selectionBehavior - beyond at-most-one

[dennispatterson]

On Zulip, @mattvarghese suggested additional values such as 'any' or 'all'. At the September connectathon in Atlanta, @alexander-ivanov expressed a similar desire to have multiple suggestions where the user may not be limited to one. I've invited them both to share more about their use-cases on this issue.

After the merge of #91 2 years ago (to limit to at-most-one), there was discussion about expanding this, as well. Opening up this issue to pick up the discussion. cc: @kensaku-kawamoto @bdoolittle @kendrapugh @travisstenerson

[mattvarghese]

I would say, the behavior of OR has been what clinicians are used to in our EHR. So I find it rather strange that unlike discussions in the cited issue, we have not really heard such a concern from clinicians.

Additionally, from an adoption by EHR perspective, I would agree with the comment by @kensaku-kawamoto in the cited issue that behavior that exists today in EHRs and is widely familiar to Clinicians should be adopted by the standard.

[mattvarghese]

Created pull request: #498

[isaacvetter]

Hey Guys,

With this proposed change of allowing a selectionBehavior = all, a cds developer must decide when their guidance is an action vs a suggestion and there’s not really guidance on how to make this choice.

Why should selectionBehavior be expanded to include all ? When would a cds developer use selectionBehavior of all, instead of simply placing multiple actions within a single suggestion?

It seems a lot simpler and more straightforward to allow at-most-one and any, but not all.

Isaac