Abstract
The District Health Information systems 2 (DHIS2) is a highly flexible and customizable platform for building health management information systems, for many developing countries. The DHIS2, by being a platform, supports innovation by third-party app developers, thus facilitating the targeting of more specific challenges upon an already robust platform. This thesis will cover the design, development, and testing of a web-based application called “Facility Burden Prototype” (FBP) as an extension onto the DHIS2 platform. The research objective of the thesis is to determine to what extent and how the presentation of a visual relationship between cases of health events or diseases (e.g. reported patients with a specific disease) and the originating health facility, affects decision making, and how it should be applied at the district level in India. The decision problems covered in this thesis are 1) How to establish the existence of a possible disease outbreak? 2) Whether to monitor, investigate or control a possible outbreak? The DHIS2, in combination with the FBP, will be assessed as a case- and syndromic surveillance system, considering strength and weaknesses in the context of the decision problems. The analysis of requirements that goes into the design and development of the FBP have roots in the information cycle, and are split into five branches based a methodology within action research, called Multiview. The main requirements from these branches cover: - Concepts and terminology must match the user level - Pursue transparency in the process of generating information, through the user interface, for reducing any mistrust or concerns. - Data should be organized and limited to “need only data”. - Data should be comparable, thus have a context. - Possibility for disaggregation of data. - Ill-structured spatial decision problems should be supported by flexibility in the GIS. As descriptive epidemiology, which is considered to be a working method of the end-users, concerns an area and not a set of health events derived from a facility, datasets in the FBP must be combined carefully to support outbreak detection in a geographical area. This makes the tracked entity-facility model a complementary model, focusing on an assessment of facility disease burdens, rather than the dominating model for establishing the existence of outbreaks. The FBP arguably has the potential of facilitating a close feedback loop between the district and sub-district/facility level – A feedback loop that could provide district-level decision makers with complementary information that addresses the decision problems of concern. Also, the FBP, by displaying a visual relationship between health events and facilities, backed by a GIS with high spatial resolution, has the potential of targeting field investigations by having a reference to a residential location and enrolled facility. The result is a high level of spatial traceability that may lead to less resource-consuming investigations. The modeling of the relationship between health events and the originating facilities also shows a potential of reducing the cognitive and phycological workload of the end-users, thus the strength of visualizing the relationship.