No code changes in OC
All configuration should be via UI. No backend or direct SQLs etc.
Integration should happen in real-time as much possible. CRF sync'ing can be nightly if there are technical challenges.
For every object added from OC to OS (patient, site, visit, crf, etc) there should be an entry in an audit log table (this can be a central table e.g. object_type, OC_ID, OS_ID, created_date, last_updated_date)
OC or OS database should not be accessed directly. All data access should be via APIs.
An error log table should be present to log all errors.
Error should be emailed to the system admin too.
Error should be accompanied with meaningful errors.
It should be possible to link an OC study to OS collection protocol. This can be done either by displaying a dropdown of OC studies (via an OC API to get all studies) OR just giving a text box to enter the OC study code/title.
Configuration should include OC URL/instance. This will enable different studies to have different OC instances.
Same for Site.
This is one way link from OC to OS.
When a patient is registered in OC, it should get automatically registered in OS for the linked study.
The “study ID” field in OC should be duplicated as PPID field in OS.
It should be possible to map fields to OC to either static domain object fields or extension fields of OS
Ditto to Visits.
When a study is linked, its CRFs in OC should be replicated in OS under that study.
In OC these will be visit level CRFs.
If a new version of CRF is updated in OC, it should be synced in OS
Each time a CRF is CRUD in OC, it should be synced in OS