Preferences of Women for Maternal Health Services in Sidama Region, Ethiopia: Discrete Choice Experiment

Skilled maternal health services, including antenatal care (ANC) services, childbirth services and postnatal care (PNC) services, can save the lives of women and newborns. However, women in resource-limited countries, such as Ethiopia, tend to have a relatively lower uptake, partly because current services may not align with their preferences. This study assessed the stated preferences ofwomen for maternal health services in the Sidama region, Ethiopia.

A multi-stage sampling technique was used to select reproductive-aged women (15-49 years) from the Central Sidama Zone, Sidama region, Ethiopia. A D-efficient discrete choice experiment (DCE) with two blocks of 24 choice tasks was designed. Six attributes, including healthcare provider attitude, availability of medication and supplies, distance to the health facility, privacy, waiting times, and service cost, were identified through literature reviews, focus group discussions and in-depth interviews with women and other stakeholders. Respondents were randomly assigned to one of the three interrelated but distinct maternal health services (ANC, childbirth and PNC). A latent class analysis (LCA) model was used to estimate preferences and estimate the relative importance of attributes, uptake probability, and willingness-to-pay (WTP). The association between individual sociodemographic characteristics and latent classes was examined using a class membership model.

A total of 1558 women, including 526 for ANC services, 537 for childbirth services and 495 for PNC services, were included in the analysis. For each service, three classes were identified, and notable differences in preferences were observed across the classes within each maternal health service. For ANC services, Class 1 was a medication-sensitive group (34.3%) that prioritised the availability of essential medications. In contrast, Class 2 was a price-sensitive group (21.9%), exhibiting high disutility for costly services (7500 ETB ≈ US$137). For childbirth services, Class 1 was a time-sensitive group (21.1%) exhibiting strong aversion to long waiting times (24 hours), while Class 2 was a price-sensitive group (19.3%). For PNC services, Class 1 was a medication-sensitive group (48%), whereas Class 3 was a price-sensitive group (25%). The predicted uptake probability in ideal conditions was above 96% across all maternal health services; however, it ranged from 60% for class 2 PNC services to 99% for class 2 childbirth services in average conditions. Overall, women showed a higher willingness to pay (WTP) approximately 8604 ETB (US$159) and 5118 ETB (US$95 ), to avoid childbirth services that lacked medications and supplies or had long waiting times, respectively. Sociodemographic factors, including age, residence, and education level, significantly shaped women's preferences for maternal health services.

The cost of service, availability of medication and supplies, and waiting time were the most important attributes influencing maternal health service preferences for most of the classes of all services, while privacy and distance to the health facility were frequently the least important. Maternal health services should address the specific needs of women, such as subsidising costs, ensuring sustainable medication supply, and optimising service efficiency to enhance maternal healthcare uptake and outcomes in resource-limited settings, including Ethiopia and similar contexts. The findings will provide policymakers with valuable insights and inform future research on maternal health services in other contexts, including the Afar and Somali regions, where nomadic lifestyles are common, and data are sparse.