Big Data is transforming health care in multiple ways, from patient management to diagnostic and treatment methods. These new technologies are changing the health and public health landscapes, offering improved public health and clinical care. However, careful oversight of proposed uses of Big Data technologies is needed to protect against discrimination and increasing health inequities. In this blog, I propose that governments should undertake human rights impact assessments, including assessments that integrate right to health impacts, before using Big Data driven technologies in health. These assessments provide a structured approach to examining multiple ways in which the right to health could be at risk, including, but moving beyond, privacy issues.
Examples of the use of Big Data in healthcare include personalised medicine where a patient’s treatment is tailored to their genetic and environmental profile, DNA sequencing (which results in vast amounts of data stored in bio banks), forensic, genetic or medical databases, including data from public health studies and clinical trials – and all of which can be re-purposed for various technical inventions. The artificial intelligence (AI) industry in health care is booming, with growth rates in economic terms of around 40% per annum, reaching over $US6 billion by 2021. All this depends on access to huge data sets.
The concerns about the use of patient data, whether for patient management or clinical purposes, have focused predominantly on privacy and breaches of security. Although this is crucially important, here I examine broader social and economic rights issues, through the use of an abridged right to health framework (see the Oxford Textbook of Global Public Health, chapter 3.3, new version due in 2020).
Key elements of the right to health in the context of Big Data
1 Human rights laws
While some health data sets are collected and controlled nationally, particularly people’s medical records, much of the data feeding into global databases is derived from individuals contributing their own data beyond the borders of their resident state. This includes fitness and lifestyle data derived from wearable devices, or personal information uploaded into Facebook and other social media, as well as far more clinical and medically detailed data such as DNA sequencing, or clinical trials data. The rules, regulations, and laws governing how data can be used, what protections are afforded to the individual providing such information, are not universal, although there are some regional regulations, such as the European Union’s General Data Protection Regulation (GDPR). Collection and use of data that ultimately comprises Big Data may be overseen by professional ethics, developed to guide fair use of data, but such codes of ethics are not legally binding.
National laws are needed to ensure that personal data is collected, stored, and used in ways that respect, protect and fulfill human rights entitlements; but how can this be achieved beyond national borders? Human rights law, and treaties such as the Equally, all usages of Big Data, regardless of how the data sets are initially developed, even if ethically or with individual consent, must not subsequently be used in ways that violate human rights, including the right to health. States must be held to account for any breaches of rights relating to Big Data, through the various UN monitoring committees, including the Human Rights Council Universal Periodic Review.
2 Cooperation and international assistance
There are vast global inequities in relation to data: from the individual level (being literate, owning mobile phones, having internet access), to national information systems (functional national statistical offices, electronic information systems, including for health data, and a Big Data driven sector with the resources and capacity to develop, analyze, and use electronic applications). Those countries, usually low-income countries, without sophisticated and extensive electronic infrastructure will not benefit from the uses of Big Data, will be left further behind and this will increase global inequality. Specific international assistance and technical support is needed, and is a right to health obligation, as well as being a Sustainable Development Goal commitment, from wealthy countries to address information systems deficits.
3 Dignity, equality and non-discrimination
Big Data enthusiasts cite its usefulness in many health domains, going beyond clinical purposes. An example is excitement about social media’s role in monitoring and tracking epidemics and monitoring SDGs. However, it is a human rights imperative that inequality must not increase and delivery of healthcare must not be discriminatory. Therefore, it is important to disaggregate data collected to determine who has, and who has not, contributed to the data sets. When considering the use of Big Data to track epidemics, for example, there are still large pockets in the world where communities live without internet access. Global access to the internet at the end of 2017, was 54.4%, ranging from 95% in North America to 35.2% in Africa. Several African nations have less than 10% coverage, with Eritrea the lowest on 1.4%. If traditional ways of monitoring epidemics were replaced by methods based on social media, communities or countries living beyond internet penetration are at risk of invisibility during health or humanitarian crises. This would have severe, and discriminatory, health outcomes, leading to increased global health inequity.
Over the past few years, use of Big Data has been plagued with scandals: for example, in the United Kingdom in 2016 over one million confidential patient files were given freely, and without hiding personal identities, to a Google-owned company to develop privately owned apps for clinical use; throughout 2016-2018 details were released about the extent of data mining through Facebook networks to influence voters and US and Brexit election results. Exposure of these scandals has tended towards a narrow focus on privacy which has thwarted attention to related human rights concerns. As a result, rules have now been introduced to help control aspects of data collection and to protect privacy, but such rules focus more on rights to privacy than they do on the longer-term ownership and use. Transparent, effective, and accessible accountability mechanisms are among the most crucial characteristics of the right to health, and these extend far beyond privacy, to all aspects of people’s entitlements.
It is impossible to predict how data collected now will be used in the future, but our social and economic transactions, including health care, are becoming ever more dependent on data. As a result, the value of the data is ever increasing and it is largely owned by the private sector. Governments will have to enter into agreements with the private sector to provide data-dependent social services, including health services. There are clear risks, including data availability, cost of access, and quality.
Governments have a responsibility to plan for this future to ensure that the right to health is protected in a future data economy, and to be held to account for it. One form of protection could be that governments should stop giving data away without demanding free access to its future uses. Otherwise, ownership of data by the private sector will almost certainly lead to inequitable outcomes. Data processors must also be held to account for the human rights implications of the products they develop: for example, will data be gathered in discriminatory ways (perhaps from poorer communities and countries who use public health services), but will the resulting products be used only by individuals or their insurers who can afford to pay for them?
6 Core obligations and progressive realization
The use of Big Data should contribute to the fulfillment of people’s right to health and not interfere with their entitlements. New Big Data driven technologies should not be adopted at the expense of, or with funding that would otherwise go to, services identified as ‘core obligations’ or services that will increase health inequities.
All states are obligated to make continuing improvements with regard to respecting, protecting and fulfilling people’s health and other social rights – putting their maximum available resources towards this effort. When considering Big Data in the light of this principle, and that of equality and non-discrimination, it becomes important to ensure that any benefits arising from the use of Big Data do not accrue exclusively to privileged groups in society (perhaps urban elites) while poorer people in rural areas remain beyond the reach of data driven clinical services.
7 Available, accessible, acceptable, and good quality
The benefits of Big Data should be available, accessible and acceptable and based on good quality information and methodologies. This covers such issues as ensuring that the datasets from which applications develop are built from quality, robust, non-discriminatory data, with privacy protected. Is the use of the data agreed to and acceptable to the people from whom the data was collected? Services that arise from the use of Big Data must meet the ‘available and accessible’ criteria and avoid increasing inequities. For example, if patient information and notifications switch to digital systems, does this discriminate against people without smart phones, and result in them losing access to these services?
8 Strengthening health systems
The health system is the core institution through which the right to health is fulfilled.
Data is an essential component of health systems, used for improving patient management, providing health information, and importantly, to monitor the health system’s performance. As data sets increase in size, and as multiple data sets are integrated, it is essential to monitor that this use of Big Data continues to strengthen the health system. Questions about Big Data-driven technologies and their impacts on health systems cover both technical infrastructure aspects, such as the robustness of the new systems and their vulnerability to hackers or power outages, through to ownership of the systems and the quality and reliability of data contained within them.
An impact assessment cannot hope to anticipate all future uses of Big Data, and to therefore guarantee the respect, protection and fulfillment of the right to health in all circumstances. But impact assessments do help conduct a systematic consideration of the broad range of possible impacts arising from the donation, collection, and reconstitution of data. To comply with their human rights legal obligations, governments should carry out human rights impact assessments, including assessments that integrate right to health impacts, prior to adopting and implementing policies based on Big Data.
Disclaimer: The views expressed herein are the author(s) alone