FAIR Data Services Version 1.23.0

We want to make the DRE fit into a wider ecosystem, and help customers create their own platforms. As part of this approach, we have added a new option for extensibility: customers can reconfigure the existing link to the Aridhia knowledge base with a link of their own - for example to a community wiki or similar. Links to the knowledge base are still available from the search box in the banner at the top of the page. In addition, customers can define additional links that can appear in the results of the banner search box, if there other systems to connect users with.

For UK customers and others, sharing metadata with the HDR UK Gateway is an important step to building up the research community. Syndication, the process of sharing updated content with partner sites, is a key form of collaboration in a healthy research ecosystem.

We have extended the FAIR API to facilitate syndication from the DRE to the Gateway. Standard FAIR metadata is presented in the format required for the Gateway, and customers can use existing role based access control and API keys to control access. This will provide the Gateway with the ability to pull metadata on a regular basis.

Cohort Builder is a major new feature in FAIR that adds the ability to select and request subsets of data. This release introduces the preview “alpha” of the Cohort Builder. Customers can selectively test the functionality using role based access control and we will be gathering feedback while we work to develop the user experience.

Data owners remain in control of how data is accessed. In this version, they can set whether to allow the cohort builder module for their dataset. They can also set whether to allow visualisation tools and estimated record counts. We are interested in feedback on use cases and governance for this feature - should there be a researcher approval step before accessing aggregate summaries?

Researchers may filter dataset across multiple tables, run simple visualisations and get estimated record counts and entity counts. Filters are defined using a form of boolean logic - they are built up as multiple clauses in a query. Multiple ‘sub-cohorts’ can be defined, allowing comparison of different groups.

Estimated record counts support feasibility analysis. If the data owner has marked a field as representing an entity, then unique counts of entities are also provided as well as record counts. This allows the researcher to see the number of individual participants or patients, or the number of samples in a filtered dataset.

Researchers can visualise distributions of data. Categorical variables can be shown in a bar chart, continuous variables in a histogram. A box plot provides statistical properties of a variable, and can be split by categorical variables. Visualisations can be downloaded. We would like feedback on the access control and features around

A cohort definition can be saved for future use, and it can be used for a new data access request. We are working on having the selection delivered to a target workspace on approval. We are interested in feedback on how best to deliver sub-cohorts - as labels in the same table or as separate tables.

Please note there are some known limitations with the current release

• Cohort Builder visualisation and record counts depend on the existence of data and alignment between the data stored and the metadata (dictionaries).
• Data Access Requests can be initiation from a Cohort selection but the whole dataset will be transferred for now. If you want to test the selection delivery, please drop us a request at servicedesk@aridhia.com
• Sharing of cohorts is experimental - these should only be shared with data stewards.

We are rolling out a set of features to allow customers to define catalogues templates with domain-specific features that go beyond the core. This change tries to clarify the language used to describe these templates. The default catalogue template will now be known as the “standard” catalogue template in the system and documentation. (An older core template is now referred to as the “legacy”)

Customers can then extended the foundation provided by the standard catalogue to define their own template and apply that across all datasets.

Password managers and browser autocomplete features are increasingly important and convenient for users. On some browsers, we have been able to fix an issue with the dataset editing form where browser autocomplete could cause an unwanted edit, replacing the dataset name when the author name was being edited. Due to some limitations in the underlying Chromium browser engine, this fix will not apply to Google Chrome or Microsoft Edge browsers.

Audit and activity logs provide data owners and other stakeholders with important information about the use of their metadata and data. This enhancement improves the way data access requests are linked to a dataset and presented in the dataset history. This allows a data owner to navigate directly to the detail of a request (and the requesting user) when it is logged.

FAIR provides an extensible workflow engine for the business process used in handling data access requests. This enhancement improves the way emails are sent by the workflow engine. For example, in some of our integrations, this is important in case the user who requested the data needs to take a manual step.

We fixed a bug where some users had difficulty uploading metadata, attachments or data because of session time outs. This might occur if there was a long period of inactivity but the browser window was left open, or if a data transfer was taking too long. If a session has expired before upload, or during the upload, the system will attempt to refresh the session and ensure upload can proceed.