On Sat 25th & Sun 26th of January, http://nhshackday.com was held at Cardiff University. Previous events have been in London, and Oxford and Cambridge unis have had their turn – thanks to a very driven Anne-Marie Cunningham it came to Wales for the first time. Given that Nye Bevan was from these parts, it seems particularly appropriate.
If you’re not familiar with hack days, essentially you get a bunch of domain experts in an area (medical in this case) together with a group of developers for a day or two, and see what can be produced. The domain experts can be optional, some just have a theme. But I like the mix of technical and non-technical people – it normally produces much more interesting and inventive results. The people with good ideas potentially get a prototype or proof of concept (occasionally a finished product), and the developers get to develop software without the usual projects and paperwork.
I’ve done quite a lot of these hack days and bar camps now, and indeed our own Comp Science students run one each year. Though it tends to be more as a judge than developer those days, due to slowly deskilling myself in senior management. So judge it was for this one too. I was pleased to see such a broad panel of fellow judges with differing backgrounds.
The new Haydn Ellis building was the venue. It hosts some of our research institutes, so is an appropriate location. And eagle-eyed viewers may have seen it moonlight as a hospital in the last episode of Sherlock. I know there was also a lot of behind the scenes work from professional services colleagues at Cardiff Uni too, making sure the infrastructure was there. Hack days do love power sockets and wifi…
I liked the pitch format at the start of the first day a lot – people with ideas had two minutes to pitch it in order to recruit people to their team. There was a great spread of ideas, across a variety of disciplines and possible technical solutions.
As they were pitching one thought kept cropping up – while it’s great to pair up cross-discipline people (indeed it’s a model we’re keen on in research), I’m sure there’s a good case for more joint honours degrees with computing. Full medicine and computing might be a little heavy going, but by thinking of computing a bit like foreign languages as an add-on, it could add so much value. Doesn’t need to be the hard-maths part of computing – just a few modules of basic code, design, and more understanding the possibilities.
Anyway, back to the hack day. After the pitches the teams started to form over coffee, then the hard work started. Lots of whiteboarding, arm waving and gatherings of laptops…
There were a dozen projects that presented (from around 16 pitches – a pretty good ratio). Apologies if any of the medical parts of these descriptions are wrong – I was there for the technical view!
A few were based around presenting existing information in a more portable or flexible way. There was an Opioid Manual app for palliative care, which replaced a long paper based guide for patients with a mobile friendly solution. Similarly there was an Out of Hours Palliative Care Toolkit. This was an Android tablet app with a toolkit of information, this time aimed at staff. Perhaps something to consider with these would be focussing on making the source information mobile friendly instead, and cacheable offline for mobile/tablets, rather than fully native apps with embedded content. Both very useful prototypes though, and serving a clear need. Of particular note was the same developer worked across both, so doubly impressive.
Others focussed on visualisation of information. Lunio was a neat web app to display content from GitHub – this could have many applications beyond just medical open source apps. Another, Colour Wheel, used some interesting D3 (think spotted a few projects using that library) visualisation. Think using sunburst diagrams. These showed the value of presenting information in different, inventive, easier to understand ways.
Open QI was a web application focussed on how you could make better use of the learning from Quality Improvement audits that doctors take part in during their training. Currently these audits are often not completed, written up or shared (a figure of only 15% was mentioned). It produced a project based web application (looked like using Bootstrap or Foundation), along with a JSON API. It was certainly ambitious, and was a solid base for potential future work. It was another polished UI too – even with their own logo and branding.
A couple were about the emotional wellbeing of patients. One, cunningly called Emotional Wellbeing, was a content driven site with a variety of modules around advice and toolkits for wellbeing. It encouraged keeping a journal of positive things and self reflection. Another, called Essence, was an online diary a patient could keep, recording their emotional state over time. If there was a prize for best looking, this would have won it – some very nice design work, multimedia and visualisations.
Those were just the application that didn’t make the final shortlist, though all excellent ideas with potential. The final seven were:
Work Based Placement Assessments. When on placements students are regularly assessed. This is currently all paper based. The project delivered an online version. Looked to be using the open source LimeSurvey from the URL formats. Another example of a relatively simple solution potentially providing a lot of value. Interestingly we (Cardiff Uni development) recently produced a very similar prototype at the request of our dental area – we used WordPress and Formidable Pro in our case. Also of note was some discussion about security, “proof” of who reviewed – largely cut short by pointing how how inherently insecure paper and signatures are. Again, a pragmatic focus on something at least being better than a current system is something many real world projects can learn from.
Wound & Dressing Selector. This was a web based application to filter and select wound dressings. This was a very nice idea, and with some more work and time (adding photos would particularly help) could be a very practical app. It’s also one where there could be a clear sponsor if needed, from dressing manufacturers…
And down to the top three… drum roll…
BCON.CC. This won best for improving patients lives. David Rankin had the idea after his brother-in-law found someone on the side of road at following a bike crash. They didn’t know his name or any information about him. The idea was relatively simple – to have a barcode or QR code on a bike helmet in order to provide information to first responders. They had a pretty fully formed idea and solution – a web app and a scanner. They’d even spoken to the Welsh Ambulance Service to find out the OS mapping information most useful to them. You can read more on the site they created – http://www.bcon.cc.
Clever Early Warning System. This won best for improving clinicians lives. The scope of the project was huge – it encompassed a web application, PDF generation, an iOS barcode scanner and hardware interfaces. Admittedly a bit was prebuilt – such as the vital sign monitoring interface, but they were open about this – the vast majority was built over the weekend. It’s a ward management system in essence. It provides patient prioritisation for ward round observations, based on an early warning score. It emails the ward manager if a patient’s early warning score escalates. There’s an iOS app where you can scan a patient’s barcode (they even produced sample tags) and it will display all the stats about the patient, including graphs and visualisations. The same reports are accessible via the website, and can be exported as PDFs for archiving. This was a strong project from all aspects – a great idea, a really good implementation, and a very impressive amount of work over a weekend.
And the overall winner was:
Cardiff iCards. Luke Anderson, an Opthamologist at the Heath Hospital here in Cardiff, currently uses printed cards for acuity measurements in toddlers aged 1-3. These are popular, but working out where a child is looking can be subjective. The team produced an iPad application that displays these images on screen instead, and tracks eye and head movement. They used the OpenCV open source library, and built an impressive new application on top of it. They’re also releasing changes they’ve made back to the community. They also thought ahead to open data and sharing, potentially linking to Open Eyes, and electronic patient record system. It was extremely close, but the eyes won by a nose (ahem) – there was a real concrete benefit to avoiding unnecessary surgery on children, cost savings on printed cards, would allow much wider distribution, and something that would be very hard to solve without using the technology now available. Technically it was seriously impressive for a weekend’s work too. It’ll be very interesting to keep an eye on this. I’ll stop now.
The judging and awards at these things are always a little academic – it’s one of those cases where it really is the taking part that counts. There was a lot of discussion behind the scenes, with good comments and a lot of respect for all of the products created.
All in all, a very successful event. There were two things that came across pretty clearly.
Firstly, it’s not always the fanciest or most complex technical solution that is most valuable. While a couple of the main winners were very impressive, many of the other were relatively straightforward – yet could deliver huge value in their own area.
Secondly, the real value in close collaboration between a business area (medical professionals), end users (patients), and developers. Put them in a room together, let them talk directly, remove long specs and contract negotiations, and deliver. Sounds familiar.
Finally many thanks to all those who gave up their time and put in so much hard work (really the judging part is easy!).