...

...

...

...

...

...

...

...

...

...

...

...

...

...

Note:

• UMCs may well provide their own local tools/software and may not need HRI for basic onboarding in the HRI Catalogue.

• If HRI does want to offer a tool for filling in the Metadata schema (core/petals):

  • CEDAR will not be used (discussed 24-10-2023, Marianne)

  • Data Stewardship Wizard (DSW) could maybe be used (Suggestion Rob Hooft).

  • Keep in mind that if the scenarios mention the DSW this could be another tool. No choice has been made.

Design

Two main types of scenarios

1. General Guidance Scenarios

   1. Actor needs information / knowledge

      1. Involves HRI reference pages (such as metroline), knowledge bases and tools offered by HRI

   2. If more help is needed, the actor contacts HRI Servicedesk

      1. It then becomes a Targeted Guidance Scenario

2. Targeted Guidance Scenarios - Involves HRI Servicedesk / experts

   1. Someone contacts HRI Servicedesk with a request for help

At the moment we’ll Focus on the General Guidance Scenarios

General Guidance Scenarios - Three main topics

1. (Guidance about) Making a Dataset Available in the Health-RI Catalogue

2. (Guidance about) Creating a FAIR Dataset from Scratch

3. (Guidance about) FAIRifying an Existing Dataset

General Guidance Scenarios

(Guidance about) Making a Dataset Available in the Health-RI Catalogue

1. Publish information about the dataset using a form-based system

   1. Core Metadata

   2. Core + Petals Metadata

2. Publish information about the dataset using technical semi-automated methods

   1. Core Metadata

   2. Core + Petals Metadata

   3. Core (+ Petals) + Custom Metadata

(Guidance about) Creating a FAIR Dataset from Scratch

1. Clinical Data - How do you make your Clinical Data FAIR from Scratch?

   1. Build Clinical Dataset in Castor EDC

      1. Simple, couple of fields

   2. Build Clinical Dataset in Castor EDC and Publish it in Castor's FAIR Data Point

      1. Simple, couple of fields

      2. How about we show the steps involved in setting up VASCA?

   3. Build Cancer Clinical Dataset using some form of predefined items (Plateau 2 if I recall correctly)

   4. Some other ideas:

      1. Publish the metadata it in a different FDP? Data in some triple-store?

      2. Multi-lingual - build an English and a Dutch dataset in Castor, publish them both, query them

      3. Different EDC, e.g. REDCap? Does it need a scenario or perhaps we can list relevant changes if there are few?

2. Omics Data - How do you make Omics data FAIR from Scratch?

   1. Build Simple Omics Dataset

   2. Genomics, Proteomics, etc. These could all have a separate scenario if necessary and we could then link to metroline pages for e.g. recipes on how to tackle the specific problem.

3. Imaging Data

   1. No Idea how this works

4. … Data - How do you make … Data FAIR from Scratch

   1. Build Simple … Dataset

(Guidance about) FAIRifying an Existing Dataset

1. Clinical Data - How do you transform existing non-FAIR Clinical Data to be (more) FAIR?

   1. Make an existing non-FAIR clinical dataset more FAIR

2. Omics - How do you transform existing non-FAIR Omic data to be (more) FAIR?

   1. Make an existing non-FAIR omics dataset more FAIR

3. Imaging Data - How do you transform existing non-FAIR Imaging data to be (more) FAIR?

• … Data - How do you transform existing non-FAIR … data to be (more) FAIR?

  • Transform …

Targeted Guidance Scenarios

• This will probably follow this type of structure (probably up to Services):

  • User contacts Servicedesk via a Form on the HRI website

  • Servicedesk does some form of intake

    • (maybe confirm whether the solution is not already available as general guidance)

  • Necessary HRI experts get involved (or perhaps bring in contact with external experts?)

  • Necessary next steps are decided (multiple iterations if necessary)

  • Problem solved

We need to keep track of the question / solution to see whether the question+solution occur more often and could be turned into a general guidance

...

Stop reading here… Below is information which probably has to be incorporated in the above. It also serves as scratch and contains previous ideas.

Scenarios - Making data Findable in the National Health-RI Catalogue

...

Proposal: have two main solutions. To stick with the Metroline Analogy: a Basic and an Advanced Track:

1. Basic Track - A fillable form-based solution, such as CEDAR or the DSW

   1. Most researchers will probably want to have a solution that takes little effort, but does meet the requirements imposed by employers, funders, etc.

   2. Health-RI provides easy to use forms with mandatory and optional fields as required by (initially) the core metadata schema and (later) extended metadata schemas

      1. Forms must have guidance to help researchers properly fill in the necessary data. Guidance could e.g. be (sections of) metroline pages

   3. Output should be something suitable for a FAIR Datapoint or Catalog (harvestable by Health-RI)

      1. Maybe provide the output directly to an FDP as an export option?

         1. E.g. You’re an Amsterdam UMC researcher it could suggest an Amsterdam UMC preferred FDP

         2. It would need some form of quality control / permission before the “ok, send it” button could be pressed?

2. Advanced Track - A manual solution [Dena: Advance track cannot be fully manual]] So I am wondering as we previously discussed we need to have a scenario where we define semi-automatic approach for extracting data, mapping and transformation and api creation]]

   1. The Basic Track does not meet your demands or you just prefer to do it manually

   2. Health-RI provides guidance (knowledge, references, recipes, perhaps scripts/software where possible)

      1. Core can probably have a follow-along example; for extended maybe pick one extension for a follow-along example

      2. Metadata elements outside of the core and extensions should have clear guidance where possible with also recipes for domain specific problems.

         1. If custom things can be generalised to apply to e.g. more domains, that would be better

         2. Also, it may be useful to keep track of custom items? If lots of projects add the same custom item, it could be useful to add it to a new version of core.

Scenario List

Basic Track

• BT1 - Make Simple Dataset Findable using Datastewardship Wizard - Core

• BT2 - Make Simple Dataset Findable using Datastewardship Wizard - Core + Extension

• BT3 - Make Simple Dataset Findable using Datastewardship Wizard - Core (+ Extension) + Custom

  • Probably Unfeasible: custom does not seem realistic in a DSW approach as HRI would need to create new forms for every project with custom items

Advanced Track

• AT1 - Make Simple Dataset Findable using Semi-automatic approach - Core

• AT2 - Make Simple Dataset Findable using Semi-automatic approach - Core + Extension

• AT3 - Make Simple Dataset Findable using Semi-automatic approach - Core (+ Extension) + Custom

...

• AT2.1 - Clinical

  • AT2.1.1 - Cancer

• AT2.2 - Omics

  • AT2.2.1 - Proteomics

• AT2.3 - Funders

  • AT2.3.1 - ZonMW

• AT2.4 - Rare Diseases

Scenario Ideas - Creating new FAIR Datasets and Transforming Existing Data to be more FAIR

• Clinical Data

  • Build Simple Clinical Dataset in Castor EDC and Publish it in Castor's FAIR Data Point

  • Transform an existing Simple Clinical Dataset in Castor EDC and Publish it in Castor's FAIR Data Point

  • (We could look into other EDCs as well if necessary. Design principles for FAIR should be the same)

  • Advanced versions: E.g. Build Cancer Clinical Dataset…

    • Could this be something for Plateau 2?

• Omics Data

  • Build Simple Omics Dataset

    • Genomics, Proteomics, etc. These could all have a separate scenario if necessary and we could then link to metroline pages for e.g. recipes on how to tackle the specific problem.

  • Transform an existing Omics Dataset

• Imaging Data

  • Build Simple Imaging Dataset? No idea how this works

• … Data

• Build Simple …

Transform …

Other types of Scenarios

• What does a researcher have to do if he/she wants to find data and then gain access to it?

  • Probably Architecture has this covered? TBD

...