@ssmusoke that sounds like a pretty complicated process, that leaves quite a bit of room for error.
I would like to see a set of scripts or utilities that take CSV input and can upload/import them to OpenMRS. The CSV would be something you can repeat. (Ideally you can lead an effort to create this as a shared library. Or even better would be if the link that Pascal shared is already such a library.)
@darius CSV is the simplest and easiest, why not a FHIR based output since it can easily be imported into OpenMRS - from a design call I remember Bhamni already has some implementation, and in the future can be used to share and move patient data between installations?
CSV is easier to produce for an average-skilled person trying to get data
from another system. FHIR is quite a bit more complicated, and also it’s
possible to express the same thing in multiple ways, hence more room for
error.
It depends how sophisticated the system you are migrating data from is. If
it’s an EHR then maybe FHIR if a better way to express its data. If it’s a
simple registry, FHIR is probably overkill.
Maybe an even better approach is to combine:
Document a few FHIR templates that can be used to reliably load OpenMRS
with patients, visits, encounters. If it’s feasible for the person doing
the work, they can directly produce this format.
Define a simpler CSV format, and a script to transform from this CSV to
your FHIR format. When directly producing FHIR is not feasible, go via this
CSV format.
@darius I would like to explore this more at the conference in December (hope that you are coming) would you be interested in having a brainstorming unconference session on this?