Duplicate Detection

[sleidig]

Detecting (possible) duplicates is an important aspect for reliable data and consistent reporting stats. Double registration of the same person / item can happen when

• someone doesn't notice that the same person has been registered before (by themselves or another user)
• someone doesn't have access permission to the existing record and therefore no chance of knowing it is already registered

The system should help users

• see a warning while recording a new entity if the data is similar to an existing record (and also see that warning when viewing a specific entities' details, not only during creation)
• list all (possible) duplicates and review them
  • merge two duplicates into a single record
  • mark two suspected duplicates as distinct to hide the warning

Approaches and related concepts

• see https://guides.ohie.org/arch-spec/openhie-component-specifications-1/client-registry
• see GovStack

Implementation steps

--> Implementation Plan

• Algorithm to detect (possibly) duplicate records
  • (incl. research of approaches, testing & analysis on real data, ...)
• Show duplicate warning while creating a new record
• List possible duplicates to review for admin
• Merge two notes/events into a single record #1414
• Discard duplicate warning / mark as distinct