The limited availability and low-quality of national-level data to understand stillbirth determinants in low- and middle-income countries is a major barrier to stillbirth prevention efforts globally. Data for producing stillbirth estimates in these settings generally come from national household surveys due to the high proportion of women that give birth at home, and because of inadequate or absent civil and vital registration systems. There are several quality concerns with stillbirth data from household surveys, particularly around under-reporting and misclassification, which affect the reliability of estimates. What also remains underexplored is to what extent social and cultural factors affect the disclosure and reporting of stillbirth, as these have important implications for stillbirth data quality. This thesis sought to review the availability of country-level data on stillbirth from nationally-representative household surveys in low- and middle-income countries to outline what data exists, where the limitations are, and potential for improvement. Analysis of data from a national mortality survey in Afghanistan is used to describe the key determinants of stillbirth in this setting and demonstrate how better data could be generated to understand stillbirth risk factors globally. Qualitative methods are used to further explore how community and healthcare provider experiences, perceptions and practices might impact stillbirth data quality and contribute to pathways leading to stillbirth in Afghanistan. The findings of this thesis can inform future efforts to improve stillbirth data in low- and middle-income countries and identifies where interventions can begin to focus to facilitate stillbirth reduction in Afghanistan.