Geographies of uncertainty in the health benefits of air quality improvements

Assessing the long-term benefits of marginal improvements in air quality from regulatory intervention is methodologically challenging. In this study, we explore how the relative risks (RRs) of mortality from air pollution exposure change over time and whether patterns in the RRs can be attributed to air quality improvements. We employed two-stage multilevel Cox models to describe the association between air pollution and mortality for 51 cities with data from the American Cancer Society (ACS) cohort (N = 264,299, deaths = 69,819). New pollution data were computed through models that predict yearly average fine particle (PM2.5) concentrations throughout the follow-up (1982-2000). Average PM2.5 concentrations from 1999 to 2000 and sulfate concentrations from 1980 were also examined. We estimated the RRs of mortality associated with air pollution separately for five time periods (1982-1986, 1987-1990, 1991-1994, 1995-1998, and 1999-2000). Mobility models were implemented with a sub-sample of 100,557 subjects to assist with interpreting the RR estimates. Sulfate RRs exhibit a large decline from the 1980s to the 1990s. In contrast, PM2.5 RRs follow the opposite pattern, with larger RRs later in the 1990s. The reduction in sulfate RR may have resulted from air quality improvements that occurred through the 1980s and 1990s in response to the acid rain control program. PM2.5 concentrations also declined in many places, but toxic mobile sources are now the largest contributors to PM in urban areas. This may account for the heightened RR of mortality associated with PM2.5 in the 1990s. The paper concludes with a three alternative explanations for the temporal pattern of RRs, each emphasizing the uncertainty in ascribing health benefits to air quality improvements.