Disclose healthcare data

Owned by Health-RI
Aug 23, 2024
4 min read

datE: 23-08-2024 Status: ADOPTED

nlvlag48x34.png

 

Within the health data infrastructure for research, policy and innovation, we distinguish two main streams of health data

  1. Healthcare data; Real World Data from healthcare information systems intended for the implementation of care. (see #4 from Overall picture of the joint health data architecture model)

  2. Other health data (see #5 from the overall picture of the joint health data architecture model); This is the same type of data as healthcare data, but

    1. Data from other information systems; for example, data from Electronic Data Capturing tools for prospective research.

    2. As enriched, prepared and interpreted Real World Healthcare data

The Health-RI ecosystem is in line with the solutions described in the National Coverage Network initiative of the Ministry of Health, Welfare and Sport. This provides a single, standardized interface with the Unity of Language and Technology, with which healthcare data is made available. This page describes the main features of the National Coverage Network and what needs to be added to it for research, policy and innovation.

Nationwide Healthcare Data Network

The National Healthcare Data Network provides an integration layer on which healthcare data is made available in a consistent uniform way for multiple use, including research, policy and innovation. The different healthcare information sources can connect to this integration layer in different ways;  via platforms, or nodes and directly. In the short term main focus will be on the CumuluZ concept platforms. In this case, data from different health care systems (XISs) is duplicated to data platforms with a standardized vendor-neutral data model (according to Unity of Language) for one or more linked data platforms:

  • Data platforms are a central place

    • where healthcare providers can find information

    • that allows citizens to access their data

    • where the data is available for secondary use. This is in addition to the storage of data in the source systems.

  • A platform strategy does not have to be limited to primary healthcare data, but can also include data outside healthcare, such as the social domain, so that population management can be effectively implemented.

  • A data platform also offers opportunities for innovation. This is a facility where algorithms can be developed/executed and knowledge about them can be shared. Furthermore, a data platform can make (pseudonymised) data available for secondary use. However, the requirements for making data accessible for research and prevention must be met. For example, OMOP is used for medical research.

 

Disclosing data in the context of research, policy and innovation

The overall picture of the joint health data architecture model can be translated into the picture below, with an additional interface shown from the platform to the Researcher Data Platform.

image-20240503-105920.png
The blue box within the 'Multiple use' module includes the CumuluZ concept and serves healthcare applications (the blue box within the ''Apply' module) with data from 1 data subject via an N=1 FHIR interface. 'N=1' here means that it concerns making data available at patient level. For N>1 links (data about multiple patients), an API is provided that CumuluZ approaches from a different angle to make data available from healthcare systems for secondary use (as indicated in the yellow areas in the ''Apply'' module).

An interface with open standardized APIs, which are based on an application-independent data model (Unity of Language) for unlocking data in the context of research, policy and innovation, imposes a number of additional requirements with regard to unlocking data in the context of healthcare .

  • The data must be made easy to find; With a FAIR Datapoint, the data contained in a healthcare information source is described on the basis of standardized metadata. With which these can be published in catalogs and made findable.

  • It must be possible to request a dataset (= specific part of the data, needed for research, policy and innovation);

  • A dataset must be able to be interpreted and reproduced on the basis of a unique persistent identifier (PID)

  • A dataset must be made available, where the data

    • is minimized horizontally; only include those data points that are necessary for specific research, policy or innovation

    • is vertically minimized; exclude data subjects for whom an Opt-Out applies

    • is pseudonymised

    • Complies with a mapping within Unity of Language; Unity of Language stands for standardizing AND harmonizing application independent data model. Different application-independent data models may therefore apply to different application areas, which must be mapped with minimal loss of quality using a generic facility.

 

 

 

 

 

Â