Technology and risk in healthcare’s move to digital: Five lessons from the history of technology and STS

I was recently invited to speak on a panel of historians discussing new technologies and risk with an audience from the Department of Health and Social Care. Rather than a historical case study, I wanted to present some more general ideas from the history of technology and science and technology studies (STS) - hoping that although they might not represent familiar ways of thinking outside those disciplines, they would resonate with professionals who are managing change in the NHS. It was a useful exercise for me to go back to basics and try to communicate some of the core elements of how we think about data and digital. 

1. Successful technology can become invisible 

When a technology and the practices surrounding it are deemed successful and become embedded in day-to-day work and life, they can become invisible. To give an example of a data practice that we have probably all been subject to and yet take entirely for granted: one of the first things that happens in the NHS when a baby is born is that the exact time and the baby’s birth weight is recorded. These data have become essential parts of birth announcements: we are expected to exchange this information not just by entering it into health records, but in excited phone calls and messages to friends and family. But of course this wasn’t always the case. We didn’t always have precise timekeeping devices; weighing scales were not always embedded in infant care. [1] Paper case cards, with dedicated space for recording repeated measurements in pounds and ounces, were once the height of information innovation. Historical investigation can be useful in making the familiar unfamiliar and therefore better amenable to analysis: the way things are right now was never inevitable, and making visible the ongoing work that institutes and maintains technological practices is critical to understanding technological change.  

2. Digitisation and datafication are continuous processes, not discrete events 

Today’s problems are yesterday’s solutions, and (unfortunately) today’s solutions will be tomorrow’s problems. This is probably familiar to anyone attempting to build on top of layers of legacy software, or rationalise contracts with different terms and timescales. For all that past initiatives may have been intended to sit on top of, slot into, tweak, simplify, unify, or integrate existing practices, the reality for complex systems in health and care is that they are, well, complex. Digital and data work in the NHS inevitably involves some degree of untangling and tying into existing systems and practices. And once adopted, new practices become part of the tangle. ‘Futureproofing’ is never fully possible; and the work of digitisation and datafication does not have an end. Projects have start and end dates, but evolution and maintenance are continuous. 

3. Technology is always situated 

Situatedness is one of the core ideas drummed into students starting out in Science and Technology Studies and in the History of Science, Technology and Medicine. When we talk about how something is situated, we mean that we want to take into account not just its ‘intrinsic’ nature and qualities, nor simply how it interacts with its surrounding context, but rather how it cannot be understood except as part of a web of things, people, practices, and ideas. Technical specifications, user guides, standard operating procedures, and the like create the sense that a product is (in and of itself) complete, comprehensible, and transportable across time and place. But attending to how technologies are situated – how they actually work in practice, and who and what they are interdependent with – helps us to understand what we see so often: that something that works in one time and place might work differently (or not at all) in another. Failing to attend to situatedness leads to the risky expectation that, having proved successful in one situation, technological solutions can simply be rolled out and scaled up in a uniform manner. 

4. Digital does not automatically lead to better care, research, and management 

Digital and data can contribute to better care, research, and management. But they do not necessarily or automatically do so. Nor are data and digital necessarily more efficient, or cheaper, or more productive, than analogue counterparts. Digitised and data-driven practices require new forms of interpretation and labour, work that will be visible in some parts of the healthcare system but not others. There are a couple of risks here. Firstly, it is easy to overestimate the return on investment and underestimate the resource needed to make data and digital work. Secondly, data and digital often require lots of change and disruption for staff ‘at the coal face’, who need to be seen and listened to – this isn’t reflected enough in policy and management discussion that present digital as a panacea. 

5. Data and metrics are always partial 

When you’re in possession of a huge amount of data, it can seem like you have sight of all the information you need. Dashboard interfaces and other compelling visualisations promising ‘real-time’ updates, [2] for example, can provide a powerful sense of knowledge and power for users, creating an illusion of omniscience. But digital and data isn’t everything, and not everything can be digitised and data-driven. Not every aspect of health and care can be captured as data; and not every healthcare practice can (or should) be digitised. Numeric data alone can never tell a full story; it requires not only interpretation but also supplementation. Some of the most serious risks for digital and data follow from failing to recognise that they can never capture everything and tell the whole story. Good healthcare includes caring practices that are emotional and embodied, that can’t be rendered into data and are therefore at risk of being unseen and undervalued. In research, there’s a risk of directing resources to areas that are more digitised and datafied on the assumption that research problems in data-rich areas are more tractable, when this may not be the case. And in management, there is a risk of working to metrics developed from data that can only ever give a partial account of the world, and which may become increasingly divorced from everyday health and care practices through technocratic feedback loops. 

If I had to bring these points together as concisely as possible, I would say: systems are messy, work is never over, and digital and data dissimulate as much as they reveal. 

Based on remarks prepared for ‘Ask the Historians’ – Analogue to Digital (17th February 2026), a Department of Health and Social Care History Network event with History & Policy. I introduced myself to the attendees as a historian and social scientist drawing on my past research as well as our work in DARE; as a patient with a mix of chronic illnesses and disability; and as a former NHS staff member (Grades 4-5) in Leeds (NIHR CRN) and Manchester (Vocal). 

References

[1] Weaver, L.T., 'In the Balance: Weighing Babies and the Birth of the Infant Welfare Clinic', Bulletin of the History of Medicine, 2010 Vol. 84 Issue 1 Pages 30-57

[2] Shaw, L.P., & Sugden, N.C., 'Portable sequencing, genomic data, and scale in global emerging infectious disease surveillance', Geo: Geography and Environment, 2018 Vol. 5 Issue 2