I have to confess I was a little skeptical about attending any conference which claimed to be about the future of NHS IT in Scotland. The last time I went to something like this was the 21st Century Health conference in June 2007. Apart from inspiring me to start this blog, that conference was a huge disappointment. I was standing there trying to convince people about the advantages of collaborative working, RSS, openness, etc while everyone looked at me like I was some sort of lunatic.

So today turned out to be something of a surprise. The keynote speaker was Jim Campbell, a GP from Irvine who has done some pioneering work on giving patients access to their own health information. His solution was simple and pragmatic: no patient-identifiable information is delivered through the website, the patient logs in with a username and then their health info is delivered from the back-end database without any demographics etc. Elegant and simple. Of course it could be a problem in a small GP Practice such as those on our remoter Orkney islands, but that’s an extreme case. In general I would imagine it’s very difficult to use health info to work out who the patient is. Jim mentioned the tendency for us to use the consent issue as a convenient way to say “no” when we feel like it. I was intrigued to hear that the patients help the GP Practices to maintain better quality data because they tend to contact the practice if their info is wrong.

Following the keynote, the conference split into two streams: clinical/patient and technical. Unsurprisingly, the technical stream was in the basement. Ken commented to me that it was like being on The IT Crowd. Actually, it was a better room: more space, cooler, more natural light. Bob Jarvis was the first speaker up, talking about the Microsoft Connected Health Framework. I’d heard of this before, but was intrigued to hear that they’ve recently expanded it to include community health stuff as well – that’s an area bound to get pretty interesting as more health boards and local authorities try to connect their systems together. It was encouraging to hear some these concepts being mentioned: use cases; SOA (a much-derided concept not so long ago, ISTR); “render content to any device”. Bob finished by giving us the (first of many) definition(s) of a portal. It’s both an iGoogle-style aggregation of plugins for patients to view their health record, and it’s a place where health professionals can “manage their business processes”.

Next up was Ian Herbert talking about IHE. I really enjoyed his talk, which was clear and to the point. His initial few slides in particular were so good that I went up to him afterwards to ask if I could plagiarise them! AIUI, the IHE is about bringing together multiple vendors and guiding them to implement some of the complex health standards such as HL7. Ian was at great pains to emphasise that IHE is *not* a standards body. As a programmer, this approach makes total sense to me: basically programmers are lazy and love shortcuts (that’s why they’re programmers!), so if you help them to understand how to implement standards and how to get their stuff working with other people’s stuff with minimal effort, they’ll jump at it. The UK is only just getting started with IHE, but it’s been successful in other countries. I particularly liked Ian’s description of it as “detailed, gritty work”.

Next was Eric Williams. This was a pretty well-delivered, tongue-firmly-in-cheek semi-sales pitch for Sun’s healthcare products. I’ll forgive them given that the products are open-source! There were some other nice quotes I took (probably inaccurate) note of: “allow collaborative workers access to the original data source, not copy it”; the “governance envelope” of authentication, authorisation, administration and audit. What came next in Eric’s presentation didn’t surprise me, but it did disappoint me. He described the “integration platform”, a sort of magic glue which will mean a complete, reliable and regularly-updated body of information can be gathered from a plethora of different systems. This is a seriously hard thing to achieve. Along with nearly all the later presenters, Eric glossed over this with an Apple-style “it just works”. One of the delegates asked about concurrency: what if data was updated in one system but not in another, how does the platform work out what to use? It occurred to me that this is especially true in the case of the multitude of different sources of demographic data sources in use in NHS Scotland systems at present… Finally we got our second definition of “portal” – a customisable 2-way website, enabling collaborative working both between users and between software systems. (Liked that last point.)

Andy Hardy’s presentation was about the fledging patient portal system in A&A which allows patients to monitor and manage their health info. He began with an interesting list of the most-requested features from patients: lab results, repeat prescriptions, personalisation, links to trusted info, reminders/alerts. Booking appointments was not a high priority. Essentially, Andy’s design is similar to that pioneered by Jim Campbell. The underlying tech is very familiar to me: web services coded in C#, SQL Server 2008 and a front-end written in Silverlight. (That last is perhaps a bit of a worry – I hate the use of Flash or similar in any website.) Then Andy brought out the “magic glue” again: their GP systems and hospital PAS systems feed into a database which makes sure a complete and accurate record is assembled and presented to the patient. I wonder if I can get that wonder glue in Woolies… A pilot is scheduled for the end of the year.

The final presentation of the morning was also my favourite of the conference so far. Ewan Davis talked about “internet technologies” – which he told us meant whatever he wanted it to. Ewan actually defined what are for me the *real* problems we have to face before we can even get close to all this lovely portal business. Paraphrasing: “Making this stuff work is *difficult*: it’s time we acknowledged this fact and made it clear to our bosses how hard it is. It’s difficult because of the scale, the complexity (fuzzy data), high impact (life/death) and the loaded social context of what we’re trying to do.” I especially liked his comment about the several different ways in which our data is fuzzy: the relationship *between* data items is fuzzy, so is the context of the data, and even the relationships between sources of data is fuzzy. No other industry faces these kinds of challenges. Ewan’s toolset for addressing these problems? Basically a shopping-list of the things I’ve been wittering on about in this blog: collaborative working; SOA; social networking; semantic web; agile development;open-source; non-proprietary tech. In contrast, what we’ve actually mostly got right now are monolithic systems: lowest-common-denominator functionality; limited interoperability; long development cycles. The final point – one I’d never considered before – is how difficult this all makes it for innovative new software developers to gain entry to the healthcare market. This talk felt in some way like a call to arms, inspiring despite (or perhaps because of) facing the difficulty of the problems we face.

(It occurred to me that patient demographics information is an area which is crying out to turned into some sort of wiki accessed through standard API which all system get read/write access to…)

This is far too long, shouldn’t have drunk so much coffee today, sorry.