Analytes and same-instance (bioref interface)

[Fasjul]

It is unclear how analytes take part in the cohort selection filter. In particular, it is unclear how they should behave (and actually behave) with the "same instance" panel timing.

Wanting to plot results of measurement A that were gathered using device D, a user would naively drop device D in the selection filter and measurement A in the analytes zone.
While the query would include only patients who at some point underwent a test involving device D, measurements for these patients of test A performed with another device than D could be included in the resulting plot.

This is due to the distinction between patient selection done in the explore filter and record selection forming the statistics query ; a solution would be to push to the latter the selection criteria involvng the same concept as the analyte. (in the example above, it would mean filtering the records using the device criterion)