jamesagnew

🧙Our Wizard release is jam-packed with updates, and @jamesagnew can't wait to tell you about it on Feb 16 at 10 am ET! Want a sneak peek? hubs.la/Q01Bgrjc0 Our releases are live-streamed on our YouTube channel. Subscribe so you know when we go LIVE! youtube.com/smilecdr

This online class introduces the #HAPI #FHIR framework for developing applications in Java. Only offered once this year on May 24-26, so java developers, don't miss out! Instructor is James Agnew, Project Lead of HAPI. Register at 👇bit.ly/36gcxvl

Hey @gitlab - Your status page shows everything green, but we can't log in this morning. cc @gitlabstatus

#DHIS2022 Make #fhir not paper. Thanks Brian Chapman

Oh no! You missed our quarterly product release webinar. That's okay, we recorded it. Check it out and learn about new improvements made to Smile CDR and #HAPI #FHIR #openstandards hubs.la/Q01404bM0

This morning we're releasing the first in a series on optimizing and benchmarking Smile CDR. First up: EC2+Aurora on AWS. smilecdr.com/benchmarking-s… @smilecdr #fhir #performance

A huge congratulations to the whole team who worked so hard on this. ISO 13485 certification ain't easy.

A quick note that HAPI FHIR and Smile CDR are not vulnerable to the #log4shell RCE issue that is making waves today (CVE-2021-44228). smilecdr.com/our-blog/a-sta…

@niquola @GrahameGrieve @ChristinGorman @linesele @Ettenaej The query cache is actually a drain on performance, it helps ensure fully stable searches even if data changes during paging. People at high volume generally disable it. This specific 10k/sec deploy is a big and fairly traditional cluster of on-prem VMs backed by Oracle RAC.

@jamesagnew @GrahameGrieve @ChristinGorman @linesele @Ettenaej Probably that's how you got 10K RPS - #search-result-caching" target="_blank" rel="nofollow noopener">hapifhir.io/hapi-fhir/docs… :)

Health IT people - has anyone managed to use FHIR at scale? With extensions? Is it physically possible to get decent performance when dealing with hundreds of thousands of messages simultaneously? If so how do you do it?

@ChristinGorman @GrahameGrieve @linesele @Ettenaej Yeah, that problem certainly resonates. Weve had existing models that were hard to map to FHIR because they had so many unusual data points that didn't fit cleanly and needed lots of extensions. Sometimes this is a smell that you've picked the wrong resource. Sometimes it isn't

@jamesagnew @GrahameGrieve @linesele @Ettenaej I don't know the details, but I think the performance issues I have been hearing about at work are mainly due to how cumbersome it was to map the external formats to FHIR. Lots of extensions were required, the data volume grew, complexity grew, performance suffered.

@ChristinGorman @GrahameGrieve @linesele @Ettenaej That said, if you're talking national EHR infrastructure, you probably don't control both sides of the connection. The trivial added cost of scaling up an extra server or 2 gives you an easy-to-adopt protocol. Personally I'd use FHIR every time. Horz scaling HTTP infra is easy.

@ChristinGorman @GrahameGrieve @linesele @Ettenaej Well, personally scales of 100k/sec sounds to me like streaming realtime device data where you have a ton of devices and control both sides of the connection. I'd probably personally be using an ultra compact binary protocol for that scale (parquet, protobuf, etc).

@GrahameGrieve @ChristinGorman @linesele @Ettenaej 2/2 Those are 10k operational transactions with a healthy mix of read/write in POS applications. I'm certainly not aware of anyone using HAPI FHIR at 100k+/sec scales but that also feels like an unlikely requirement for anything you'd use FHIR for in the 1st place.

@GrahameGrieve @ChristinGorman @linesele @Ettenaej Largest HAPI FHIR deployment I'm aware of peaks at servicing roughly 10k FHIR REST requests per second. Parsing am average JS resource typically takes on the order of 5-10ms, if you're seeing multiple seconds of latency for that something is very wrong with your setup.

Have your say on #hapifhir development and prioritization for the next year by filling out our 2021 HAPI FHIR Community Survey: bit.ly/hapi-survey @smilecdr

Looking to learn more about #FHIR? There are 2 great introductory and refresher courses available for FREE on @OHIEAcademy by @DigitalSQR and @jamesagnew with @smilecdr Check them out here: academy.ohie.org

HAPI FHIR 5.5 (Code name: Quasar) is only 1 week away! You still have time to register and join us live at 10am on August 19th. hubs.la/H0V1b830

We are proud to announce that Smile CDR has secured $20 Million in Series A funding led by 30 North Group and UPMC Enterprises to develop new innovations in healthcare and data management that extends beyond interoperability. Read the full story here: hubs.la/H0TPvsC0

@iamhasibrahman Sure thing. I'm jamesagnew on GitHub and my email address is on my profile there. Feel free to shoot me a note!

@jamesagnew I'm leading a large FHIR initiative and would love to speak to you about an area where your solution can be a great addition. Can we chat?

As rumors swirl around Apple Watch & BG monitoring, of course it depends on how accurate it will be. But, getting more glucose monitoring into more peoples' hands will be a good thing for societal health & wellness. *AND* the real big deal here is that data will be interoperable.