List of questions. for TREs to accompany release from sprint 2

[jim-smith]

To be included in document:

As researcher:

• what do you need to know
• what would make you use it
• something else
• should we call finalise if you forget?
• how do you feel about being prompted to provide more inputs to help SDc
• would you like a tool that took your outputs and turned them int the acro process?
• how do they feel about Data exports being unformatted? is. this a problem, do they ant to submit both?

As a TRE

• should we call finalise automatically if the user forgets
• should we prompt them more for inputs to help the process
• what do you look at when checking for secondary disclosure / differencing
• would a list of what variables a. used in which outputs be helpful?
• if the user just provided you with a directory full of docs/files, would you like a tool that added them into the acro process?
• or maybe this should be done stand-alone in the tool
• how would you like to handle/support a dataset-specific risk appetite file? Could you provide it as a read-only file in the same directory and you put the data? How much of a priority is that for you? What if the researcher is working on multiple datasets with different sensitivities? What do you do at the moment?
• how do they feel about Data exports being unformatted? is. this a problem, do they ant to submit both?
• do you have a set of filetypes you do/don't let out?

[jim-smith]

we need talk a little about how to handle different sessions via sub-foldfers

[jim-smith]

Do they record / log the interactions between TRE staff and researcher that lead to changes in an output prior to release?

• and if so , how?
• not hard to imagine this might be very different for each TRE, but let's see
• • feeds in to whole issue of long-term storage of artefacts and 2ndary disclosure/differencing

[mahaalbashir]

Here are some of the comments mentioned on the feedback from DaSH and DataLoch

Comment	Status
Prefer the SDC results to be written to an excel worksheet as the default format	Will further discuss this with TREs
Each run of acro.finalise() produces a separate timestamped folder containing all the outputs from that session	What is the plan for handling two acro session’s outputs?
Researchers should be strongly encouraged to rename their files. Having the outputs numbered was confusing	Do we want to do anything about this?
A table explaining the potential check outcome and action might be useful	We already have the check outcome in the summary field in the SDC results. Do we need to include an action as well? Something like (allow release, don’t allow release?)

[jim-smith]

people got confused by missing numbers when users had removed outputs

[jim-smith]

need to know more in practice about whether people use things from scripts