The continued prevalence of Tuberculosis (TB) in the United States, particularly the incidence of multi-drug resistant strains of the disease, is a major public health concern. In Virginia, the statewide rate has been declining (down from 5.5 cases per 100,000 population in 1995 to 4.9 in 1999). However, in northern Virginia, which has over forty percent of the state's cases, the number of new active TB cases has been increasing. Fairfax County has the largest number of TB cases in northern Virginia. The high incidence of TB in northern Virginia is tied to its relatively large immigrant population within which the disease is concentrated. About 85 percent of active TB cases in the County are from persons with a country-of-origin other than the United States. Asians and Hispanics account for about 60 percent of the County's foreign-born cases. The state of Virginia maintains identifying information for all active TB cases in its Epi Info data base. Epi Info is public domain database and statistics software for public health professionals available over the Internet through the Centers for Disease Control and Prevention. Local TB controllers from across the state collect the information that populates the statewide data base. Controllers collect the information in the process of providing testing, treatment and education services to TB patients.

We have been analyzing the Epi data base in order to characterize the relationship of TB to the immigrant population. Our aim was to gain an understanding of TB patterns and relationships that would support local health departments in their management of TB control including locating facilities, assigning outreach resources and forecasting and justifying future service resource demands. We have shown that TB is concentrated within immigrant populations in particular neighborhoods and corridors. Further, these areas are associated with particular socioeconomic and demographic characteristics. The significance of various immigrant groups has fluctuated over the 12 years of information in the data base. These findings have strengthened arguments for targeted testing and have helped to clarify and articulate staffing requirements.

There is a strong spatial context associated with our data analysis and presentation of results. ArcView is the geographic information systems (GIS) tool for visualization of geocoded TB incidence data that we have linked to US Census socioeconomic and demographic data. Our effort has required teamwork between specialists in statistical and spatial analysis. We extracted data from 12 years of Epi records into the SAS System. We encountered data incompatibilities across those 12 years due to new releases of Epi together with varying data entry practices at the local level. The power of SAS facilitated merging compatible annual information and analyzing TB incidence at the neighborhood level to identify corridors with the highest TB risk. First, ArcView was used to geocode incidence data and link the observations to Census tract block groups. Then SAS was used to merge the TB data with US Census socioeconomic and demographic information. Through statistical and GIS analyses, we ranked Census tract block groups according to their TB rates and immigrant concentrations. The relationship between high TB rates and geographic, socioeconomic and demographic information was characterized and then visualized via GIS maps.

Our presentation will describe the interdisciplinary work of our team consisting of a County TB Coordinator, a GIS analyst and a statistician. Three distinct roles are represented: the subject area expert and practitioner, the spatial analysis expert and the data analysis expert. The interaction of the three areas of expertise led to meaningful analysis products that are proving useful in supporting policy and planning for TB control.