Child health services utilization in Nepal: Modeling physical accessibility with geographic information systems and spatial network analysis

In order to improve child survival, it is important to understand the level of health care accessibility and the determinants of healthcare use so that appropriate policies can be developed to maximize health services access and utilization. This study undertakes techniques of spatial network analysis to develop a geographically-derived measure of health services accessibility based on a combination of maps, administrative records, and the 1996 and 2001 Nepal Demographic and Health Surveys (NDHS). We evaluate the effect of this measure at three levels of aggregation: individual, household, and cluster, in a multilevel logit model of health services utilization for child vaccination (DTP3), and children suffering from diarrhea and/or respiratory illness in Nepal. The results reveal: (1) important travel time effects; (2) important impedance effects of elevation and poor road composition; and (3) the relevance of health facility type. The analysis also highlights the usefulness of NDHS 1996 and 2001 tabular and spatial data to conduct health services research, and in developing empirically-based and spatially-relevant policies and programs to address child health inequalities and improve access to underutilized child health services.