From data science to drinking water: love of the old and new

How will our data-rich world change medicine? It will certainly help us as we begin to master and make sense of the masses of data collected routinely. But blind advocates of harnessing new data, like England’s current health minister, need to code, analyse and conclude with caution. While the true benefits of data may seem promissory and distant, the harms are with us in breaches of privacy, confidentiality, and misuse and manipulation. Data fall into the same trap as technological innovation and genetic profiling by offering solutions that we find hard to critique largely because they are beyond our understanding and experience.

This is not to say that data or technology or genetics must be opposed. Far from it. But many of us are attracted to what’s new, and in medicine what’s new is understudied to the extent that the prospect of better cures or better health and wellbeing is unable to be balanced by sufficient evidence on harms and difficulties in implementation. Those cautions come much later and allow a wide window of opportunity for people seeking to ride on the bare back of the latest fad to financial, political or academic advantage.

The developing field of data science offers one route to better understanding. ¹ And while new methods of data collection will pave the way for new methods of analysis, the values of evidence-based medicine must be retained. A promising world of new data must not become a Wild West of quick and dirty analysis. How to do this will be hard. We have enough difficulty ensuring best practice in handling data in our existing system of observational studies and randomised trials. There is already too much information, and too much of it is based on data that are flawed, irrelevant or inaccessible. Before we make matters worse by submitting ourselves to the gods of artificial intelligence, and machine and deep learning, clinicians who train in the mysterious arts of computing and data science will be better placed to serve their patients.

Data science can help us too in our earnest endeavours to reduce medical error and learn from it, ² bring social prescribing on par with pharmaceutical prescribing ³ and revitalise generalism. ⁴ It should help us test clinical hypotheses, like a novel approach to surgery for Eisenmenger syndrome, ⁵ and learn more from familiar ones. ⁶ If talk of data, technology and innovation is too much, however, I recommend a plea for reviving an ancient public health intervention: free, clean drinking water in public places. ⁶ In some ways, medicine doesn’t change so much after all.