Client Registry Implementation Guide - Local Development build (v0.1.0) built by the FHIR (HL7® FHIR® Standard) Build Tools. See the Directory of published versions

Overview

The primary purpose of the LK Client Registry (CR) is to manage patient identities. Managing identities is composed of the following major responsibilities:

1. Providing an accessible, authoritative and performant data repository for client demographic information
2. Identifying duplicate records using deterministic or probabilistic matching algorithms
3. Prevent duplicates prior to creating new records and resolve known duplicates
4. Maintaining linkage of records across source systems to indicate these records represent a single individual
5. Possibly creating a 'golden' record representing the most up to date, accurate demographic information about an individual amalgamating data across disparate systems

Not all of these responsibilities will be implemented; the above is a summary. The details of how the CR is behaving is documented in this guide.

Interface Conformance

Connected systems SHALL use the base FHIR Restful specification to access Patient instances, limited to patterns as described in this guide. Specific FHIR Operations, as defined in this guide, SHALL be used to interact with the CR. The limited interfaces offered by the CR are described in the detailed specification page and in the capability statement.

Duplicate Records and Linkage

A source system is an organization, like a hospital or clinic, that have a medical record computer system. A source system should send client demographic updates to the CR and also send new Clients to the CR. Duplicate client records represent the same individual from a single source system. Resolving duplicate records including merging will not be part of the initial CR requirements.

Any Client record can be linked to other Client records if the CR determines, through deterministic and automatic matching, they are the same individual. The primary focus for Sri Lanka is an automated, straightforward linking algorithm that can run unattended (no review necessary). This may be links within a source system or between source systems. There is also an operation defined, LinkClients, that allows an user to link Patients from a point of service.