Health Foundation report on Learning Health Systems

Realising the Potential of Learning Health Systems

The Faculty of Clinical Informatics was established to advance the health of the public through clinical informatics. I used to think this was fairly straightforward; after all, #DataSavesLives. However, the reality is very different. 

My original view was reinforced as I explored the literature and listened to enthusiastic conference speeches about the digital revolution. I was, and still am, taken by the idea of a health system that employs informatics to learn from every patient treated – a Learning Health System.

But to my surprise, I soon learned that clinical informatics is littered with failed projects. My own endeavours took longer, cost more and produced fewer benefits than expected. Where are the reports about these projects? They probably aren’t much fun to write.

Luckily, The Health Foundation has funded us – the Newcastle University Learning Healthcare Project – to investigate how we can Realise the Potential of Learning HealthSystems, work that involved consultations with a wide range of national and international experts. We have just published our report. Our conclusion is much more complex than we originally expected.

First, we need to understand why we are building a Learning Health System. It could be to improve outcomes and experience for patients, to improve value, reduce variation, to generate new knowledge, to apply existing knowledge, to make better use of our IT investments or to improve clinical performance, among other things.

Next, we need to put the technical building blocks in place:

·         Data must be captured from clinical practice: This requires good data quality, interoperability and information governance.

·         Knowledge must be generated from the data: This may involve quantitative analytics or human deliberation within learning        communities.

·         Knowledge must be put into practice: This can happen gradually via books, journals and clinical guidelines. Alternatively, knowledge can be made computable and delivered to the front line – in decision support systems, for example.

Often informatics projects fail because they stop at this point, having delivered the technology platforms to support the steps outlined above. There are many ways that such projects fail to be adopted, despite delivering a functional piece of tech. We used the NASSS Framework to think about this.

Often the illness is more complex than assumed, with challenging co-morbidities or socioeconomic associations that the technology can’t cope with. Often the value proposition doesn’t make sense for the organisations involved. Often those organisations are under too much pressure or have limited capacity for innovation. On many occasions, we have unrealistic expectations of the patients, carers and clinicians who will actually use the system. The outside world can also intervene with regulations or political realities that wreck the project. And even if you overcome all these challenges, any individual element could still change over time.

So what are we to do? We must set a carefully considered strategic direction, then build the organisational structures to deliver it. We must create a culture of learning and a scientific approach to implementation. We must change behaviour and continually evaluate our progress. Perhaps most importantly, we must co-design all of this with our patients, colleagues and key stakeholders.

We brought all of this together in a single framework.


If all of this sounds fascinating – or even intimidating – then please check out our report, which contains lots of links to models, tools and case studies, that can help your organisation create a Learning Health System, whether you are a GP, a Trust, an ICS or a national health service.