I am in the process to integrate microbiology too, and I haven´t found any news/choices about this question. From several reading below I think @TBanokina proposal is a good move and I took it as a basis with some slight modifications:
- added specimen_id to measurement table as a clear foreign key to specimen
- used Modifier_of_event_id / Modifier_of_field_concept_id to store the relation between records (susceptibility derive from of culture)
- I am thinking about a microbiology_era table which could synthesize the information and ease the analysis (related to https://github.com/OHDSI/CommonDataModel/issues/281)
This has the benefit to not use the fact relationship table. Still the self joins on measurement to link the culture is a performance problem as mentioned by @cukarthik. I also wonder if a dedicated culture table is not a better choice.