Sormas-project: Observation on duplicate person detection

[Iheanacho2027]

During import, once a user selects an about to be imported case as a new person, the system doesnt allow the user to make selection on the case because of the difference in a variable such as age between an existing person and an about to be imported person. This will result in a duplicate because a different selected person might be an already existing case on the platform. The previous method whereby users can make their selection by comparing name, age, region, district, case classification and Date of report is better than this new algorithm.
If the aim of this new algorithm is to track reinfection, we can use a variable that asks about reinfection, and also links the reinfection to an existing person. In future, for developed nations where national IDs and health IDs are compulsory for citizens, this can be used as a tracker @bernardsilenou

[bernardsilenou]

@Iheanacho2027 Some few comments:

• When system detect duplicate person and do not suggest duplicate case, it means there is none. This is the goal of duplocate person detection, having a duplicate person does not mean there is a duplicate case or contact
• The goal for duplicate person detection is not only to track reinfection but basically not do have duplicate persons in the system
• Variable like national ids are used only if available, if not then what ti available is used, so no problem here

• What i will like you to test with 1,45 is the to record the time it takes to import n cases. how many duplicate persons or cases found during import.
  Hope this helps

[Iheanacho2027]

@bernardsilenou when creating a new case, it takes about 3 minutes to spot a person duplicate because the system has to check the whole person database which is choked up on the nigerian server

[bernardsilenou]

there is another issue to address the speed, I hope when it is implemented the it will be faster

[Iheanacho2027]

@bernardsilenou i believe the higher the data on the database, the slower the performance. Putting in more process checks further reduces the speed with regards to the heavy data on the database. In Nigeria, we are having so much issues as regards this newly introduced duplicate person checks during case entry. It takes so much time to check for person duplicates as it's frustrating

[bernardsilenou]

@Iheanacho2027 I absolutely understand you, as you can see the team will work on it in this sprint.

@MartinWahnschaffeSymeda @markusmann-vg this issue has many relatives, please read Matés comment on issue #2539. We need to find a way to all chackes in the db

[bernardsilenou]

Same as #2539

[Chinedar]

@bernardsilenou when are we able to resolve this? Users are unable to create cases easily. The system keeps trying to search for duplicates and this takes a long time. We have cases coming in 1000s on a daily basis. The implication of this is that we are unable to pride on real-time as the system is designed to run. Urgent resolution is required please.

[bernardsilenou]

@Chinedar New version on 26 will solve this issue. We will know the next step after the release.

[ChristopherRiedel]

@bernardsilenou @Chinedar as you can see is this issue assigned to version 1.48.0 (milestone) and will likely be implement in the next sprint.