Assessing risk in focal arboviral infections: Are we missing the big or little picture?

Abstract

Background: Focal arboviral infections affecting a subset of the overall population present an often overlooked set of challenges in the assessment and reporting of risk and the detection of spatial patterns. Our objective was to assess the variation in risk when using different at-risk populations and geographic scales for the calculation of incidence risk and the detection of geographic hot-spots of infection. We explored these variations using a pediatric arbovirus, La Crosse virus (LACV), as our model. Methods and Findings: Descriptive and cluster analyses were performed on probable and confirmed cases of LACV infections reported to the Tennessee Department of Health from 1997 to 2006, using three at-risk populations (the total population, the population 18 years and younger, and the population 15 years and younger) and at two geographic levels (county and census tract) to assess the variation in incidence risk and to investigate evidence of clustering using both global and local spatial statistics. We determined that the most appropriate at-risk population to calculate incidence risk and to assess the evidence of clustering was the population 15 years and younger. Based on our findings, the most appropriate geographical level to conduct spatial analyses and report incidence risk is the census tract level. The incidence risk in the population 15 years and younger at the county level ranged from 0 to 226.5 per 100,000 persons (median 41.5) in those counties reporting cases (n = 14) and at the census tract level it ranged from 50.9 to 673.9 per 100,000 persons (median 126.7) in those census tracts reporting cases (n = 51). To our knowledge, this is the highest reported incidence risk for this population at the county level for Tennessee and at the census tract level nationally. Conclusion: The results of this study indicate the possibility of missing disease clusters resulting from performing incidence risk investigations of focal diseases using inappropriate at-risk populations and/or at large geographic scales. Improved disease surveillance and health planning will result through the use of well defined at-risk populations and the use of appropriate geographic scales for the analysis and reporting of diseases. The finding of a high incidence risk of LACV infections in eastern Tennessee demonstrates that the vast majority of these infections continue to be under-diagnosed and/or underreported in this region. Persistent prevention and surveillance efforts will be required to reduce exposure to infectious vectors and to detect new cases of infection in this region. Application of this study's observations in future investigations will enhance the quantification of incidence risk and the identification of high-risk groups within the population. © 2009 Haddow et al.