Methodological considerations in developing local-scale health impact assessments: balancing national, regional, and local data

National-scale health impact assessments (HIAs) have been conducted for many years and have become reasonably systematized. Recently, there has been growing interest in utilizing HIA methods at local scales, in the context of Environmental Public Health Tracking and in other settings. This paper investigates the data and analytical challenges to estimating the incidence of health effects associated with changes in air pollution concentrations at the local scale, focusing on ozone and fine particulate matter. Although it could be argued that the local-scale HIA is simply a more geographically discrete version of the national-scale assessment and, therefore, has similar challenges, in practice, many key inputs in national-scale assessments are assumed to be spatially uniform or vary only at coarse geographic resolution. For a nationalscale assessment, this assumption may not contribute appreciable bias, but the bias could be significant for any individual location. Thus, local-scale assessments require more geographically resolved air quality data, concentration-response (C-R) functions, and baseline incidence rates than are often used. However, comprehensive local data may not be available, may be incomplete, or may be time-intensive and resource-intensive to develop, especially for C-R functions for which small-scale epidemiological studies will often be underpowered. Given this context, this paper considers how best to develop credible local-scale HIAs, identifying factors that contribute to variability across geographic areas, study designs, and time periods. This paper also describes which key sources of analytical uncertainty change as the scope shifts from the national to the local scale. These challenges notwithstanding, the paper concludes that a well-designed local-scale HIA, following key principles and recommendations, can be both informative and defensible.