Thanks Simon!
I’m happy to share the details of our current setup. And I’m also curious to hear your opinion and suggestions on how to improve the process.
Partners In Health in Malawi has an implementation of OpenMRS catching patients District-wide across all facilities (~15). Besides this patient-level data now screening activities happen both at community level (e.g. in a village, or through dedicated events) and at the Outpatient department of every facility. During these screening events (potential) patients can be referred to their nearest health center / hospital providing the required services.
Now unlike the typical (?) use case of CommCare where a frontline health worker has his/her own dedicated phone, we will have tablets that are travelling around the district to the various screening activities. These ‘registration’ devices will be use to open the CommCare case. And then there will be matching tablets at the health centers/hospitals to keep track of who came in when (aka closing the CommCare case).
Now the majority of the referrals will remain within their 'catchment area’. So most people screened at one screening activity will be referred always to the same nearest facility. But there will be cases where either the patient actually lives closer to another facility or where a patient requires special services not available at the nearest facility. And these cases leave the 'default boundary’ and the CommCare cases need to be available not only at the nearest, but in theory at every other facility.
With this the devices are not bound to a specific user, but more to a use case (referral registration or tracking-who-visited-through-a-referral) and to some degree the physical location. Hence I have a hard time dividing the CommCare cases into disjunct user groups.
I don’t have exact numbers, but I would guess that 100 referrals for one screening activity are the upper boundary. But multiple of these activities can happen throughout the district (but I doubt that from the beginning they all will get close to this upper boundary). But the referred (potential) patients will only show up a week or 2 after their referral. Additionally everyone who was referred will actually show up (and then require a potentially lengthy tracking process), so I expect that over time we will have a growing number of open cases.
This is not implemented yet, but I could see that for patients where the tracking-process was started, the CommCare case might also be closed. So that these ones, which might potentially never be resolved due to lost-to-followup, are actually no longer part of the open cases in CommCare.
Hope this makes sense and happy to learn more about other ways to implement it.
christian