Abstract
Devolution, which is becoming common in developing countries, refers to the creation of subnational governments that are substantially independent of the national level. Devolved entities have a clear legal status, recognized geographical boundaries, elect their leaders and raise their revenue. Potentially, the process of devolution generates or modifies existing institutions. Arguably, the new institutions, the incentives they create, and the behaviour of agents in the face of these incentives affords or constrains the implementation of Health Information Systems (HISs). Identifying these incentives and proposing remedies to improve the HISs is the theoretical motivation of this study. The empirical foundation for this thesis comes from studying the implementation of the District Health Information Software (DHIS2) in Kenya. During its implementation, the country adopted a new constitution, which introduced the devolution of decision making to the lowest levels of the government systems. The study attempts to answer two research questions.
RQl: What institutional arrangements arise from the interactions between devolved health systems and the health information systems implementation in the context of advanced information technology?
RQ2: How do the institutional arrangements generated by the devolution of health systems create institutional incentives that shape the effectiveness of health information systems in terms of data quality and information use?
Using a longitudinal case study design, I collected data for this thesis from extensive desk reviews, participant observation notes from the implementation of the DHIS2 and key informant interviews. To understand the implications of devolution of health systems to the implementation of HIS, I drew upon concepts from institutional theory, particularly those relating to institutional incentives. Taking an interpretivist epistemological position, I used both qualitative and quantitative analytic approaches to understand the role of institutional incentives in shaping the effectiveness of the health information system in terms of data quality and information use in devolved health systems of developing countries.
The key findings from this study show that the use of advanced Information Technology (IT), specifically central server and cloud computing enhances data access, information sharing across all levels of government and facilitates the integration of fragmented HISs. By enabling data sharing between national and devolved governments, technology becomes a centralizing factor in a decentralized environment. Concerning the role of institutional incentives, I argue that simple and practical incentives improve accuracy, timeliness and completeness of data in information systems. For instance, in the free maternity project in Kenya, the national government reimburses hospitals for deliveries using HIS data. The number of hospital deliveries are therefore important to all stakeholders, explaining the observed high accuracy of data transfer from maternity registers to DHIS2. The quest for re-election of the local leaders is a major institutional incentive driving data demand and use for planning. This thesis contributes to the institutional theory, specifically the concept of institutional incentives by claiming that successful devolution of power and authority from national level to sub-national levels creates institutional arrangements, which generate institutional incentives that facilitate multiple individuals to utilize their time, skills and knowledge to jointly, create valued HIS outcomes such as high data quality for informed decisions. However, while there are many incentives at play in a devolved health system, it is also important to step back and examine how some of these incentives might impede rather than enhance HIS outcomes. For instance, elected local leaders receive more recognition from the electorate when they implement visible projects like constructing roads, buying ambulances and building new hospitals than by printing reporting tools for HIS.