FAIR data: What is it (and what isn’t it)?

Your research data is most likely to have a significant impact if you manage it in accordance with the FAIR principles. But what exactly do the FAIR principles stand for? And why are they particularly important for dementia research?

Do you work with data or collect it as part of your dementia research? There’s a good chance you’re already applying some of the FAIR principles – even without realising it. DEMPACT has summarised all aspects of FAIR data for you.

FAIR stands for Findable, Accessible, Interoperable and Reusable. This essentially means that research data must be findable, understandable and usable – by both people and computer programmes capable of analysing the data. But what exactly does that mean for your data?

1-Findable: retrievable forever

Data must be easy to find, both for people and for computer programmes that analyse data. To make this possible, you need to provide your research data with two things. Firstly, what is known as ‘rich metadata’. In other words: information about the data that provides context, making it understandable, easy to find and more useful. Secondly, you must assign your data a ‘persistent identifier’. This is a globally unique and permanent label, such as a Digital Object Identifier (known as a DOI). This is important to ensure that data remain findable at all times, even if the original link (such as a URL) changes.

How to do it

Many data repositories (physical storage locations for your raw data) and catalogues (centralised inventories) automatically assign a persistent identifier to your data when you publish your (meta)data. But not all of them do. To ensure good discoverability, therefore, choose a repository or catalogue that does this.

So which repository or catalogue is a good choice?

• Check with your research institution: most have a Research Data Management (RDM) page with advice on which repository to use.
• You can also contact the data stewards or RDM support team at your own research institution.
• Make use of databases such as FAIRsharing (see fairsharing.org or re3data.org); there you will find many repositories and catalogues.
• Keep an eye on dempact.nl: DEMPACT compiles the most relevant catalogues and repositories for dementia research in a guide.

2-Accessible: accessible data is not the same as open data

Accessible dementia data, in the context of FAIR data, is not necessarily public or open data. Privacy-sensitive (patient) data, for example, must remain protected. The guiding principle is: as open as possible, as closed as necessary.

This conditional accessibility is also something the repository facilitates. This is achieved through a technical protocol that regulates access to the data.

3-Interoperable: easily exchangeable

You make data exchangeable by storing it in a format that can be used by different systems. To achieve this, it is important to use standardised language and terminology.

This is particularly essential in dementia research because dementia data often comes from different sources. Consider, for example, data from different research groups, hospitals or GP practices. Or different types of data, such as brain scans, genetic information and so on. The use of standardised language and terminology is also crucial if you wish to analyse data from different cohorts in conjunction with one another.

To date, no standard exists for this. This makes combining data in dementia research (as yet) difficult. You can read about why this is the case and what DEMPACT intends to do about it in ‘Making research data FAIR… How?’.

4-Reusable: reusable in the future

Data are reusable if you document them in such a way that they can be used without difficulty for future research. For example, for a meta-analysis or for testing new hypotheses.

To achieve this, you must release the data under a clear and accessible licence that specifies what you may and may not do with it. In addition, you must ensure that the origin of the data is transparent (so-called provenance metadata). And your metadata must comply with so-called domain-specific standards; examples include SNOMED CT for clinical terminology and the OMOP Common Data Model for clinical and observational health data.

Benefits of FAIR data

Managing your data in accordance with FAIR principles does therefore require some attention. And you’ve already got enough on your plate… Do the benefits really outweigh the extra work? Absolutely.

Increase your impact

FAIR data management increases the scientific impact of your dementia research. Well-documented data are – quite literally – more visible. This ensures that your research is mentioned more often; for example, you will be cited more frequently in publications. Visibility, in turn, opens the door to valuable collaborations with fellow researchers.

Accelerates your dementia research

Collecting research data is a time-consuming process. Some studies collect data from people with and without dementia over the course of decades. If this data is well documented and accessible, you and fellow researchers can make use of it. This means you no longer need to collect any, or very little, data yourself. As a result, you can arrive at new insights more quickly. Avoiding duplication of effort also means significant time savings for dementia research as a whole. And that is no luxury now that the number of people developing dementia is on the rise.

Greater chance of breakthroughs

Major breakthroughs in dementia research often stem from international consortia. FAIR data management makes such breakthroughs a practical possibility more often. Thanks to FAIR data, datasets from fellow researchers worldwide can be combined; this allows you to measure smaller statistical effects. This is crucial, for example, for research into certain rare causes of dementia.

Obstacles? DEMPACT can help

Is managing your data in accordance with FAIR principles essential? Yes. Easy? No. But bear in mind that you’re not alone, and that there are resources available to help you (Making research data FAIR? How?). The methods for managing research data in accordance with FAIR principles are also evolving rapidly.

Are you facing obstacles? Please contact Didi Lamers, data management coordinator at DEMPACT, at d.lamers@dempact.nl. She’ll be happy to help.