Reanalysis of the association between reduction in long-term PM2.5 concentrations and improved life expectancy

Table 2 Estimates of the increase in life expectancy associated with a reduction in PM_2.5 of 10 μg/m³ over approximately 20 years between 1979–1983 and 1999–2000, adjusted for socioeconomic, demographic, and proxy indicators for prevalence of smoking across 211 U.S. counties, using PM_2.5 concentrations based on Metropolitan Statistical Area (MSA) averages of Inhalable Particulate Network (IPN) or Federal Reference Method (FRM) monitoring data vs. county averages of predictions at census tract centroids and national grid coordinates estimated by the historical prediction model

Model^a	N of counties	Regression coefficient ± standard error
		Measured PM_2.5	Modeled PM_2.5
		Measured PM_2.5	Census tract	National grid
1	211^b	0.72 ± 0.29^c	1.39 ± 0.55^c	1.04 ± 0.56
2	211	0.83 ± 0.20^c	1.14 ± 0.49^c	1.02 ± 0.38^c
3	211	0.60 ± 0.20^c	0.54 ± 0.40	0.64 ± 0.31^c
4	211	0.61 ± 0.20^c	0.69 ± 0.31^c	0.81 ± 0.29^c
5	127	0.55 ± 0.24^c	0.06 ± 0.41	0.54 ± 0.35
6	51	1.01 ± 0.25^c	0.16 ± 0.94	0.34 ± 0.64
7	51	0.94 ± 0.23^c	0.46 ± 0.76	0.63± 0.50

^a Model 1: PM_2.5
Model 2: Model 1 + income + population + 5-yr in-migration + high-school graduates + urban residence + black population + Hispanic population
Model 3: Model 2 + lung-cancer mortality rate + COPD mortality rate
Model 4: Model 1 + income + population + black population + lung-cancer mortality rate + COPD mortality rate
Model 5: Model 4 in 127 counties with the population greater than 100 thousand
Model 6: Model 3 in 51 counties each of which had the large population in each MSA
Model 7: Model 4 in 51 counties each of which had the large population in each MSA
^b The same number of counties to that in Pope et al. (2009)’s analysis including some combined counties to have sufficient numbers of deaths for computing life expectancy as described in Pope et al. (2009)
^c P < 0.05

ISSN: 1476-069X