Veronica Escamilla1, Kristen H Hampton, Dionne C Gesink, Marc L Serre, Michael Emch, Peter A Leone, Erika Samoff, William C Miller. 1. From the *Department of Obstetrics and Gynecology, University of Chicago, Chicago, IL; †Department of Epidemiology, Gillings School of Global Public Health, University of North Carolina, Chapel Hill, NC; ‡Dalla Lana School of Public Health, University of Toronto, Toronto, ON, Canada; §Department of Environmental Sciences and Engineering, Gillings School of Global Public Health; ¶Department of Geography; and ∥Division of Infectious Diseases, Department of Medicine, School of Medicine, University of North Carolina, Chapel Hill, NC; and **Division of Public Health, North Carolina Department of Health & Human Services, Raleigh, NC.
Abstract
BACKGROUND: Identifying geographical clusters of sexually transmitted infections can aid in targeting prevention and control efforts. However, detectable clusters can vary between detection methods because of different underlying assumptions. Furthermore, because disease burden is not geographically homogenous, the reference population is sensitive to the study area scale, affecting cluster outcomes. We investigated the influence of cluster detection method and geographical scale on syphilis cluster detection in Mecklenburg County, North Carolina. METHODS: We analyzed primary and secondary syphilis cases reported in North Carolina (2003-2010). Primary and secondary syphilis incidence rates were estimated using census tract-level population estimates. We used 2 cluster detection methods: local Moran's I using an areal adjacency matrix and Kulldorff's spatial scan statistic using a variable size moving circular window. We evaluated 3 study area scales: North Carolina, Piedmont region, and Mecklenburg County. We focused our investigation on Mecklenburg, an urban county with historically high syphilis rates. RESULTS: Syphilis clusters detected using local Moran's I and Kulldorff's scan statistic overlapped but varied in size and composition. Because we reduced the scale to a high-incidence urban area, the reference syphilis rate increased, leading to the identification of smaller clusters with higher incidence. Cluster demographic characteristics differed when the study area was reduced to a high-incidence urban county. CONCLUSIONS: Our results underscore the importance of selecting the correct scale for analysis to more precisely identify areas with high disease burden. A more complete understanding of high-burden cluster location can inform resource allocation for geographically targeted sexually transmitted infection interventions.
BACKGROUND: Identifying geographical clusters of sexually transmitted infections can aid in targeting prevention and control efforts. However, detectable clusters can vary between detection methods because of different underlying assumptions. Furthermore, because disease burden is not geographically homogenous, the reference population is sensitive to the study area scale, affecting cluster outcomes. We investigated the influence of cluster detection method and geographical scale on syphilis cluster detection in Mecklenburg County, North Carolina. METHODS: We analyzed primary and secondary syphilis cases reported in North Carolina (2003-2010). Primary and secondary syphilis incidence rates were estimated using census tract-level population estimates. We used 2 cluster detection methods: local Moran's I using an areal adjacency matrix and Kulldorff's spatial scan statistic using a variable size moving circular window. We evaluated 3 study area scales: North Carolina, Piedmont region, and Mecklenburg County. We focused our investigation on Mecklenburg, an urban county with historically high syphilis rates. RESULTS: Syphilis clusters detected using local Moran's I and Kulldorff's scan statistic overlapped but varied in size and composition. Because we reduced the scale to a high-incidence urban area, the reference syphilis rate increased, leading to the identification of smaller clusters with higher incidence. Cluster demographic characteristics differed when the study area was reduced to a high-incidence urban county. CONCLUSIONS: Our results underscore the importance of selecting the correct scale for analysis to more precisely identify areas with high disease burden. A more complete understanding of high-burden cluster location can inform resource allocation for geographically targeted sexually transmitted infection interventions.
Authors: Linda M Niccolai; Niiamah Stephens; Heidi Jenkins; Wanda Richardson; Stephen Q Muth; Richard Rothenberg Journal: Sex Transm Dis Date: 2007-03 Impact factor: 2.830
Authors: Kyle T Bernstein; Frank C Curriero; Jacky M Jennings; Glen Olthoff; Emily J Erbelding; Jonathan Zenilman Journal: Am J Epidemiol Date: 2004-07-01 Impact factor: 4.897
Authors: Neela D Goswami; Emily J Hecker; Carter Vickery; Marshall A Ahearn; Gary M Cox; David P Holland; Susanna Naggie; Carla Piedrahita; Ann Mosher; Yvonne Torres; Brianna L Norton; Sujit Suchindran; Paul H Park; Debbie Turner; Jason E Stout Journal: PLoS One Date: 2012-10-02 Impact factor: 3.240
Authors: Kimberly M Fornace; Henry Surendra; Tommy Rowel Abidin; Ralph Reyes; Maria L M Macalinao; Gillian Stresman; Jennifer Luchavez; Riris A Ahmad; Supargiyono Supargiyono; Fe Espino; Chris J Drakeley; Jackie Cook Journal: Int J Health Geogr Date: 2018-06-18 Impact factor: 3.918