Skip to main content

Table 4 Attributable deaths in Milan, Italy (2003-2006) calculated using different approaches

From: Assessing short-term impact of PM10 on mortality using a semiparametric generalized propensity score approach

SourceNatural DeathsRespiratory DeathsCardiovascular Deaths
 Estimate90% CIEstimate90% CIEstimate90% CI
GPS- AD(40)1157(689, 1645)312(210, 418)771(580, 961)
GPS- DAD(40)1857(1479, 2233)403(293, 504)1014(815, 1221)
PS matching 401079(116, 2042)305(17, 593)716(117, 1315)
Regression 40358(156, 560)    
  1. GPS- AD(40): deaths attributable to daily exposure levels above 40 μg/m3, setting the counterfactual exactly to 40 μg/m3.
  2. GPS- DAD(40)): deaths attributable to daily exposure levels above 40 μg/m3, sampling the counterfactuals from the exposure distribution below 40 μg/m3.
  3. PS matching 40: from Baccini et al. [25], deaths attributable to daily exposure levels above 40 μg/m3, calculated according to a PS matching.
  4. Regression 40: from Baccini et al. [37], deaths attributable to exceeding the limit of 40 μg/m3 for the annual average level of exposure, calculated from a regression approach