Spatial analysis and factors associated with low birth weight in Ghana using data from the 2017 Ghana Maternal Health Survey: spatial and multilevel analysis

Objective Low birth weight (LBW) is an important indicator of newborn health and can have long-term implications for a child's development. Spatial exploratory analysis provides a toolkit to gain insight into inequalities in LBW. Few studies in Ghana have explored the spatial distribution of LBW to understand the extent of the problem geographically. This study explores individual and cluster-level distributions of LBW using spatial exploration components for common determinants from nationally representative survey data. Design We used data from the 2017 Ghana Maternal Health Survey and conducted individual-level and cluster-level analyses of LBW with place and zone of residence in both bivariate and multivariate analyses. By incorporating spatial and survey designs methodology, logistic and Poisson regression models were used to model LBW. Setting Ghana. Participants A total of 4127 women aged between 15 and 49 years were included in the individual-level analysis and 864 clusters corresponding to birth weight. Primary and secondary outcome measures Individual and cluster-level distribution for LBW using spatial components for common determinants. Results In the individual-level analysis, place and zone of residence were significantly associated with LBW in the bivariate model but not in a multivariate model. Hotspot analysis indicated the presence of LBW clusters in the middle and northern zones of Ghana. Compared with rural areas, clusters in urban areas had significantly lower LBW (p=0.017). Clusters in the northern zone were significantly associated with higher LBW (p=0.018) compared with the coastal zones. Conclusion Our findings from choropleth hotspot maps suggest LBW clusters in Ghana's northern and middle zones. Disparities between the rural and urban continuum require specific attention to bridge the healthcare system gap for Ghana's northern and middle zones.