Walk about on the wards.

Chat with nursing, medical or pharmacy staff. Chat with physiotherapists, occupational therapists, managers.

They all have problems that need to be solved, problems that make carrying out their day-to-day work more difficult.

Ask them:

“What would make your job easier?”

“What would enable you to provide better care?”

At their most fundamental, their problems and the solutions to those problems relate to data.

You may scoff. Oh here he is again, going on about data as the answer to everything.

Well it is.

That’s why clinical research, such as a clinical trial, used to answer questions such as “Does this treatment work?”, is built on data; carefully acquired, carefully managed and carefully analysed data.

We have a nursing staffing crisis! We should be solving it, with data.

We need to run an audit to see whether our service is performing well. Solve it, with data (and in a computer on not on paper).

We want to identify that patient not responding to conventional treatments; we want to show them their progress compared to their peers, because we need to have a discussion about more aggressive treatments. Solve it, with data.

How many nursing staff do we need on that ward, with those patients, for this shift? How can we identify the deteriorating patient? Show the admitting team looking after the patient with muscular weakness that the GP had already checked the acetyl-choline receptor antibodies two weeks ago and the result is already available tonight. How do patients like me fare after that procedure?


We now have virtually unlimited computer resources at our disposal, both in terms of storage and computational power. We need to think big and have grand aims. It’s then our job to develop strategies to carefully take advantage of that technology to envisage better ways of working and to iterate our way to achieve those aims. As I mentioned in my disruptive influences blog post, we may change the detail, but not necessarily our direction of travel.

We don’t have all of the answers…. but we don’t have all of the questions either

The key message, and I cannot say this too many times, is:

We need to build an infrastructure that explicitly recognises that we don’t have all of the answers, and therefore must support innovation. Such innovation may simply be incremental improvement of existing technology, but also must cope with and embrace the introduction of transformative and disruptive technologies.

I wrote that during debates within NHS Wales about the best way to formulate an information technology strategy. I’m reviewing a “functional requirements” list for a project currently which will end up with a procurement. It is hard work and it brings into sharp focus an important question: how can I know what will happen when our systems finally meet their users?

Instead, we should be focusing on data first, as data as the solution to a problem. Data, structured, understandable, semantically interoperable data are the foundation on which clinical applications and our analytics can run.

And that is why you should be learning:

And start thinking about immutable, structured and understandable data stored in vendor-neutral archives or exposed using an open application programming interface.

And for homework, start thinking about how you might create an infrastructure for real-time analytics… for example, show me some simple survival curves in real-time for patients having emergency surgery for any type of pancreatic surgery in the last 30 days stratified by age, gender and co-morbidities… Or, for that matter, the data that might be needed to solve the nursing staffing allocation problem…

If we’re going to make “health information universally accessible and useful”, then let’s get the right information in the first place. After all, there are two words in “information technology”…


Further reading:

Disruptive influences in healthcare information technology: part 1, part 2 and part 3