Targeting Interventions for Youth and Pregnant Women with Chlamydia through Multilevel Modeling and Mapping Spatio-Temporal Trends

Background: Chlamydia incidence has increased in Kalamazoo County, MI, since the late 1990s and has now reached a rate of nearly twice the state average. Rates in the 15-24 year olds are particularly high, as are rates of infant morbidity and mortality, particularly in disadvantaged populations. Also, sexually transmitted infections (STI) can co-occur in this population. The purpose of this analysis was to combine individual level records from several different data sources with census tract level socioeconomic status and built environment to create a multilevel model and geographic information systems (GIS) maps of spatio-temporal trends in the disease to tailor local resource planning and intervention placement to specific populations, including youth and pregnant woman. Methods: Using multilevel statistical modeling and GIS maps, we analyzed nine years (2006-14) of chlamydia laboratory cases reported to the local health department and reported chlamydia cases among women giving birth (2006-2012 vital birth records) by race, age and gender (individual-level), and poverty, education, unemployment and urban/rural characteristics at the census tract level (U.S. Census) to identify spatio-temporal patterns. Results: Over 13,000 chlamydia cases were geocoded. Those less than 18 accounted for over 1,000 cases in 2012-14. Of these cases, 70% were within 1 of the 12 school districts in the county.  In addition, 7specific geographic hotspots for chlamydia were identified among those giving birth. Conclusions: Through examination of combined data sources, including vital records, laboratory results, and census data, we were able to obtain detailed information to target STI interventions across different high-risk populations.