DeepMind, a British-based artificial intelligence company that was acquired by Google in 2014, are currently piloting a new tool working alongside NHS clinicians. Although DeepMind specialise in AI systems, this project does not currently incorporate any such technology and instead seems to be all about data access. The thing that stands out though is the way in which they have gone about creating it.
Working with any health data can be a tricky task, not just for obvious reasons such as the huge part that security plays when working with people’s medical data, but also the amount of data involved can be a sticking point. As technology has evolved it has meant that we very rarely run into problems with aspects such as the amount of data that we can transfer, store and manipulate, even the phones in our pockets can now crunch and store much more data than the average desktop PC could 10 years ago. The sticking points that we often face now, when working with health data, is how to present the vast amounts of data to the people whom need it. Finding a way to do this without overloading the user with useless information and making sure that the things that need to stand out do, has become a major consideration.
The way DeepMind are solving this problem is a simple one, but one that is often overlooked or underestimated when creating both apps and web-based solutions. It’s simply this: start with the end user. Whether that is a health professional, a home user or administrative staff that needs access to certain reports, these are the people who need the data so these are the first people who should be consulted as to what the output should look like. What do they need to see? What isn’t as important? What things should they receive alerts about? It’s refreshing to see a design process that is designed from the ground up by the people on the ground. Understandably, this can be difficult to do for smaller projects, getting the correct people together can be difficult, getting them to think about a system that doesn’t yet exist is hard and sometimes knowing the right questions to ask can be just as difficult as trying to answer them. The hard work is likely to pay off though. What a doctor wants is an app that tells them exactly what they need when they need it, and who is in a better position to say what that need is, than a doctor?
As a developer, I certainly know that an understanding of how a system is to be used and what is needed is a lot nicer to work with than having to make things up as you go along and hope that you’re getting it right.