# Healthcare Interoperability Architecture - Seeking Expert Feedback

**Category:** [Implementation](https://discourse.openehr.org/c/implem/39)
**Created:** 2025-08-15 14:12 UTC
**Views:** 96
**Replies:** 2
**URL:** https://discourse.openehr.org/t/healthcare-interoperability-architecture-seeking-expert-feedback/7046

---

## Post #1 by @Juan_Fajardo_Reyes

Hii everybody!!

We're designing a healthcare interoperability platform to integrate legacy Laboratory Information Systems (LIS) with modern Electronic Health Record (EHR) systems and external applications. The goal is to create a **semantic-preserving, standards-based architecture** that can scale across multiple hospitals.

Current Context:

* **Source Systems**: Multiple LIS systems producing HL7 v2.5 messages (ORU^R01, ADT^A08, etc.)

* **Target Consumers**: EMR systems, mobile apps, HIE networks, regulatory reporting

* **Requirements**:

  * 100% clinical semantic preservation

  * Multi-standard interoperability (HL7, FHIR, openEHR)

  * Longitudinal patient records with versioning

  * Scalable architecture for national deployment

Could you help me to improve this dataflow diagram??? Thanx so much!!
![image|690x412](upload://5bMpjHw7Nq3xkO2Q7O5oU3Bb6Ub.png)

---

## Post #2 by @ian.mcnicoll

That looks about right to me.

You might want to add in a reporting (probably SQL) DB to which you offload reporting and analytics via an ETL process. Native clinical data is tree-shaped and does not play as nicely as you might like with BI tools and population queries. Interestingly, ChatGPT tells me that EPIC does exactly the same - it has an operational datastore based in Cache but a separate RDBMS reporting datastore. AQL makes it fairly easy to populate the ETL, with a little post-processing.

---

## Post #4 by @Juan_Fajardo_Reyes

Thank you so much for your time and detailed explanation. I completely agree — clinical data in tree-shaped formats can be tricky for BI and aggregate queries, so having a relational reporting database makes a lot of sense. Using AQL to extract data into an ETL, with a bit of post-processing, seems like a very practical approach.
Best regards!!

---


**Canonical:** https://discourse.openehr.org/t/healthcare-interoperability-architecture-seeking-expert-feedback/7046
**Original content:** https://discourse.openehr.org/t/healthcare-interoperability-architecture-seeking-expert-feedback/7046