We report on the successful development of a multi-disease data management system that can be distributed without the user incurring licensing costs, and that has exceptional potential for adaptation by the user to local circumstances, allowing resource poor countries to benefit with minimal expense. The system incorporates user-friendly functionalities for data entry, data storage, data query, mapping and reporting. We consider it likely that use of the system will lead to improved continuous surveillance, intervention monitoring and evaluation of control program performance. However, this needs to be confirmed through operational implementations of the system.
As the Results
section already extensively highlights the strengths of the system, the Discussion
section focuses more on system limitations and potential for improvement.
Operational Testing of the System
Plans are now underway to trial the system operationally with control program partners so that it can be rigorously evaluated. Ideally, the system should be evaluated in parallel with existing management systems for a 1–2 year period in multiple settings for each relevant disease. This will also provide valuable feedback on system performance and lead to improvements in later versions of the system.
System Complexity and Adaptability
The system has exceptional potential for adaptation by the user to local circumstances without the involvement of software developers. The cost of this versatility is increased complexity for the system administrator. Implementation and operational use of the system without a highly competent local system administrator is likely to result in poor system performance, especially with regards to synchronization of data. The complexity is most apparent during the initial system configuration.
The system includes extensive capacity for data import, including import spreadsheets that are tailored to specific functional components such as entry of individual disease cases, entry of data for insecticide resistance bioassays, entry of survey data, entry of data for distribution of insecticide-treated nets, etc. This allows the user to rapidly populate the system with historical data. However, it should be noted that the import process, for data quality purposes, is unforgiving when it comes to poor quality data. Therefore, it often will be necessary to clean historical data that originate from other sources, which may not enforce input of high quality data, before importing the data into the system. Furthermore, the behavior of the import process for a given import spreadsheet is linked to the behavior of the corresponding data entry screen, for example with regards to data fields where an entry is mandatory. Thus, it becomes important that all personnel executing data imports also have a working knowledge of the corresponding manual data entry functionalities in the system.
The data import/export functionality also allows for linkage to existing health information systems by import or export of relevant disease case data. Initial incompatibility issues are expected, especially with regards to names of geo entities, due to inconsistency in spellings or name changes. For geo entities, this can be addressed through an import synonym tool which is included in the system and assists the user in finding names for a given geo entity that closely resembles the one the user is attempting to include in the import. The user then can add the name of the geo entity from the external health information system as a synonym for the same geo entity in the geographical entity tree in our system, which will facilitate subsequent data imports. Finally, the system still lacks capacity for mass-deletion of data, which can become an issue if the data import process is not handled carefully.
The information trees (geographical entity tree, universal tree and term tree) provide tremendous adaptability in the system but require careful consideration when they are configured for a local implementation. There is no question that poorly configured information trees will lead to downstream problems with the operational use of the system. It is important that the team executing the initial configuration of the information trees has domain expertise relating to vector and disease surveillance and control practices in the local environment, as well as a clear idea of what the local user wants to get out of the system in terms of specific outputs to support decision-making and reporting.
The system comes with a default term tree and the universal tree is not onerous to configure. Data for geo entities can be imported into the system by the user to build a locally relevant geographical entity tree. This is, however, restricted to import for a single geographical hierarchical level at a time, which can make the process of building the geographical entity tree time-consuming. Alternatively, the data making up the geographical entity tree can be mass-imported but this currently requires assistance by software developers.
Statistical and Spatial Analysis Capacity
The system was developed primarily to support operational disease control programs and therefore has very limited statistical and spatial analysis capacity. Statistical operations that are directly supported in the system are restricted to 1) query builder calculations of sums, averages, and minimum and maximum values and 2) pre-configured query builder custom calculations that relate to specific system functionalities, such as disease case incidence or mosquito abundance indices. Other statistical operations require the user to export data for subsequent import into a statistical software package.
The system supports, through the use of GeoServer/OpenLayers, basic mapping functions but essentially lacks spatial analysis capacity. The system is capable of producing map overlays to illustrate spatial patterns, for example disease case incidence in relation to percentage coverage by a given control intervention, but lacks capacity for applying spatial statistics to further explore these patterns. To achieve this, the user needs to export a shapefile from the system for subsequent import into a GIS software with spatial analysis capacity.
The most important short-term future directions are pilot implementations of the system, including assessments of the cost for system set-up and operation, and the inclusion in the multi-disease data management system platform of additional important vector-borne diseases such as Chagas disease, human African trypanosomiasis, leishmaniasis, lymphatic filariasis and onchocerciasis.
Additional future plans include 1) making the system directly compatible with hand-held mobile data capturing technologies, such as Personal Digital Assistants and smartphones, 2) developing an over-arching query builder to make it easier to combine data from different parts of the system, 3) developing additional user-configurable functionalities such as configurable indicator surveys or knowledge, attitudes and practices surveys and 4) determining the potential for expanding the system to include infectious diseases with other modes of pathogen transmission than arthropod vectors (e.g., through ingestion of water or direct human-to-human contact).