News: The HAPI FHIR Blog

HAPI FHIR 4.1.0 (Jitterbug)

Published: 2019-11-13T15:00:00
By: James

It's time for another release of HAPI FHIR!

This release brings some good stuff, including:

  • Structures JARs have been updated to incorporate the latest technical corrections. DSTU3 structures are upgraded to FHIR 3.0.2, R4 structures are upgraded to FHIR 4.0.1, and R5 draft structures are upgraded to the October 2019 draft revision.

  • ValueSets are now automatically pre-expanded by the JPA server into a dedicated set of database tables. This "precalculated expansion" is used to provide much better performance for validation and expanion operations, and introduced the ability to successfully expand very large ValueSets such as the LOINC implicit (all codes) valueset.

  • Support for the FHIR Bulk Export specification has been added. We are now working on adding support for Bulk Import!

  • First-order support for ElasticSearch as a full-text and terminology service backend implementation. At this time, both raw Lucene and ElasticSearch are supported (this may change in the future but we do not have any current plans to deprecate Lucene).

  • Live Terminology Service operations for terminology file maintenance based on delta files has been added.

  • Binary resources and Media/DocumentReference instances with binary attachments stored in the FHIR repository can now take advantage of externalized binary storage for the binary content when that feature is enabled. This allows much better scalability of repositories containing large amounts of binary content (e.g. document repositories).

As always, see the changelog for a full list of changes.

Thanks to everyone who contributed to this release!

Also, as a reminder, if you have not already filled out our annual user survey, please take a moment to do so. Access the survey here: (note that this URL was originally posted incorrectly. It is now fixed)

Tags: #Release

HAPI FHIR 4.0.0 (Igloo)

Published: 2019-08-14T15:00:00
By: James

The next release of HAPI has now been uploaded to the Maven repos and GitHub's releases section.

This release features a number of significant performance improvements, and has some notable changes:

  • A new consent framework called ConsentInterceptor that can be used to apply local consent directives and policies, and potentially filter or mask data has been added.

  • Initial support for draft FHIR R5 resources has been added.

  • Support for GraphQL and the _filter search parameter has been added.

  • The ability to perform cascading deletes has been added.

As always, see the changelog for a full list of changes.

Thanks to everyone who contributed to this release!

Tags: #Release

The Growth of HL7 FHIR

Published: 2018-02-11T15:06:00
By: James

One of the things we often talk about in the FHIR standards development community is where FHIR currently sits on Gartner's Hype Cycle. The hype cycle is a coarse measure of the trajectory of new technologies on a journey from being "new and exciting silver bullets" to eventually being "boring useful technologies".

When you are a proponent of a new technology (as I certainly am with FHIR), probably the most important aspect to remember about the hype cycle is that you really only ever know where you are at any given time long after that time has passed. In other words, it's fun to ask yourself "have we passed the Peak of Inflated Expectations yet?" but you really won't know until much later.

Speculating is perhaps a fool's errand. I probably shouldn't try but I can't help but wonder if we have passed the peak yet.

The trajectory of HAPI FHIR's growth is interesting. FHIR has been growing over the last few years by all kinds of metrics. The connectathons keep getting bigger, the number of vendors participating keeps on getting bigger, and FHIR DevDays keeps on getting bigger.

If I look at our website in Google Analytics, I am curious about the trajectory.

While HAPI FHIR has seen pretty steady growth over the last few years, that growth has been either tapering or at least very unstable over the last 8 months.

Certainly I don't think HAPI FHIR has stopped growing. The number of messages on the support forum and the number of people with big production implementations these days certainly doesn't suggest that; however, things have certainly been weird the last 8 months.

Let's look at interest in FHIR overall. The next thing to look at is the FHIR Google Trends graph, which measures the number of people searching for terms on Google (a pretty decent indicator of general interest). The following graph shows the last 4 years for FHIR.

It would seem that FHIR itself saw a crazy explosion of interest back in May, too. That makes sense since FHIR R3 was released right before that peak.

Let's compare that with the graph for IHE. I don't think anyone would disagree that IHE sits firmly atop the Plateau of Productivity. Most people in the world of health informatics know what can be accomplished with IHE's profiles, and certainly I've worked with many organizations who use them to accomplish good things.

The FHIR and IHE Graph shows interest in FHIR in BLUE and IHE in RED.

So what can we take from this? I think the right side of the graph is quite interesting. FHIR itself has kind of levelled off recently and has hit similar metrics to those of a very productive organization.

I probably shouldn't attach too much meaning to these graphs, but I can't help but wonder...

Tags: #Hype Cycle, #HL7, #HL7v2, #FHIR
Newer Entries Older Entries