Common Metadata Library¶

Introdução¶

This document describes a metadata package that provides a baseline of core DHIS2 metadata for use across multiple programs or datasets. The metadata included are not programme/use case specific, but are generally expected to be shared across tracker programs as part of a harmonized HMIS.

One key advantage of DHIS2 is that it can be used for multiple health programs, within a single, harmonized instance. There are several practical reasons to share metadata across tracker programs:

• To reduce duplication of data entry and improve data flows, e.g. where a person’s details recorded in one program can be carried over to another program;
• To improve analytic capabilities across programs by re-using common metadata and its associated codes for core attributes like Sex;
• To facilitate maintenance of the metadata in the system

DHIS2 metadata packages are designed and harmonized in a way which allows multiple packages to be imported into the same DHIS2 instance. This might be done to link an individual’s enrollments in different programs, or create synergies across reporting flows for different health areas. For example, a patient may be enrolled in both HIV and TB case surveillance programs, but the patient’s Sex is recorded only once at the time of first enrollment. To do this, the programs need to share the same metadata for the tracked entity attribute (Sex) plus its associated option set and options (Male, Female, and Unknown).

Importing those common metadata together as a baseline package ensures that the Unique IDs, names, and codes for objects (e.g. tracked entity attributes, option sets, options, and data elements). In addition, if a custom program is developed within the same instance, it could re-use those metadata from the baseline package, thereby bootstrapping configuration of the new program, preventing namespace clashes and other complications during metadata import/export, and facilitating data exchange across platforms. This becomes particularly important when aligning tracker implementations that may exist across multiple instances into a shared, harmonized environment.

System Design Overview¶

Common tracker metadata are made available as installable metadata files that can be imported into a new or existing DHIS2 instance. These installable files contain metadata that are aligned to a few common use cases for DHIS2 Tracker:

Core Case Profile¶

The core case profile is a small collection of metadata objects in the tracker domain that represent identifiers and common data variables typically used across case-based registries. These metadata are re-used across all WHO case-based surveillance tracker packages (e.g. HIV, Malaria, TB, VPD, and COVID-19 case-based surveillance). The Core Case Profile was developed through an internal WHO harmonization process for identifying and harmonizing naming conventions for shared data variables across person-centered monitoring and case-based surveillance programs. There are many practical benefits of harmonizing these metadata in a DHIS2 instance, such as enabling the registration of a TEI 'person' in one program and making their personal details available for searching to enroll in another program (such as enrolling an HIV case into the TB program).

The core case profile includes the following objects:

- Tracked Entity Type (Person)

- Tracked Entity Attributes (e.g. given name, family name, date of birth)

- Options (e.g. male, female)

- Option Sets (e.g. sex)

Common Tracker Metadata¶

This library contains all objects from the core case profile above, plus other common metadata in the tracker domain such as data elements, option sets, and options. This library is not exhaustive, but contains metadata concepts that are included in official DHIS2 tracker packages, and are generic enough that they could be conceivably shared across programs (e.g. option sets for a list of countries or a data element for ‘date of death’). The library is intended to provide a selection of common metadata that implementers can take at their discretion in the development of tracker programs. This library will be more frequently updated and expanded on a semi-annual basis as metadata packages are developed and deployed.

All metadata in the Common Tracker Metadata Library comprise metadata groups containing the prefix GEN (for General). The prefix helps to signify to the implementer or system admin that the metadata object is being shared across programs, and impact across programs should be considered before making changes/adjustments to shared metadata.

• Tracked entity type (Person)
• Tracked entity attributes
• Option sets, including ISO 3166-1 alpha-2 country codes and country names
• Options
• Elementos de dados
• Custom attributes associated with the above (e.g. LOINC, Snomed, ICD-11)
• Attribute for aggregate data exchange

The diagram above illustrates a few points about the modular structure of the package. The Common Tracker Metadata library includes the entirety of the Core Case Profile, in addition to other metadata objects. A DHIS2 Metadata Package may utilize metadata found within the Core Case Profile, within the broader Common Tracker Library, or its own program-specific metadata (see HIV CS Package). Meanwhile, a custom, locally-designed tracker program may also utilize metadata found within the Core Case Profile, within the broader Common Tracker Library, or develop its own program-specific metadata.

Imagine a scenario where the NTD Surveillance program above is developed independently. The implementers then decide to import the HIV CS package into the same DHIS2 instance. There would be a number of conflicts during metadata import, because objects in each program have the same name and different UIDs (phone number, national ID, and country of birth). If the Core Case Profile had been imported prior to the configuration of a custom NTD program, there would be less work involved to configure new metadata, and fewer clashes when the HIV CS package is imported.

Note that the GEN prefix is used internally within the package environment to signify that the metadata is shared across programs; however, the GEN prefix will not be visible to end users of your programs. These metadata come translated for different locales, and can be further translated to other local languages without interfering with the baseline UIDs, names, and codes.

Common Aggregate Metadata¶

The purpose of these metadata are to store the most commonly used population denominators.

This file includes 6 different indicator types

• Apenas numerador (número)

• Por 10.000

• Por 100.000

• Por 1000

• Percentagem

• Taxa (factor=1)

Total population and populations under 1 year, under 2 years, and under 5 years are included, as are the number of live births, and total number of expected pregnant women. Indicators included to process and display population data are "numerator only".

A variety of category options for common age and sex disaggregations of population are included. Before being used as dimensions of population data, these categories would have to be manually added to categories, and then those categories added to category combinations and fianlly to data elements.

Complete Common Metadata Library¶

This file includes ALL metadata from "Common Tracker Metadata" and "Common Aggregate Metadata" files.