Health systems strengthening for maternal and child health: Real-life effects in rural Guinea-Bissau

Introduction: Guinea-Bissau has among the highest burden of maternal and child mortality globally. In response, the nation-wide health system strengthening initiative ‘Integrated Programme for the Reduction of Maternal and Child Mortality’ (PIMI) has been implemented. PIMI’s theory of change is to improve maternal and child survival through increased coverage of essential quality maternal and child health (MCH) services. The aim of this thesis was to assess PIMI’s impact on MCH service coverage and perinatal health and to explore facilitators and persisting barriers to timely and quality care during labour, birth, and immediately postpartum (henceforth: peripartum care) during PIMI in rural Guinea-Bissau.

Methods: We applied an explanatory sequential mixed-methods approach with a systems thinking lens, based on three studies. The studies were nested in the Bandim Health Project’s nationally representative rural health and demographic surveillance system (HDSS) which monitors pregnancies, uptake of MCH services, and maternal and child survival in an open cohort of >50,000 women and children. To assess PIMI’s impact on MCH service coverage and perinatal health (impact evaluation), we defined three two-year cohorts comprising 23,828 HDSS-registered births: (i) pre-PIMI, (ii) during PIMI’s pilot phase where PIMI was implemented in 4/10 rural health regions, and (iii) during PIMI’s full-scale phase where PIMI was implemented in all health regions. In line with PIMI’s stepwise roll-out, we differentiated early vs. late implementation areas and assessed coverage of antenatal care (ANC) and facility births as well as perinatal mortality (i.e., stillbirths and early neonatal deaths <7 days of age) by area over time. To explore women’s perspectives on facilitators and persisting barriers to timely and quality peripartum care (demand-side assessment), we conducted 258 structured interviews and 12 in-depth interviews with HDSS-registered women in 19 randomly selected HDSS villages. To explore service providers’ perspectives, we conducted 8 in-depth interviews and participant observations (192 hours) at the target health facilities of the previously in-depth interviewed women (supply-side assessment). Quantitative data were analysed using generalised estimating equations and descriptive statistics; qualitative data using thematic network analysis aided by theories of social practice.

Results: The impact evaluation revealed that coverage of ≥4 ANC consultations (ANC4) and facility births increased from approximately 1/3 pre-PIMI to 1/2 during PIMI’s full-scale implementation in both areas. Perinatal mortality remained stable at approximately 80 deaths per 1,000 births in both areas over time. While higher village-level coverage of ANC4 (both areas) and facility births (early implementation area) was associated with a tendency towards lower perinatal mortality pre-PIMI, we found no related associations during PIMI’s full-scale implementation. In the demand-side assessment, women commonly expressed a pronounced preference for facility births over home births. Accordingly, 42% of the women who participated in the structured-interviews and had given birth at home or en route to the health facility reported having planned a facility birth (54/130). Yet, at the same time, the women interviewed in-depth reported ubiquitous geographical and financial barriers to care including a multitude of direct and indirect out-of-pocket payments despite PIMI’s official user-fee waivers. Similarly, the supply-side assessment revealed that providers regarded health facilities as the only reasonable place to give birth. Yet, timely and quality peripartum care provision was severely compromised by geographical, material and human-resource constraints. Providers stressed that essential medicines, consumables, appropriate equipment, and staff were key bottlenecks, and explained lacking material supplies by discontinued donor support. To navigate these circumstances, providers applied several strategies including asking women to purchase lacking materials, omitting diagnostics, and involving birth companions and support staff in care provision. These strategies further compromised financial service accessibility, patient and occupational safety, while causing delays to care, and diffusing health worker responsibilities. Meanwhile, both the demand- and supply-side assessment indicated that amid persisting barriers to care, women conditioned care seeking on perceived individual risks of birth complications.

Conclusions: While we found increases in the coverage of essential MCH services during PIMI, coverage remained suboptimal, and developments appeared secular rather than associated with PIMI’s implementation. At the same time, perinatal mortality stagnated on a high level. This is despite service providers and women valuing facility-based service provision and regarding health facilities as the ideal place of birth. We identified pervasive geographical and financial access constraints which likely explain suboptimal coverage developments while causing delays in obtaining emergency care. At the same time, delays to care, along with severely compromised quality of care due to material and human-resource constraints, likely explain persisting high perinatal mortality. Our findings raise severe equity concerns and call for a rigorous monitoring of health system strengthening initiatives in parallel to their implementation to ensure the early detection and mitigation of unintended developments.