In this section I seek to describe how the philosophies, methods and processes described in the previous section has been inscribed into the DHIS software. DHIS as an artifact is very important to the spread and practical applicability of the HISP approach. This analysis is based based on the MS Access line of DHIS, but I will try to make it sufficiently abstract to avoid getting into the technical particularities of the different versions of DHIS. That I will come back to in the empirical part of this thesis.
Important to the spread of HISP is vertical scaling, relating to the hierarchy of standard seen in figure 6.1. To make DHIS scale to all levels within a county’s health system an organisational hierarchy are inscribed into the system. Data can be fragmented down to the level of the reporting unit, and data can be aggregated for all levels above the reporting unit. To facilitate local use of information, and at the same time facilitate reporting to the level above, a unit in the hierarchy can have any numbers of data elements and indicators. Reporting is done by exporting data for data elements and indicators required by the level above. This can then be imported into a DHIS database at the level above.
To empower local health worker through local use of information, DHIS offer support for data analysis at all levels. DHIS offer support for calculating indicators and making charts. Raw data, indicators and charts can be shown at all levels of aggregation from the reporting unit and up. A district information officer having received data from all the other districts in the region can compare data from his district to that of the regional total and with other districts. As mentioned data needs to travel both up and down in the hierarchy to make this possible. DHIS also offer support for validating data to avoid mistakes and inaccuracies in reporting. This is needed to make the data more reliable.
To facilitate horizontal scalability, that is transfer of DHIS to new countries, DHIS is designed to be flexible. DHIS is not a ready made system, but needs to be configured to be useful in a new country. The health hierarchy and all the health institutions have to be specified. All the indicators, semi-permanent data (non-routine data used for indicator calculation etc.), data elements, reports etc. have to be specified. DHIS is designed to make building a database tool for a new country easy. To facilitate translation of the user interface into new languages strings are not hard coded into the system, but retrieved from a database containing the strings in different languages. DHIS is basically divided between two parts; a stable core and a flexible exterior. The stable core is the data structures and processes being stable across different health systems. An hierarchical organisation with clinics, hospitals etc., and administrative units are found in all modern health systems. The core parts of DHIS are the health system hierarchy, the general concepts of data elements and indicators and the processes of reporting and data analysis.
The MS Access line of DHIS (DHIS 1.x) is made with tools available in MS Office. It uses MS Access, Visual Basic for Applications (VBA) and MS Excel. The reason for choosing MS Office as the platform for DHIS was based on pragmatic considerations. The development of DHIS was initiated as a part of the first pilot phase of HISP in South Africa in 1996 - 1998. MS Office was already in widespread use in South Africa, with or without a license, so the cost of obtaining MS Office was not considered a problem. MS Office is an expensive office suite, but software piracy is so widespread in many of the countries HISP operates that this cost have not become a serious problem yet. Because MS Access is a RAD tool, it is a good tool for rapid prototyping. The ability of HISP to get a working database tool made within a relative short time was an important factor in making HISP a relative success to the other HIS initiatives in South Africa. DHIS together with pivot tables in MS Excel made DHIS more than a database tool. Pivot tables offers a powerful tool for data analysis, giving user the ability to make charts and diagrams of the data. The ability of MS Access to integrate with other MS Office programs, like Excel, made the use of MS Access seem like a natural choice. The availability of programmers skilled in VBA also came into the consideration (Braa and Blobel 2003).