A retrospective of the ODP Hack Week by NHSIC

The NHS Information Centre (NHSIC) held a Hack week on the 10th -14th October. They invited suppliers with relevant experience within the industry along to highlight various problems that they will face when trying to accomplish their strategic vision for information – the Open Data Platform (ODP). The goal is to build an open, easily accessible, central view of a patient’s medical journey throughout their lives. This journey encompasses interactions with a host of providers from secondary to primary care and in its current form the NHS IT estate is unable to provide this view.

Many issues were understood before and highlighted during the event including

  • Data quality across the entire estate

  • Data consistency across systems

  • Linking heterogeneous, currently unrelated,  data sources

By running the event in a controlled environment, providing multiple data sources and a simulated infrastructure to build and deploy some applications the hack week provided some insights on these issues and some potential approaches to solving some of the problems.

I was sceptical at first that the event would provide any significant insights to the problems that the NHSIC face. However, as the week progressed I changed my view and think that the added value the event provided was mostly in the lessons that have been learned, cheaply, during this week and the conversations that would otherwise not have taken place. That is, cheap compared to the alternative of building something for real to test its relevance or maybe just not taking the risk and build nothing at all. Either way, real requirements became apparent around the interest both in data and easily consumable services.

Why did this format work?

Well partly because the attendees were all keen to produce something useful (and show off) and partly because the room was not just full of techy types. There was a real cross section of people that you might find on a project and believe it or not people talked to each other because they were in the same room.

I met with Robbie Foy, a GP and health services researcher from Leeds University. He was saying that, despite the best efforts of most health professionals, there is currently a lot of variation in the health care that patients receive. Robbie’s research focuses on understanding and closing the gap between evidence and practice – or, put another way, ensuring that more patients receive effective health care. For this he needs patient data to identify current problems with healthcare delivery, understand why variations occur and measure the impact of strategies to improve the uptake of effective practice. It is now often possible to use existing anonymised patient data in such studies, captured from primary care information systems. Researchers like Robbie are always on the lookout for new sources of data that they can use and sometimes link to existing data to get a more complete picture of the healthcare that patients receive.

Of course, with some hard work, it is often possible to overcome technical difficulties with data. However, there are still ‘language barriers’ between the information technical, clinical and research worlds – each has its own set of jargon and perspectives – which can hinder communication and progress. Events like Hack Week therefore provided an opportunity to bridge some of these gaps – and work out ways of working together to improve healthcare. Robbie’s passion to improve the flow of really useful information to practicing clinicians is clear and hopefully the ODP will provide a strong basis for taking this forward. It’s these sorts of conversations with real users that will ultimately determine how successful the ODP will become.

By the way I did actually build something at the event and didn’t just attend as an observer.

My first port of call was to demonstrate an easily consumable, platform independent service view of the world for all providers and suppliers. I did this by building a very simple REST interface over a number of the reference data tables, which would be scattered across many data source in the real world. This gave a single source of information rather than various suppliers having to negotiate (or worse, duplicate) nasty integration problems. I provided a simple web interface to the reference data to demonstrate consuming the interface. More importantly, other teams confirmed that they were able to consume the information from these services using various REST clients. I presented this model, which, can be viewed online. Of course one of the benefits of an interface like this is the ability to change the representation of the data at the drop of a hat. So, team who wanted XML got it and teams who wanted JSON got it.

I fell into the next piece of work by accident.

I had a chance conversation with James Barrett of PRIMIS during the last day and that conversation resulted in us deciding to build something together. By working closely with James, who has worked in this area and understood the useful trend information that was buried in some of the data sources, we managed to produce something that would be really useful to the medical community.

We wrote a query on the ODP looking at the number of emergency admissions for chronic obstructive pulmonary disease (COPD) for a given time period. Users based in Primary Care could use this service to add value to their data and review the impact of reconfiguring local services and the introduction of local care pathways.

We created a REST interface fronted up by some simple HTML to show off the resulting stats with a nice shiny graph using HighCharts. Underneath the interface we just plugged in the necessary queries to the database and voila!

Unfortunately, it wasn’t complete enough to show until after the last presentation. Certain we would have won a prize for that one.

Overall conclusion

My overall conclusion was that, although there were teething problems (Guess what? This happens on every project. Retrospectives anyone?) I think these events are worth serious consideration for any program that exhibits a complex problem domain. If, for no other reason than idea generation. More significantly, though, as an opportunity for the suppliers to talk to real users rather than proxies – they are the only people who really understand the wider value of the work and the reasons why the work is important to them.

 

Watch the video of my presentation here:

 

 

Leave a Reply

  • (will not be published)

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>