SDU Business School Gains International Recognition

Associate Professor Anna Schneider-Kamp from SDU Business School has plenty to celebrate these days. Together with an international team of researchers, including colleagues from some of the world's highest-ranked universities, Harvard and MIT, she has published the results of her latest digital health research project in The Lancet, widely regarded as the world's leading journal in health and medical research.

Many of us rely on digital health apps several times a day to monitor our health and well-being. There is great optimism about the potential of health technologies to help address health inequalities by making health information more accessible. However, the research project reveals that these technologies may not be contributing to greater health equity at all. In some cases, they may even be reinforcing existing inequalities. As such, the findings are highly relevant to a broad range of people.

The leadership of SDU Business School is delighted to see the international recognition generated by the research to which Anna Schneider-Kamp has contributed.

Whose Health?

The researchers examined the algorithms behind the diagnoses generated by health technologies. These technologies are largely based on AI, and this is often where the problem lies: AI is only as good as the data it is trained on. According to the research team, the data underpinning many health apps is heavily shaped by Western, Eurocentric perspectives that overlook other cultural understandings of health.

For example, algorithms used to assess skin conditions are primarily trained on images of light skin, which can lead to misdiagnoses in people with darker skin tones. Because these algorithms form the basis for treatments that generate data for the next generation of algorithmic models, a skewed feedback loop is created. This can reinforce cultural inequalities in health. In some cases, such biases in algorithms may even cause harm to population groups that are underrepresented in the data.

Health and Illness Are Cultural Too

The study published in The Lancet offers concrete recommendations for improving the development of health technologies through a deeper understanding of the people who use them and a broader conception of what health actually means. According to the researchers, achieving greater health equity requires cultural insight, not just medical expertise. They argue that the Learning Health System (LHS) models that have informed the development of health apps need to be critically re-examined.

- Our research highlights the importance of rethinking and redefining the approaches and underlying assumptions of research and innovation in order to create a more equitable and inclusive foundation for health and well-being, says Anna Schneider-Kamp.

The cultural bias embedded in AI algorithms means that health technologies can perpetuate assumptions and prejudices about health and well-being, about what constitutes a "normal" body, and about which forms of knowledge are considered legitimate. In doing so, they overlook the fact that there are many different ways of understanding health and well-being.

Who Holds the Power?

When it comes to the design of health technologies, tech companies, investors, and research institutions are the ones sitting at the table. They decide which problems matter and whose knowledge is considered legitimate. Marginalised communities are typically not consulted during the development of health technologies, and when they are, the most important decisions have often already been made.

Developers do not usually examine the broader social and historical contexts that shape people's health and well-being. As a result, their technologies may cause unintended harm. Assumptions and biases about marginalised communities often stand in the way of developing technologies that effectively address their health and well-being. This, in turn, helps perpetuate health inequalities. The project's findings also clearly show that what investors and technology developers want from a health technology, such as technological advancement and profit, is not necessarily what matters most to the people who will ultimately use it.

The research team argues that integrating social and cultural considerations into the development of health technologies could make a significant contribution to reducing health inequalities.
- To make that happen, we need to have a serious conversation with investors and developers before a single line of code is written.