search

IFRCGo / cbs

Red Cross: Community-Based Surveillance
https://cbsrc.org/
Other
101 stars 110 forks source link

Have another look at identifying case reports #962

Open woksin opened 5 years ago

woksin commented 5 years ago

When a phone number is added to a data collector the policy kind of re-reports the case report where the data collector is unknown but the origin matches the added phone number. This should not be as it communicates to the CBS system that a new case report has been received, this can be hard to work around in the other bounded contexts (Alerts, Analytics).

A better idea would be to have the information necessary to update the entry of the case report in question on the CaseReportIdentified event

woksin commented 5 years ago

This could have vital implications for Alerting and potentially other bounded contexts, am I correct @Velcrow81 ?

Velcrow81 commented 5 years ago

I think that it is important that case reports are not made available to other bounded contexts before they are complete. If we do not know who sent the message or if the message makes no sense, it is worthless for alerts. If however, the data collector was identified the next day, we might have enough information and it could still be valuable (timestamp will show that it is one day to old). If it was to old, say six months old, it could still be worth something to analytics and they could run scenarios like what if we were able to get new phone numbers quicker, would we save more life. If we let case reports that are incomplete through with the same event as proper reports, all other parts would have to deal with these kinds of messages. Holding them back is the only safe guard we have against random people sending garbage into the system. I think it is right that we re-report since that is the frist possible time that we have enough information to consider that message as a case report. What do you think about this?