance
Potential benefits to health
- New insight into health data and medical advances
- Saving time and accuracy of diagnosis
Potential risks to privacy
- Development of accurate algorithms requires large volumes of sensitive data
- Algorithmic decision-making may be biased and/or hard to challenge
Big Tech in healthcare
While some digital health tools are designed and delivered by public healthcare providers, many are also dependent on the private sector. Digital health tools are often complex and depend on specialist technical expertise developed within the private sector. While this is not inherently problematic, public-private partnerships do create additional risks such as:
- lack of open, inclusive and transparent decision-making processes;
- excessive data collection;
- data being used for purposes other than healthcare.
It’s not just new or specialist companies working in health tech either. Big name companies such as Google and Amazon are engaging in healthcare as part of their business and investment strategy in various ways.
But Big Tech’s business often relies on the acquisition and processing of vast amounts of people’s data. The centrality of data and the value of large datasets to both healthcare and to Big Tech creates incentives for them to work together. The large global expenditure on healthcare only magnifies these incentives: the World Bank’s records on health expenditure show that globally over US$1000 per person is spent on healthcare each year (this is around 10% of GDP).
To give a sense of the scale of activity in the healthcare sector by Big Tech firms, we’ve collected just some of their products, services, acquisitions and investments in the explorer below. Through these, we can see Big Tech engaging in the healthcare sector in the following ways:
- Selling data management and analysis services to healthcare organisations;
- Collecting health-related data to combine with and enhance existing datasets;
- Using capital and expertise to test, explore and innovate new ways of diagnosis and treatment; and/or
- Developing tools to enter as healthcare providers themselves.
Click the arrows to expand and see the examples of Big Tech in healthcare
Alphabet/Google
- AMIE: an artificial intelligence that is claimed to have better bedside manner and make better diagnoses than human doctors
- Anthem: a health insurance company working with Google Cloud to generate synthetic data aimed at detecting fraud
- Bayer: collaborating with Google on drug discovery and radiology tools
- C2D2: a machine learning approach to colonoscopy screening
- Calico: an Alphabet-founded research and development company working to discover and develop potential new therapies for patients with age-related diseases, including neurodegenerative disorders
- CareCloud: working with small and medium healthcare providers to inform clinical decision making
- Care Studio: provides tools to give clinicians an integrated view of patient records
- COTA: using NLP to process unstructured oncology data
- DeepMind: a history of health interventions, including:
- Eye diagnosis: AI that can detect over 50 eye diseases as accurately as a doctor
- Project Nightingale: gathered health data on millions of Americans
- Royal Free Trust: highly controversial use of patient data that faced legal challenges. The related Streams app has been discontinued.
- UCLH: a project to improve treatment on head and neck cancers
- Derm Assist: an AI dermatology tool to assess skin conditions
- Google Fit: app for tracking health and fitness data
- Google Health Studies: an app that collects data from users to contribute towards medical research
- Google Ventures (GV): making a range of investments into healthcare companies
- HCA Healthcare : a strategic partnership that includes using generative AI to help with administrative workflows in hospitals
- Highmark Health: a six year strategic partnership to develop Highmark’s ‘Living Health’ model (also working with Verily)
- iCAD: medical device manufacturer Incorporating Google AI technology into mammogram tech
- MedLm: a family of foundation models designed to help clinicians and researchers carry out complex studies, summarize doctor-patient interactions and more
- Med-PaLM: a large language model specifically designed to work in healthcare settings
- Nest Hub: device that can monitor and track sleep patterns – recently acquired Senosis, a health-monitoring startup
- Northwell Health: New York State’s largest health care provider using Google Cloud to equip clinicians with decision making support and to help the health and hospital network operate more efficiently.
- Sanofi Global: partnered with Google to create a healthcare innovation lab and more recently working together on diabetes
- Verily: an Alphabet company with a focus on healthcare. Initiatives include:
- Diagnosis: tools from Verily that screen for diabetic eye disease
- Study Watch: a sensor-based wearable device that sends data to researchersl
- Granular Insurance Company (formerly Coefficient Insurance Company): employer health insurance
- John Hancock Aspire: Diabetes life insurance plan combined with a technology-enabled wellness program
- Project Baseline: platform for patients to contribute their data to clinical research
- Vertex AI Search: has been made available for healthcare purposes
- Wearables including Fitbit: (whose data will be analysed by AI to provide personalised health recommendations, Google Pixel Watch and other ways to detect your physical behaviour.
Amazon
- 1492: a secret “skunkworks” lab dedicated to opportunities in health care, including new areas such as electronic medical records and telemedicine
- Alexa products and features:
- Alexa Care Hub: a set of Alexa features you can use to remotely check in on older relatives in their homes
- Alexa Together: a remote caregiving service to help care for friends and family
- NHS-Alexa: NHS’s website content as the source for the answer given to medical question
- Amazon Care: formerly provided primary care for Amazon employees
- Amazon Comprehend Medical: a natural language processing (NLP) service that extracts health data from medical text, such as prescriptions, procedures, or diagnoses.
- Amazon Pharmacy: medications delivered in 60 minutes or less via drone.
- PillPack: acquisition 2 years before launch of Amazon Pharmacy
- Eli Lilly: drug delivery, including weight loss medication
- Working with companies like BlueCross BlueShield and Prime Therapeutics
- AWS Health Services:
- HealthScribe: uses speech recognition and generative AI to automatically create documentation from patient-clinician conversations that can be entered into an electronic health record.
- HealthLake: aggregates, standardises and structures health data
- HealthImaging: stores medical images and analyses them with AI
- HealthOmics: stores and analyses genomic and other similar types of data
- Healthcare Accelerators: programme supporting startups using AWS to develop health related products (for example ClosedLoop.ai, used for predicting patients at risk)
- Health Condition Programmes: virtual care between doctor visits, working with providers such as Omada Health and Maven
- Halo: now-abandoned fitness band and subscription service, at one point had plans to integrate an AI-powered personal trainer
- Haven: a health plan for company employees at JPMorgan/Berkshire Hathaway
- Health Navigator: online symptom checking and triage tools
- One Medical: previously Amazon Clinic, provides users with 24/7 access to third-party healthcare providers directly via Amazon’s website and mobile app
Apple
- Beddit: sleep tracking app acquired by Apple
- CareKit: an open source software framework enabling developers to build apps that help you manage your medical conditions.
- Gliimpse: a personal health data platform acquired by Apple that collated user data from labs, hospitals and pharmacies.
- Health app: allows users to store, visualise and share their health data. Health records from medical institutions can be combined with patient-generated data and shared with providers.
- Health Gorilla: Reportedly working with this health records and diagnostic data company
- John Hancock Vitality Programme: collaboration with insurance company
- Project Casper: reported plan to open Apple primary healthcare clinics
- ResearchKit: An open source framework for building apps that makes it easier to enroll participants and conduct studies.
- Tueo Health: app to monitor asthma symptoms in children (acquired by Apple)
- UnitedHealthcare: fully insured members in most US states will be eligible to sign up for Apple Fitness+ and get a 12-month subscription at no additional cost
- Wearables:
- Apple Watch: tracks and notifies about heart rate, cardio fitness, blood oxygen and falls
- Air Pods: these now include hearing test and hearing aid features
Microsoft
- Azure: Microsoft’s cloud has purpose built tools for digital health tools, including:
- Azure Health Bot: conversational healthcare chatbot
- Flywheel: a medical imaging AI available on Azure
- Microsoft Cloud for Healthcare: cloud platform designed for healthcare providers
- Microsoft Genomics: genomic data analysis on Azure
- DAX Copilot: uses AI to write up clinical summaries integrated with electronic healthcare records
- Healthcare NExT: an earlier initiative focussed on the development of digital health tools
- Microsoft Fabric: storage and analysis of healthcare data, used by Northwestern Medicine and SingHealth (Singapore’s largest network of public healthcare institutions) among others
- Microsoft Health Bot: Empowers healthcare organizations to build compliant, AI-powered chat bots
- Nuance: Speech recognition, conversational AI and cloud-based ambient clinical intelligence for healthcare providers.
- Paige: collaboration to develop cancer diagnosis and treatment tools
- Sentillion: 2009 acquisition of company providing healthcare software
- Teladoc Health: Using Microsoft AI to automate clinical encounters and documentation
- University of Pittsburgh Medical Center: investing $2 billion in three new hospitals being designed in collaboration with Microsoft.
Tech industry and health data exploitation
It might be easy to get excited about the potential of some of the healthcare tech listed above (as well as futuristic sounding things like robot surgeons and using AI to detect cancer), but it’s also important to remember just how much sensitive and personal information is at stake when using digital health tools. If these tools do not properly respect privacy and treat people as people rather than data-sources, they have the potential to do just as much (if not more) harm than good.
As well as the harms that can result from data breaches and the loss of confidentiality, anonymity and trust, there is a bigger picture to consider too. Health is central to our humanity as living, breathing beings: how we manage healthcare can be revealing about how we believe we should treat each other. It can even indicate the sort of society we want to live in. Below we consider some of the potential implications of taking the wrong pathway for digital health.
Big Tech and a data first, people second, approach
Companies’ involvement in healthcare is just as dependent on opportunities for business growth and diversification as it is on improving public health. The sort of healthcare this delivers may therefore foreground an increased reliance on tech and data collection as the way to meet our everyday needs. How to improve public health and how to use data and technology to deliver improvements and/or savings in healthcare are not necessarily the same challenges. If done wrong, Big Tech’s approach to healthcare may end up putting data rather than people first.
The first priority in healthcare must be meeting everyone’s inalienable right to health. New technology is just one tool for doing so. Fulfilling that right for all means providing a diverse and inclusive approach to healthcare that accounts for wider societal norms and needs, running a variety of facilities and programmes, and involving people in decision-making. Technology must follow rather than lead these activities.
A techno-driven form of healthcare may in fact have negative impacts on people’s health – whether because of the anxiety induced by use of apps and social media or the risk of poor quality pseudohealth (or ‘wellness’) advice. App providers may prioritise continued interaction with their service rather than improved health (and so disengagement from the app).
Putting tech development first may also prioritise the most profitable aspects of healthcare. The experience of people who are willing to pay to make their healthcare more convenient and comfortable may be enhanced ahead of ensuring that healthcare is available, accessible, acceptable and good quality for all.
Implications for autonomy
An approach to digital health that is dependent on increasingly intrusive data collection will also affect people’s autonomy. People may be forced into uncomfortable decisions: being asked to constantly provide intimate details of what they eat, how they exercise and where they socialise to their healthcare providers via wearables, apps or other tracking technologies. Insurers or other private companies may also seek access to this data.
At the same time, your very actions themselves may be influenced and restricted by algorithmic demands. Healthcare providers might start commenting about how frequently you exercise, or the chocolate and wine in your (online) supermarket trolley, as just one example of how future scenarios might turn out.
Those who already have less freedom over their lifestyles because of existing socioeconomic inequalities will have the hardest choices. Choosing increased healthcare premiums, lifestyle changes or surrendering more data are harder choices for those with already less agency and a greater fear of oppression. One of three unpalatable options must be taken: a cost to your privacy, your wallet, or your health.
But human rights are not trade-offs or luxuries. They are not only for those who can afford to enjoy them, and you should not have to plug into the datasphere in order to access healthcare.
Implications for society
Because good health is so central to who we are – and looking after others so emblematic of our humanity – how we organise healthcare has big implications for society. The creation of modern welfare states and universal healthcare came alongside the development of human rights as international moral and legal standards after the atrocities of the Second World War. These movements were based on the promotion of solidarity and of dignity: that everyone deserves a good life, and that we can achieve that by supporting one another.
Putting too much reliance on digital tech may mean subordinating ourselves and our future prospects to unaccountable algorithms that cannot know what it is like to have an ill child or to be afraid in a pandemic. The undeniable physical reality of health demonstrates that data alone cannot define us and how we (ought to) live. Human contact matters in healthcare. This is true whether privacy is respected or not.
Staying in shape and eating nutritious food should not become a reaction to a reminder from your health insurance company that next month’s insurance premium is about to go up. There might be pure pleasure and personal gain to be found in them instead. Secretive ‘black-box’ decisions and impenetrable algorithms should not intervene in how you relate to your body and how you interact and connect with others and your community.
Conclusion
Technological research and innovation has always been at the forefront of medicine. The drive to design new tools and gain deeper understanding about what makes us ill – and what makes us better – is instinctual. To try to hold back that tide would be futile and counter-productive.
But delegating care of our bodies and our minds to Big Tech’s vast databases and silent algorithms is risky. It risks those who stand to be discriminated against because of their health status in particular, but it risks us all heading into a world where our autonomy is curtailed and our humanity diminished. We must avoid any scenario that differentiates who gets to enjoy which rights and at what price. Human rights belong to us all equally, based on what we share as humans.
Human rights are also indivisible and interdependent. That means that they cannot be traded off, and instead are mutually strengthening. Realising everyone’s right to health can only be done by respecting their right to privacy. Other approaches misunderstand the complexity and diversity of being human.