Towards Web-based Representation And Processing Of Health Information

Sheng Gao; Darka Mioc; Xiaolun Yi; Francois Anton; Eddie Oldfield; David J Coleman


Int J Health Geogr 

In This Article


The HERXML can be used to share the cartographic representation of health information (able to consider a variety of health activities), and describe health data sources and statistical methods. Our implemented HERXML parser can utilize the representation styles in HERXML documents to generate health maps. The HERXML can be shared by users through the Web in many ways such as email, web sites, web forums, and web services regardless of platform or system (See Figure 11). Thus, health information can be easily represented and shared while keeping the secret of raw health information. If the users are interested in the detailed health information, they can contact the data source manager.

Figure 11.

The sharing of HERXML. HERXML can be shared via Internet by web forums, web sites, web services, emails, etc.

Similar to GML and SVG, HERXML is pure XML (using ASCII file) and in a vector format. The ASCII file format makes it easy for humans to read, search, and edit. Although ASCII file format is much larger than the binary format, a number of XML compression techniques methods, namely gZip, XMill, XGrind, Xpress, and XComp, have been developed to improve the performance of transferring XML over the Web.[23] The vector format enables scalability and resolution independence. A file in raster format usually would be much larger in volume than a file in vector format at the same resolution.[24] The HERXML documents can be interpreted as view-only maps (e.g., JPEG) or interactive maps (e.g., SVG) with the attributes and defined representation style in them. Thus, the cartographical representation will be the same in any platform or system.

SVG is designed for computer graphics, and it lacks point feature representation elements and uses inverted y-axis coordinate system, making it unsuited for Web-based cartography.[25] In addition, SVG uses the graphical coordinate, and this leads to some problems in integrating data from different sources together if they do not have the same coordinates. Meanwhile, GML is able to model, transport, and store spatial information, but it can not provide the cartographical representation of spatial information. HERXML utilizes GML in modeling geospatial features and provide point, line, polygon, and chart styles, making it satisfactory in Web-based cartography. Moreover, HERXML integrates many kinds of attributes information together in a well-formatted structure, with the ability to be represented as text, maps and graphics.

Taking advantage of XML, HERXML is extensible, with the potential to add more health information tags to the representation in defining new health parameters or methods. HERXML is simple and well-structured, and the comprehensive description of health data representation will need more extensions. Meanwhile, to improve the semantic meaning in understanding the health data representation, a well defined ontology should be generated to represent shared vocabularies. The development of XML databases, which facilitate the efficient management of XML, will support the storage and manipulation of HERXML documents.

The WPS standard can support both synchronous and asynchronous requests in the execution process. An asynchronous request is very flexible for users in health data processing, especially when the process is computation expensive. During the processing, the dynamically updated execute response document enables users to know the processing status. The results of WPS could use direct data output or a URL which points to the processing results in the server. For health data processing, it is possible for national health organizations to host some processing functions as well as some basic data (e.g., census data) in their server. If a local health organization wants to use the processing and mapping power, they only have to purchase it and make its data accessible to processing servers, and then they can get the processing results conveniently. To reduce the hardware or software investment costs at every local organization, it would be feasible to build a public health infrastructure to support processing power on the Web. In this way, users can flexibly choose the required processing services and assemble them based on their needs. However, regarding the data used in the processing services, the standard method of accessing health data and related data (e.g., temperature data) for web processing still needed to be explored.