OpenEHR with SNOMED,LOINC,HL-7 Support

Hi,

We are a new startup and trying to implement a small EHR project by adopting an OpenEHR framework. Just wondering if most of the OpenEHR frameworks have the interfaces with SNOMED,LOINC and HL-7/FHIR by default or one has to build them?

Dear Prabhat,

The simple answer is no.

To answer the question about SNOMED & LOINC, OpenEHR archetypes and templates allow you to manage data coded using code libraries like SNOMED CT, LOINC & ICD. It is upto the applications to decide what codesets to use for coding their data.

HL7/FHIR are data exchange standards. You have to use OpenEHR AQLs to extract data from an OpenEHR repository and expose them as HL7/FHIR interfaces for sharing data in your OpenEHR server. For incoming data, you will have to create custom tools to map them to your templates and commit.

regards

1 Like

Dear Dileep,

Thank you for the reply.

Allow me to lay out my basic plan. I am helping a senior citizen group members to maintain their “Personal Health Records” in the electronic format and access them any time. So it’s primarily from the patient/user POV system. Naturally, they will maintain the records from the multiple providers as they come in on different dates. We will keep building the system organically as the service reports flow in from different providers for different members. We will maintain the paper trails for next 2 years so that the members can refer to them anytime if they have any doubt at any point of time and we will keep building the system organically. But from the day one, we would like to introduce the standard coding systems i.e SNOMED,LOINC and store data in HL7/FHIR format, even if we have just one member. We will be hosting on a HIPAA compliant platform like AWS. Ofcourse, we would like to integrate the standard API’s so that the providers can consume the data of our members seamlessly.

Please suggest an OpenEHR platform, if any, that I can start with immediately and get the members onboard quickly so that they start storing their lab test reports. I personally would prefer a heavily API driven, JSON based system which must be modular and allow us build procedurally.

Regards

Prabhat MOHANTY

Hi Prabhat,
from the above, I would say you have a major design decision in front of you. The normal role of HL7 FHIR (and HL7 v2, etc) is to provide a way to represent data extracted from source systems, if no convenient native format is available. However, to persist a longitudinal patient health record, you will need an architecture such as openEHR rather than a message format like FHIR. If you want to persist your patient-centric data in FHIR, you are building some kind of FHIR cache, but it is unlikely to function as an EHR. Reasons for this are described in this blog post.

1 Like

Dear Prabhat,

OpenEHR is not an FHIR repository in the sense that it does not follow a resource paradigm like FHIR. OpenEHR is more flexible with the way you store and process data. As I had mentioned the HL7/FHIR support can be built on top of an OpenEHR repository for sharing data with non-openehr systems.

Please have a look at https://openehr.org/downloads/platform/ for some opensource and commercial implementations. In India, we also provide a cloud based OpenEHR platform called EHR.Network aimed at application developers.

https://openehr.org/ckm/ will be another great resource for you. Here You can get a large number of peer evaluated and curated archetypes that can be used with OpenEHR systems.

regards

2 Likes

Thanks. The article was very informative.

But I also had this nagging feeling about the 80-20 usage - the regular users used barely 20% of the Microsoft Word feature sets to be productive. The doctors, in real life situations, also mostly use the snapshots, not the continuous data, for the decision making in the 95% cases.

If our members could discard the paper based reports after sometime and are able to share their reports electronically with their doctors, we would achieve our immediate goal. They are also excited about storing the data from their health devices on the cloud.

Thanks for the help. We definitely need to investigate and deliberate further.

Regards

Prabhat

Hi Prabhat,

we are currently working on support for FHIR Terminolgy Service in EHRbase. It will hopefully be available within the next three months and would allow you to validate your data entries (so called compositions) based on LOINC, SNOMED etc.

The colleagues from the national platform in Scotland have tested that EHRbase runs on Azure (using advanced features/add-ons for Postgres like Citus). Hence, using AWS might work as well. It has all needed endpoints of the openEHR REST API and a client library can also help to use a human-readable data structure in Java respecively JSON.

This might not be everything you need right from the start but there is steady progress. There will also be some FHIR Interfaces but the data is stored in openEHR format.

Hope this helps,

Birger

3 Likes

Thanks Birger. With the input like this, things are becoming clearer. Highly appreciated. I am studying EHRBase closely.

Regards

Prabhat

Hi, we had in fact integrated SNOMED expressions into the EHRServer openEHR Query Component, check this guide that includes a demo video https://www.cabolabs.com/blog/article/openehr__snomed_ct_a_perfect_combination_for_data_querying-5a440acd0f763.html

3 Likes

I am curious to know how do you search the SNOMED CONCEPT ID through your API for a given description? Any sample API?

You mean like a partial or total search over a snomed term that matches with one of the synonyms?

If I understand correctly, you are looking to search through snomed using some terms and find the relevant snomed.concept id.

That would need a snomed terminology server.

Regards

1 Like

Yes, right!

We need the relevant SNOMED ConceptID for the complete unambiguity. Else, what exactly do we input into the EHR?

Check out our API based SNOMED Mini search engine (Beta Version) implementation at www.24access.in/smsearch_new.html . It’s very recent. Please give your feedback. Tell us if the data inputing into the EHR is easier with our system or not ?

[Check Disorder > “rumatic heart” ] for example.

Select “Disorder” as the Semantic TAG and query “rumatic heart” and get the Concept ID instantly.

Searching a terminology needs some kind of terminology services. That is not related with openEHR per-se since that is mainly integrated with clinical recording applications, not with openEHR backends.

Then when you have the codes recorded in an openEHR backend, querying for coded terms usin SNOMED CT constraints might be related to openEHR, but depends on the use case. In the EHRServer we support to query by SNOMED CT individual codes, lists of codes or by using SNOMED expressions, and by exact or partial rubric/concept name (free text search will be available soon). The backend doesn’t care on how those codes were recorded.

1 Like

Last year the SNOMED team had worked very hard and the free text search is now available on the official SNOMED International Browser. It’s really awesome now.

Also check our own mini SNOMED search engine at www.24access.in/smsearch_new.html and give your feedback.

You are answering your own question with a reference to your own product… strange.

Honestly, there’s nothing strange about my approach. I am actively looking for a reliable, standard based and extendable platform.
Whenever, I get a suggestion I go through the site and check the features available and try to visualise how the existing features could be extended to match my vision. I am familiar with SNOMED because I wanted a standard based system from the very beginning and since there was no readymade solution available I quickly designed an interface. When you mentioned about your interface with SNOMED I got curious thinking I might not have to work on a connector for SNOMED with your system. Hence I asked if I could check your API.

Please do not misunderstand me. I am an entrepreneur, not a regular programmer. A year back, I came back to IT after 25 years and digital healthcare is the first project I have picked up. I have been looking for a platform that’s closer to my vision of “Standard based easy healthcare management system”. In the recent weeks, I have been working on the connectors for different standards in the domain. I do not have any product yet. Just a few components that will come together later on a platform. If I get a pure JSON based EHR platform, then it will be easy for me to integrate the missing pieces.

Hope this clears any misconception about my interaction with you and the beautiful people in the OpenEHR eco system. I genuinely need guidance at every step.

Regards

Prabhat

Selecting hierarchy is basic, as terms with same rubric exist in snomed. Also, not every snomed term applies in a given place in your UI. Subsets, intensional or extensional defined, are something you want to take into account.
Have a look to FHIR valueset resource for the usual attributes and parameters

1 Like