Fig. 5
Steps used to develop a risk-based model of CHD burden from black carbon. a shows the exposure continuum for black carbon over all census tracts as described above. As a starting point, the overall age-adjusted average death rate for CHD for the entire county was arbitrarily assigned to the exposure metric for BC (0.975895 μg/m3) corresponding to the census tract that marks the population midpoint (ie., 50% of population lies below and 50% of population lies above this point). The overall risk of CHD death was then adjusted upward or downward based on the BC increment in each preceding and succeeding tract based on the BC effect estimate for CHD mortality reported by Gan et al. [20]. b shows the BC-dependent contributions to risk relative to the BC exposure present in the mid-population census tract. c expresses the BC-dependent risk value relative to the BC metric contained in the lowest exposed census tract. d shows the overall predicted BC-dependent CHD death burden in each tract by applying the risk values shown in panel C to the population of each census tract. When the disease burden is summed across all census tracts in the county, the model estimates a total of 48.8 CHD deaths arising from BC in excess of that present in the lowest exposed census tract, or 2% of the entire CHD deaths measured in the county