The adoption of ISO/TS 17369:2005, Statistical Data and Metadata Exchange (SDMX - www.sdmx.org) by the United Nations for aggregate data will contribute to efforts to promote computer system interoperability and data use. The SDMX-Health Domain (SDMX-HD) standard is intended to be part of countries' Health Information System (HIS) and Monitoring and Evaluation (M&E) infrastructure strengthening strategy, facilitating seamless flow of information from facility to district, national, and international levels in the absence of well-developed Internet access.
In the health domain, aggregate data is usually obtained and exchanged in the form of indicator values, typically reported as time series classified by various dimensions such as by geography (country or sub-national), by organization, by programme, etc. An indicator itself is defined by its name and various metadata such as its data type, definition, rationale, topic, measurement method, method of estimation, etc. The data structure is further complicated by the notion of disaggregated indicator data (by gender, by age group, by education, etc.) and by the fact that some indicators are calculated by combining results of other indicators, e.g. a percent indicator (prevalence of ) being calculated by numerator (count of ) divided by denominator (population).
Contrary to other domains where SDMX is in use and where all agencies involved are able to reach an agreement on the full data structure prior to exchanging data, the health domain is more evolving in nature. This has important consequences on the SDMX-HD standard, which:
- Allows for the exchange of indicator definitions and metadata without any associated data, for the sake of making publicly available standardized indicator definitions to help harmonize data collection processes;
- Defines a core set of common concepts and code-lists that must be used in the SDMX-HD message to be compatible with the standard, whereas a list of supplemental custom concepts and code-lists is provided by each agency whenever broad agreement on these concepts has not yet been reached;
- Allows the exchange of incomplete matrices of data. Even though an indicator definition may recommend data collection disaggregated by gender (males, females), a particular country may only be able to report the total figure, or data might not be available on a regular basis, thus creating gaps in the time series. SDMX-HD covers this by making use of special missing values to distinguish between an actual indicator value of 0; an indicator value that is reported to be missing (cannot be obtained); or an indicator value that is not reported but may exist (absent from the data set).
blue-infinity helped define the SDMX-HD standard in collaboration with UN organizations and is now actively developing .Net libraries to facilitate the inclusion of this new standard into applications communication stack.
At the heart of the library is an SDMX-HD Document Object Model, allowing users to import and export data as well as validating content without worrying about the details of the XML file format which involves a quite sophisticated understanding of XML and XSD technologies. For example, the tool supports dynamic generation of the XSD schema to validate the data being transmitted against the Data Structure Definition file. The tool has been integrated into Visual Studio 2008 making debugging a very easy task.
A Google Group has also been created to promote the development of a community around the SDMX-HD to promote its use. The site can be found at http://groups.google.com/group/sdmx_hd.