I’m aware from several (offline) conversations and on twitter that most developers feel a number of NHS API’s have issues.
For example on this site it’s common for people to ask questions about IM1 and Kettering XML. Some feedback I’ve seen also indicates most use these instead of GP Connect and Transfer of Care.
What developers really want from suppliers and NHS is simple RESTful, resource and OAuth2 based API’s, not complex clinical models.
For example:
When I’ve worked with EMIS and TPP, my interfaces for application developers to use is roughly the same as EPIC and Cerner - it is a plain bog standard FHIR (restful, resource based), OAuth2, API with no funnies.
All pretty straight forward and in the open. It’s also not that different to what I was doing 25 years ago in mumps and then later on in SQL.
I’ve never heard of a front line practitioner asking for anything that required anything more complex than this. Yep it does get complex but that is handled differently (for different reasons) and generally wouldn’t be an API we expose externally.
So is this a good candidate for a forum or should others be used like
I would tend to avoid specs where FHIR gets more complicated and starts to resemble HL7 v3 or CDA. I’m expecting to hear a big focus on sharing data (for all practitioners) rather than using it for ETL or record transfers processes (for individual use cases).
I’m very happy for this forum to be used as an unofficial NHS and Social Care Developer Forum, absolutely, and in some respects it is why I started the forum in the first place, as there was nowhere else on the Internet actively discussing NHS/Social Care IT in an open way, publicly. The NHS Hack Day Google Group had a few interesting discussions on it, but the owners were keen that it should only be used for discussing things that were about NHS Hack Day and not general NHS IT topics.
The official NHS Developer Forums are run by the Spine Team, who I know well and am friendly with, and I worked with them to get that forum set up. But as ever with NHSD/E, all it takes is for a key person to leave the org, or funding to be withdrawn and the damn thing will disappear overnight. So maybe we should use both?
I’m happy to make changes to this forum to facilitate any of those discussions - eg making new categories for types of discussion, and if there are any other people who would like to get involved as @moderators or @admins I would be potentially amenable to this. This forum is funded for public benefit by my company Baw Medical Ltd, although it doesn’t cost much to run, just the server fees.
I don’t think in general we are not that open and it does hinder development at several levels. For example on a current project it was difficult to find out:
What practitioners were doing with patients - we had to reverse engineer several clinical pathways. NICE, Diabetes UK, GIRFT was useful though.
What tech suppliers, architects and NHS providers were providing to support these pathways. NHS X had some useful use cases but again we had to reverse engineer and in started resorting to US open documentation from IHE instead.
What developers + business analysts were doing to support these designs. Is an element here we started following US designs, especially on the use of authorisation and API standards. Also what is the data being captured openEHR / Ian McNicoll is always helpful here.
High level clinical data models are documented, usually by PRSB but these need the above elaborated to be implemented.
Everything seems to be following organisation and so management structures, especially communication and feedback.
The end result is we found most interop was supported by email/fax interactions. Practitioners were capturing data in fairly consistent ways (using a mix of forms/templates) which were converted to PDF for sharing with other practitioners.
(note: Kings Fund and INTEROPen do seem to be doing good work either highlighting this or trying to get people together to solve it).
Where tech seems to be most integrated appears to be on patient held records (PHR). Where this links into health, it again reverts to PDF documents.
The levels of openness is quite good in citizen PHR apps/devices and in health only from US based suppliers.