Over at the Health Service Journal, Steve Black comments that the Tony Blair Institute has drunk too much AI Kool Aid, and is paying far too little attention to the woeful state of the NHS data that the AI would have to rely on.

We aren’t an AI company, but we are an NHS data quality company, and on that topic Steve Black certainly has a point. He describes how NHS data is fragmented across many silos, and why it is technically so hard to join up: “incompatibility of data storage systems; a lack of interoperability standards; inconsistent definitions of what the data means”.

Joining data up is not the same thing as bringing it together in one place. A data lake brings it together in one place, but without resolving the disparate standards and meanings in the source systems. A data warehouse brings it a bit closer with a transformation or two, but usually far from a complete standardisation.

So how about four transformations before we even get properly started, and then a further nineteen or so transformations with localisations and other business rules injected along the way? Overkill? Not at all – that is the attention to detail that we deliver with Insource’s Unified Data Layer [UDL].

Our UDL is so standardised that you can change a core IT system, even the PAS/EPR, without disrupting the UDL and while maintaining business continuity. If you connect an app to the UDL in one hospital, then it will work on any other UDL without modification [provided, obviously, that the other UDL also has the required source data connections].

The same applies to advanced analytics [and, indeed, AI]. As you may have noticed, the more advanced the analytics you try to use with NHS data, the more they tend to reveal the flaws in the underlying data, so the more important it is that the data is accurate and standardised as in our UDL.

It is not wrong for people to be excited about advanced analytics and technologies. But it is wrong to try and adopt them, without first making sure that the underlying data infrastructure is ready for them. Fortunately the Insource UDL makes all of this do-able without breaking the bank.

We’d be delighted to discuss how to get your data ready for advanced secondary uses such as analytics and AI. So drop us a line, and get ready for the future.