[MA 2025 06] How to exchange data using post-and precoordinated data in different data models

Nictiz, Hybrid, The Hague (The Netherlands)

Proposed by: Louise Witteman [louise.witteman@nictiz.nl]

Introduction

The exchange of medical health information is increasing rapidly. Much effort is put into creating standards for exchange between organizations, specialisms, and even across the EU. Each standard contains several value sets. A value set is typically used to represent the possible values of a coded data element in an information model. It is a list of terminology codes accepted for a particular dataset concept.

Information standards provide structure by defining different data fields. A data field displays variable data and allows the user to enter data; terminology code systems provide a single, unique code for a concept, e.g., ‘left foot’. The boundary between the structure provided by an information standard and terminology is fluid and depends on the use case, known as the boundary problem (Rossander & Karlsson, 2023).

Example

For instance, a plain X-ray of the left foot contains three pieces of information and thus may be recorded in different ways, e.g.:

Example data field with a pre-coordinated terminology concept

Procedure = 426908009 |Plain X-ray of left foot (procedure)|

Two examples with data fields that give combined the same information. Known as post-coordination

Procedure = 168537006 |Plain X-ray (procedure)|

AnatomicalLocation = 22335008 |Structure of left foot (body structure)|

Procedure = 1290452004 |Plain X-ray of foot (procedure)|

Laterality = 7771000 |Left (qualifier value)|

The first option records the entire concept as a single code. We refer to this as pre-coordination. Pre-coordination is easy and fast to use but may lead to many concepts to choose from in an application. Dividing the concept into multiple pieces is called post-coordination. Post-coordination requires the healthcare professional to record more pieces of information, but it gives greater expressivity with a smaller number of concepts. The examples above shows that data is entered differently but means the same. The preferred approach depends on the use case and the user preferences. A radiologist would make finer distinctions in imaging procedure types and would maybe use more post-coordination than a GP, whereas a GP would record many other procedures which could be expressed with pre-coordination. The preference for pre- or post-coordination depends on the use case.

Forcing different professionals to agree upon one way of data registration has the following disadvantages:

Agreement among the various stakeholders would take a long time

It could generate a lot of ill-will

It will produce a list that fits neither group very well.

A potential solution would be to map between the pre-coordination and post-coordination codes.

Description of the SRP Project/Problem

Nictiz develops standards for the exchange of healthcare information. These standards are composed of Health and Care information models (HCIMs). These HCIMs function as the building blocks of information standards. They set a national standard for the structure of healthcare information that must be adhered to when exchanging information.

Nictiz intends to explore in a proof of concept how HCIMs can be used by organizations that have drawn different boundaries between information and terminology standards. In other words, it is an investigation on how to convert a pre-coordination to a post-coordination and vice versa. We want to create a working demo and a manual for software vendors that wish to implement such functionality. However, we also wish to determine the limits to conversion between pre- and post-coordinations. Which problems does it solve? For which applications is it feasible, and when is it not feasible? What are the pitfalls?

Research questions

How can organizations exchange pre-coordinated data with other organisations that post-coordinate data so that the system understands that the (combined) data fields mean the same?

For which applications are conversions feasible?

Are there datatypes and situations where we should avoid conversion between pre- and post-coordination?

What limits conversions between pre- and post-coordinations?

Does this solution solve the problem described above?

What resources should Nictiz publish to facilitate both sender and recipient to adopt post-coordination.

Expected results

Proof of concept is a recorded demonstration where pre- and post-coordination are converted, including a manual for software vendors that wish to implement such functionality.

Time period, please tick at least 1 time period

November – June

May - November

Contact:

Louise Witteman

terminologie@nictiz.nl

References

Rossander & Karlsson (2023): Structure of Health Information With Different Information Models: Evaluation Study With Competency Questions, JMIR Med Inform 2023;11:e46477

URL: https://medinform.jmir.org/2023/1/e46477

DOI: 10.2196/46477

Health and Care Information Models HCIM Mainpage - Zorginformatiebouwstenen