PhD defence Catherine Birabwa

Supervisors

• Prof. dr. Lenka Benova (ITM)

• Prof. dr. Josefien van Olmen (ITM/University of Antwerp)

• Prof. dr. Aduragbemi  Banke - Thomas (London School of Economics and Political Science, University of London, UK) 

• Prof. dr. Peter Waiswa (Makerere University School of Public Health, Uganda) 

Abstract

Introduction: Well-functioning health systems that ensure timely access to quality emergency obstetric care are essential for addressing the maternal mortality burden in Africa. However, evidence on their performance in cities is limited. Identifying and understanding performance challenges are essential steps in developing effective solutions for strengthening urban health systems. This PhD investigated how Kampala’s health system manages maternal health, focusing on routine facility data quality, institutional maternal mortality and trends, care-seeking pathways for emergency obstetric care, and obstetric referral dynamics.

Methods: The research followed a sequential mixed-methods approach. First, the completeness and consistency of data on deliveries, live births, and maternal deaths for 2016-2021 were assessed. Second, institutional maternal mortality levels, based on facility data, were analysed for trends over time (2016-2021) and across facility types. This was complemented by a records review of 358 maternal death cases. Third, a survey of 433 women treated for obstetric complications was conducted to examine common care-seeking pathways to reaching emergency obstetric care. Fourth, semi-structured interviews were conducted with 45 skilled health personnel to better understand obstetric referral processes, using a behavioural approach. 

Results: The findings showed that designated emergency obstetric care facilities reported relatively good quality data, although maternal death reporting was inconsistent. Kampala showed a high and fluctuating institutional maternal mortality ratio, ranging from 109 to 258 deaths per 100,000 deliveries, reflecting overall poor health system performance. Mortality levels were influenced by poor reporting of maternal deaths and referrals from districts surrounding Kampala. The care-seeking pathways for most participants mostly involved one or two steps, and reflected possible referral inefficiencies, including referrals from designated comprehensive emergency obstetric care facilities and women returning home after referral before proceeding to the recommended facility. The two most common pathways identified significantly depended on the district of residence, who decided the mode of transportation to the facility and place of delivery, as well as the need for permission to go to the facility for emergency obstetric care and to spend money on pregnancy or childbirth care. Over 50% of cases of maternal deaths and complications managed in Kampala involved women from neighbouring districts. Qualitative insights from skilled health personnel, underscored resource shortages and process challenges as major factors influencing referral decisions and practices. These resulted in avoidable referrals, hindered critical practices such as pre-referral communication, and demotivated participants. 

Conclusions: The findings suggest that the referral system significantly influences the ability of Kampala’s health system to improve maternal health. The system was characterised by multiple interrelated inefficiencies, where underperformance in one element drives underperformance in other elements. However, there were some opportunities that could be leveraged to strengthen system performance. Overall, multidimensional strategies are necessary to improve system performance, targeting emergency obstetric care readiness, referral system efficiency, provider motivation, social and financial barriers, and governance-related constraints. Future research should explore the governance of Kampala’s health system, specifically examining resource planning, inter-district coordination, and referral system management.