The diffusion of autism spectrum disorder in Costa Rica: Evidence of information spread or environmental effects? David Schelly a,n , Patricia Jiménez González b , Pedro J. Solís c a Department of Sociology, University of Wisconsin, 8128 William H. Sewell Social Sciences Building,1180 Observatory Dr., Madison, WI 53706-1393, USA b Hospital Nacional de Niños “Dr Sáenz Herrera”, CCSS, Child Developmental and Behavioural Unit, San José, Costa Rica c University of Costa Rica, San José, Costa Rica article info Article history: Received 11 February 2015 Received in revised form 16 July 2015 Accepted 31 July 2015 Keywords: Autism spectrum disorder Cluster detection Social epidemiology Latin America abstract In the U.S., children with autism spectrum disorder (ASD) have been found to live in spatial clusters. Studies have suggested that the clustering is caused by social or environmental factors, but determining the cause of the clustering is difficult in the U.S. setting because of unmeasured variation in healthcare access and diagnostic practices. The present study explores the diffusion of ASD in a small setting in which the diagnosis is not widely publicised and there is no variation in healthcare access or diagnostic practices. Costa Rica provides universal healthcare and only has one diagnosing clinic for young children, and the diagnosis is relatively new and little known among clinicians and parents. In addition, the po- tential for mercury exposure from the source that has been associated with ASD is absent, and areas with high levels of air pollution are spatially concentrated. Focusing on all young children who underwent an ASD assessment from 2010 to 2013, we identify spatial clusters that suggest a mechanism that does not depend on information about ASD, healthcare access, diagnostic practices, or environmental toxicants. These findings provide details of the “contextual drivers” of the increasing worldwide prevalence of ASD. & 2015 Elsevier Ltd. All rights reserved. 1. Introduction In the U.S., children with autism spectrum disorder (ASD) have been found to live in spatial clusters (Mazumdar et al., 2010; Hoffman et al., 2012). From a public health perspective, these clusters mean that children in certain neighbourhoods—often overrepresented by Latinos, African Americans, and low socio- economic status (SES) families (Durkin et al., 2010)—are less likely than others to receive important services that can improve life outcomes (Liptak et al., 2008). Identifying the causes of the clus- ters will inform public health officials on how to remediate these inequities and inform us on “contextual drivers” of the increasing prevalence of ASD (Mazumdar et al., 2013). Theories about the causes of the ASD clustering are diverse. For example, sociologists have found strong positive correlations be- tween the diagnosis and both SES and parental education (Van Meter et al., 2010; Hoffman et al., 2012), and others have found anecdotal evidence of a social network effect on the diagnosis (Liu et al., 2010). In California, ASD prevalence in low SES neighbour- hoods nearly caught up to the rate in high SES neighbourhoods, and this happened as information about ASD was becoming widespread (King and Bearman, 2011). In other studies, biologists have identified associations between ASD clusters and congenital malformations of the reproductive system, which can be caused by exposure to toxicants during pregnancy (Rzhetsky et al., 2014). Others have linked the spatial patterns to exposure to heavy me- tals (Windham et al., 2006), air pollution from traffic(Becerra et al., 2013), and mercury from coal-fired power plants (Palmer et al., 2006; Palmer et al., 2009). Determining the cause of the clustering is difficult in the U.S. setting for several reasons. Specifically, there is substantial varia- tion in access to healthcare (Kirby and Kaneda, 2006; Wang and Luo, 2005; Williams, 2012; related to ASD specifically, see Liptak et al., 2008; Fountain and Bearman, 2011) and the distribution of diagnosing clinics (e.g., Mazumdar et al., 2013), and the clinics engage in highly variable diagnostic practices (Bresnahan et al., 2009; Charman et al., 2009); there are even insurance incentives that variously promote the diagnosis of ASD (Grinker, 2007). Thus, diagnostic variation across clinics that are unequally distributed across space and populations could in itself lead to spatial patterns in the diagnosis, and the clustering patterns could be indis- tinguishable from clustering related to information spread or ex- posure to environmental toxicants. All of these factors would likely create spurious correlations with neighbourhood and demo- graphic characteristics (e.g., Kirby et al., 2006). Another difficulty in the U.S. setting, particularly for hypotheses Contents lists available at ScienceDirect journal homepage: www.elsevier.com/locate/healthplace Health & Place http://dx.doi.org/10.1016/j.healthplace.2015.07.007 1353-8292/& 2015 Elsevier Ltd. All rights reserved. n Corresponding author. E-mail address: dschelly@ssc.wisc.edu (D. Schelly). Health & Place 35 (2015) 119–127