Data Science: Catastrophe or Catalyst for Health Equity? Reflections from Pint of Science

Written by Ellen Coughlan

Last week, we had the opportunity to host an evening with Pint of Science and run a debate on digital innovation, data science, and health equity. We started off by introducing Helen Butters, who gave an extraordinary account of work underway in Leeds to understand housing and health, drawing on insights from a group of Roma women who contributed to the analysis. Helen’s talk set the scene for a discussion about missing data, exclusion, participatory methods in research and analysis, and evidence-driven policy.

We wanted the evening to feel conversational rather than overly technical, so we followed Helen’s talk with an ask to the audience to stand up and vote whether they thought digital health technologies are a driver for exclusion or inclusion. With enthusiastic and curious minds in the audience, we explored both the promise and the risks of using data and AI in health and care systems. El, defending the argument that digital technologies exclude, presented an analogy comparing health data to trying to classify apples scientifically. You can record characteristics like “red or green” or “sweet or tart,” but those categories never fully capture the uniqueness of each individual apple. In the same way, healthcare datasets inevitably simplify people into measurable variables, even though human experiences, behaviours, and health conditions are far more complex than any dataset can represent.

Ellen, defending the motion that digital technologies can be a tool of inclusion, extolled the benefits of technologies to free up capacity to care, to better understand treatments and rarer conditions, and to remotely monitor patients with chronic conditions who would otherwise endure regular trips to clinics that interrupt their professional and caring responsibilities. Digital technologies that actively identify and mitigate inequalities were described, like those that seek to make organ allocation fairer and more inclusive.

Healthcare datasets inevitably simplify people into measurable variables — but human experiences are far more complex than any dataset can represent.

These ideas led into a really interesting conversation about missing data and exclusion. We discussed how some groups are often underrepresented in healthcare datasets, whether because they face barriers accessing healthcare, distrust institutions, or simply do not fit neatly into existing systems of measurement. Attendees raised important questions about what happens when data science attempts to make predictions about populations it does not adequately understand.

Privacy was another major topic of discussion. While many people could see the benefits of linking health data and using health-relevant datasets, particularly for identifying inequalities and improving preventative care, there was also understandable concern about governance and access. The audience explored the tension between using data to improve public health outcomes while also protecting individual autonomy and trust.

We also spent time discussing probabilities and statistics, particularly the challenge of communicating risk in healthcare. Data science often works in probabilities rather than certainties, which can feel uncomfortable when decisions affect real people’s lives. Several attendees reflected on how statistics can sometimes appear objective and neutral, even though they are shaped by the assumptions, categories, and historical systems behind the data itself.

Overall, the evening struck a balance between serious discussion and genuine curiosity. There was disagreement at times, but in a productive way, and it was encouraging to see so many people engaging critically with questions about technology, fairness, and healthcare. More than anything, the discussion reinforced the idea that conversations about digital health should not belong solely to technical experts. These conversations affect everyone, and everyone should have a voice in shaping them.

---

Want to write a blog post for us? Get in touch via info@datascienceforhealthequity.com