Innovations that aim to improve the prevention, diagnosis, and treatment of human disease are often intended to have broad societal benefit, but they do not benefit everyone equally. This contributes to perpetuating health disparities and raises an important question for decisionmakers: how can the health care innovation process, and the innovations that result, be made more equitable?
While innovations have contributed to the doubling of human life expectancy over the past century, there are significant disparities between countries. Populations in higher-income countries such as Spain, Switzerland, Italy, and Australia, born today can expect to live on average more than 83 years, whereas those in the poorest nations, such as the Central African Republic, Nigeria, and Chad, can expect to live less than 53 years. Within countries, inequalities in life-expectancy also exist amongst different populations, such as between socioeconomically advantaged and disadvantaged communities and between different ethnic groups; such inequalities are prevalent in higher-income countries, especially in the UK and the United States
Social determinants of health, such as education, living conditions, income, health care system access, and infrastructure, are well-established drivers of these disparities and are receiving increasing attention in terms of research, policy, and practice. Less attention, however, has been given to how innovations themselves—specifically health care innovations—can create or perpetuate health inequalities.
Take for example, the pulse oximeter, a small, non-invasive medical device that measures blood oxygen saturation by passing light through the skin (typically, on a fingertip). This device has been available since the 1980s and is widely used for assessing how well the heart and lungs are working to circulate oxygen throughout the body. The pulse oximeter gained renewed attention during the COVID-19 pandemic for detecting potential respiratory failure, allowing clinicians to intervene early and save lives. Despite the medical and societal benefits of the pulse oximeter, evidence also resurfaced during the COVID-19 pandemic that this device may over-estimate oxygen saturation in individuals with darker skin tone, particularly people of colour. In fact, this evidence first came to light in the late 1980s, with multiple studies published on this issue since. Yet, adaptations to the devices and the impact of biased pulse oximetry readings in people of colour have still not been addressed.
A single BMI threshold has been historically used in clinical practice despite evidence that people of colour tend to develop diabetes at a lower BMI.Share on Twitter
These inherent biases can be extended beyond medical devices to innovative approaches to health care service delivery, such as algorithms used to diagnose disease and to allocate health care resources. For example, while body-mass index (BMI) is used as a measure of obesity and in calculating risk for developing type 2 diabetes, a single BMI threshold has been historically used in clinical practice, despite evidence that people of colour, including Black and Asian people, tend to develop diabetes at a lower BMI. This practice can restrict timely access to treatment and management services for diabetes prevention. A 2019 study on the use of algorithms to inform decisionmaking in the U.S. health care system found that some algorithms widely used in U.S. hospitals to allocate health care were systematically biased against Black patients, assigning them lower risk scores and thus fewer resources. Gender-related biases also creep into innovation processes. For example, for many years research on treatments for heart diseases focused predominantly on changes that occur in male patients, impacting an accurate diagnosis in women and conversely, osteoporosis-related R&D has been disproportionately focused on women.
Underlying these examples of biased health care innovations and practices are biased data, which tend to underrepresent certain groups. However, the inclusion of more diverse groups and the acquisition of more diverse data is only part of the solution. Even in the presence of robust data, biases can manifest in terms of how health care problems and their innovative solutions are prioritised and implemented in clinical practice and wider health policy. Indeed, why, after more than 30 years of accumulating evidence that pulse oximeters over-estimate blood oxygen saturation in people of colour has the issue not been addressed?
For health care innovations to have the fairest and broadest possible benefit to society, efforts could be made to attend to all stages of the innovation pathway to identify opportunities to mitigate biases through diversity, equality, and inclusion: from innovation prioritisation to design, testing, and, finally, access. In practice, this raises a series of questions of relevance to decisionmakers.
1. Problem Prioritisation and Selection of Innovations for Investment
The unmet health care needs of underrepresented groups, such as ethnic minority or gender-specific groups, have only gradually risen on the innovation agenda. There is now increased attention focused on gender and ethnicity in the conduct of clinical trials, but seemingly not in prioritising which innovations to develop in the first place. For example, approximately 1% of research and innovation funding is invested in conditions specific to women, outside of oncology. Similarly, there is a need to consider how to prioritise investments in innovative health care service delivery to meet the potentially distinct needs of different ethnic groups, for example a need for culturally sensitive diabetes services (PDF). A number of questions could be relevant for health care innovation decisionmakers, such as understanding which health care conditions more prevalent in underrepresented groups lag in innovation development and how can underrepresented groups effectively participate in and better steer prioritisation processes.
2. Innovation Design
Risks of perpetuating inequalities can also occur in the design specifications for innovations, for example when using a health care technology that requires physical and storage infrastructure not available in some parts of the world (e.g. cold storage). But there may also be other population sub-group specific sensitivities and preferences that merit consideration at design stages. Some examples in the existing literature (but meriting further research) relate to the colour of medical pills and to patient perceptions across cultural contexts, or to the need to design user experience interfaces in digital and mobile health-tech in a way that considers the needs or preferences of diverse populations. The needs of different sub-groups also become important considerations for deciding on appropriate treatment formulations, which can suit local infrastructures and facilities (e.g. availability of staff to administer an injection versus possibility of patient self-administering pills or sprays). A series of design-related questions may be relevant for health care innovation funders, developers, and payers, as part of efforts to support equitable innovation. These include understanding which groups should be involved in discussions related to product design and how best to involve and incentivise innovators and historically underrepresented groups to engage in discussions about design issues.
3. Evaluation and Testing of Innovations
The inclusion of diverse groups in clinical trials for medicinal products has become more prominent but merits equal attention in the medical devices space.Share on Twitter
Intended users of health care innovations also may need to be involved in their testing, from pre-approval clinical trials and to post-market surveillance in the ‘real-world’. In recent years, the inclusion of diverse groups in clinical trials for medicinal products has become more prominent in policy thinking but merits equal attention in the medical devices space. In 2021 the UK government signalled intent to launch an independent review to assess potential biases in medical devices that can affect ethnic groups and to identify how this could be mitigated. Importantly, meaningful inclusion of diverse groups in clinical studies may need to go beyond representation as participants, and it could entail involvement of such groups in a spectrum of clinical trial activities, including study design, conduct, analysis, and dissemination. Furthermore, learning from diverse users of health care innovations (i.e. real world evaluation) may also provide valuable insights as to how innovations can be incrementally improved over time. Central questions for decisionmakers to consider might relate to understanding the types of evaluation endpoints that can enable health technology assessment agencies to identify potential biases and inequalities in evaluation processes and understanding the role of post-market surveillance in learning about innovation suitability for diverse populations.
4. Access, Accessibility, and Use
Finally, a vast body of literature considers issues of innovation affordability in the context of supporting widescale and equitable access. Somewhat less researched are accessibility and user-experience considerations related to the skills and capabilities of those delivering innovative health care and of diverse end-users. One area where this is important is digital facilitation skills to support effective delivery of digitally enabled and remote health care, and to mitigate against digital exclusion (e.g. of those who lack affordable digital access, the elderly, frail, those with visual impairment, or those who are not comfortable with digital interactions).
Making Concrete Progress
There is widespread agreement on the importance of addressing these challenges, yet progress remains slow. The time could be ripe to consider how to build diversity, equality, and inclusion into health care innovation in more systematic ways. Some examples could include establishing standardised reporting guidelines on how diversity, feeding equality, and inclusion considerations feed into innovation and access pathways; incorporating these considerations into quality assurance processes for innovation development; and using tools in impact assessment of innovations that take explicit account of their impact on diverse groups. Discussion of these issues could help prompt further efforts to maximise the benefits from investments into novel health care innovations in fair and equitable ways and support policy-related decisionmaking that places eliminating inequalities at the heart of innovation.
Robert J Romanelli is a research leader in health and well-being at RAND Europe. Sonja Marjanovic is director of health care innovation, industry, and policy at RAND Europe. The authors would like to thank Marc Cabling and Nick Fahy for early discussions and feedback.
Commentary gives RAND researchers a platform to convey insights based on their professional expertise and often on their peer-reviewed research and analysis.