Top 3 EHR Data Integration Challenges


In response to a guest post on EMRandHIPAA, we take a look at the top EHR data integration challenges faced today:


Proliferation of point-to-point interfaces instead of using a hub-and-spoke type of model (like that which Surescripts utilizes with electronic prescribing). Unfortunately, most organizations which exchange data in and out of the AE-EHR utilize highly-customized point-to-point interfaces for orders, results, documents, etc. The point-to-point model is highly inefficient and does not adhere with a “plug and play” model that so many organizations desire.

We’ve seen Allscripts make an effort to move away from this by introducing capabilities to automatically send immunizations to state registries via the Allscripts Hub by  simply modifying configuration setting (with the caveat that Allscripts has worked with the state to develop the intergration).  We’ve also witnessed companies like Medicity and its Novo Grid technologywhich offers electronic communication between physician practices, hospitals, and other health care providers. Novo embeds agents (small but powerful Java programs) in hospital data centers, physician practices and other locations. The grid component is an object oriented system that can replicate an object to multiple agents and keep it in sync across locations.


As outlined in the EMRandHIPAA post, there are no mandated standards for EHR vendors to follow, thus making it difficult to coordinate data sharing between medical devices and other systems. Allscripts does offer the Universal Application Integrator (UAI),  which facilitates extendibility to other applications and devices. However, there is a certification process that needs to be pursued. In terms of the point-to-point interfaces previously mentioned, the Allscripts proprietary (API)  Application Programming Interface(which consists of inbound and outbound stored procedures to their primary clinical DB) does not segment out the data and configuration components of clinical exchange, something touched on in detail in a previous Galen Blog post.  Lastly, most vendors have their own specifications for HL7 message definitions. For instance, Quest may send ordering provider in OBR-16 in an interfaced result ORU message while LabCorp sends this in ORC-12. Another example is communication of “Ask at Order Entry” questions – something Quest expects to receive in repeating OBX segments while LabCorp expects this across Z-segments in an interfaced order ORM message.

Adherence to HL7, proprietary approaches.


John Halamka bravely predicted that when health IT vendors and providers began adopting new standards, the cost for interoperability would plummet: “We know that we won’t get precisely plug and play—this is a journey,” Halamka told Government Health IT. “But each year, we will get more constrained. We are going from a $20,000 -$30,000 venture hopefully to $5,000-$10,000.” Unfortunately the numbers quoted are accurate – and provide a high barrier to entry for smaller groups looking to electronically exchange data. There is the flip-side to cost and that is the ROI, which could include reduction in direct annual labor costs, elimination of non-billable tests, and elimination of lost charges.


The benefits of health information exchange are well documented. As outlined in the EMRandHIPAA post, there is a need for a “consistent, secure and reliable way to capture and share patient data among all systems and healthcare providers,” especially given that benefits in improved coordination of care and reduction of medical errors.

Facebook Twitter Email

1 comment

Add yours
  1. 1
    John Lynn

    Sadly, I don’t see any of the standards you mention really reaching the type of interoperability we need. HL7 is the most widely adopted standard. However, like you mention, it’s not very much of a standard since everyone does it so different.

    I still don’t see the path to true healthcare interoperability. I’m still hopeful that something will appear.

+ Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.