Parameter and model uncertainty in a life-table model for fine particles (PM2.5): a statistical modeling study

Table 4 Comparison of the results of different life-table studies.

Study	Life-expectancy effect per 10 μgm^-3 (years)	Effect per predicted exposure (years)	Predicted exposure (μgm^-3)	Relative risk estimates (mean) used in the study (per 10 μgm^-3) (RR)
Brunekreef 1997 [7], high	1.51	1.51	10.0	1.10
Brunekreef 1997 [7], low	1.11	1.11	10.0	1.10
Nevalainen and Pekkanen 1998 [9], high	1.01	1.01	10.0	1.18^a
Present study, high	0.74	0.74	10.0	See table 1
Mechler et al. 2002 [8]	0.64	1.36	21.1	1.06
Nevalainen and Pekkanen 1998 [9], low	0.60	0.60	10.0	1.20
Leksell and Rabl 2001 [13]	0.46	0.0006	1.0^b	1.17
Present study, low	0.43	0.43	10.0	See table 1
Rabl 2003 [12]	0.42	0.38	15.0^c	1.06

The results have been scaled to 10 μgm^-3 PM_2.5 exposure by assuming log-linearity between increased particle exposure and loss in life-expectancy. If two or more predictions were reported, the highest and lowest numbers were selected for the table. In the present study, the low prediction is the mean of the life-table model; the high prediction is the mean of the model without plausibility assumptions, with an imaginary birthcohort and with years 1988–1990 background hazard rates. See text for details. a) Only cardiopulmonary mortality. b) Not a life-time exposure but episode of one year increased exposure. c) Estimated for PM₁₀. Converted to PM2.5 by using formula 1 μg PM10 = 0.6 μg PM2.5.

ISSN: 1476-069X