Contents lists available at ScienceDirect Health & Place journal homepage: www.elsevier.com/locate/healthplace Spatial variations and determinants of infant and under-five mortality in Bangladesh Oliver Gruebner a,b, ⁎ , MMH Khan c,d , Katrin Burkart e , Sven Lautenbach f , Tobia Lakes a , Alexander Krämer d , S.V. Subramanian g , Sandro Galea h a Humboldt-Universität zu Berlin, Geography Department, Berlin, Germany b University of Zurich, Epidemiology, Biostatistics, and Prevention Institute, Zurich, Switzerland c King Faisal University, College of Applied Medical Sciences, Department of Public Health, Saudi Arabia d University of Bielefeld, School of Public Health, Department of Public Health Medicine, Bielefeld, Germany e Columbia University, Mailman School of Public Health, Department of Environmental Health Sciences, New York City, NY, USA f University of Bonn, Institute of Geodesy and Geoinformation (IGG), Bonn, Germany g Harvard T.H. Chan School of Public Health, Department of Social and Behavioral Sciences, Boston, MA, USA h Boston University, School of Public Health, Boston, MA, USA ARTICLE INFO Keywords: Spatial autocorrelation Socio-ecological factors Malaria Urbanicity Child mortality ABSTRACT Reducing child mortality is a Sustainable Development Goal yet to be achieved by many low-income countries. We applied a subnational and spatial approach based on publicly available datasets and identified permanent insolvency, urbanicity, and malaria endemicity as factors associated with child mortality. We further detected spatial clusters in the east of Bangladesh and noted Sylhet and Jamalpur as those districts that need immediate attention to reduce child mortality. Our approach is transferable to other regions in comparable settings worldwide and may guide future studies to identify subnational regions in need for public health attention. Our study adds to our understanding where we may intervene to more effectively improve health, particularly among disadvantaged populations. 1. Introduction Child mortality is substantially higher in low- and middle-income countries as compared to high-income countries (United Nations Children's Fund (UNICEF), 2014a). The Sustainable Development Goals (SDG) have committed nearly all governments in the world to reduce under-five mortality, that is, the probability of dying before age five, to at least as low as 25 per 1000 live births by 2030 (United Nations General Assembly, 2015). Yet there remains much to be done to achieve this goal as mortality rates vary largely between populations of lower and higher socio-economic status (SES) (Hajizadeh et al., 2014). Those with higher SES typically experience less mortality risk due to, for example, better education and thereby better health knowledge, nutrition or health care seeking behavior, as opposed to lower SES groups (Black et al., 2003; Hossain, 2015; Khatun et al., 2012; Najnin et al., 2011; Schell et al., 2007). Research has also shown that mortality is lower in urban areas where there is usually better access to health and social care as compared to rural areas (Van de Poel et al., 2007). However, urban agglomerations go along with a large number of adverse factors, for example, higher levels of pollution, violence, that affect mostly the urban poor (Rocha et al., 2015). Over the last 25 years, Bangladesh achieved a great deal of reduction in child mortality with under-five mortality declining from 144 per 1000 live births in 1990 to 41 in 2013 (United Nations Children's Fund (UNICEF), 2014b). Similarly, infant mortality, that is the probability of dying before the age of one, declined from 100 per 1000 live births in 1990 to 33 in 2013 (United Nations Children's Fund (UNICEF), 2014b). Yet, the leading causes of death (except unspecified group) in children under 5 years in 2013 were prematurity, birth asphyxia, acute respiratory infections, neonatal sepsis, congenital anomalies, injuries, diarrhea, measles, malaria, and `HIV/AIDS (WHO, 2015), most of which are preventable. There is evidence that mortality varies largely by SES and rural- urban place of residence. For example, a study by Raihan Sharif et al. (1993) investigated spatial patterns of crude death rates in 1990 and found regional variation across Bangladesh, with the lowest mortality level in the South-western part of the country, which they attributed to main economic activity groups and higher in-migration rates into the major cities and towns in this area. Ahmed et al. (2011) found that maternal mortality varied significantly by district in Bangladesh with http://dx.doi.org/10.1016/j.healthplace.2017.08.012 Received 21 December 2016; Received in revised form 19 June 2017; Accepted 29 August 2017 ⁎ Corresponding author at: Humboldt-Universität zu Berlin, Geography Department, Berlin, Germany. E-mail address: oliver.gruebner@gmail.com (O. Gruebner). Health & Place 47 (2017) 156–164 1353-8292/ © 2017 Elsevier Ltd. All rights reserved. MARK