Background: the exposure to a coal-fired power plant has been shown to increase mortality both for cardiovascular and respiratory causes among an exposed cohort in comparison with a cohort of unexposed. Hazard ratios between 1.30 and 1.90 were found for cardiovascular and respiratory mortality.
Objectives: to estimate the individual life shortening among the exposed due to power plant emissions.
Design: survival for cardiovascular and respiratory disease in the exposed vs unexposed groups was estimated by the Kaplan-Meier method. For each gender and exposure, a fictitious cohort with a cumulative 30-year follow up was built combining three subcohorts of age at entry of 55-64, 65-74, and 75-84 years, with 10 years of follow up each. Survivals at 10 years in the 55-64-year subcohort were used as initial risks for 65-74-year subcohort; then, survivals at 10 years of the 65-74-year subcohort were used as initial risk in the 75-84-year subcohort. Eventually, 30-years cumulative follow up cohorts were obtained by gender and exposure. Individual life-shortening in people exposed was estimated as time from death of an exposed subject to the subsequent time when the unexposed cohort reached the same risk of the exposed subject at that time of the death. Here, it is proposed a method to take into account causes other than those considered.
Setting and participants: 144,018 subjects aged 55-74 years at entry of both genders belonging to the open cohort of residents of 12 municipalities (including Savona) from 2001 to 2013 in the area where the coal-fired power plant of Vado-Quiliano (Liguria Region, Northern Italy) is located.
Main outcome measures: individual life shortening.
Results: after 5 years of follow up, the individual life shortening due to cardiorespiratory causes varied between 972 and 1,822 days for males and from 612 and 1,578 days among females. Taking into account other causes of death, reduces slightly (3% for males of 75 years at death) the estimate of life shortening found in this study. The comparison between the cohorts requires that the exposed and unexposed groups are comparable, except for the exposure, and that causes other than those considered are taken into account. Socioeconomic status had been found to have little effect on cause-specific death risk indicating that, at least in terms of socioeconomic status, the exposed and unexposed groups were similar. Taking into account causes other than those considered slightly reduced the found estimates (3% at age 75 in males). According to the proposal, the life-shortening for the considered causes is easy to calculate and provides an individual indicator of damage. Inferring from group statistics individual estimates could be the most controversial point of this approach. The proposed estimates are the most credible estimate of individual damage for each occurred death among the exposed people.
Conclusions: an increased hazard ratio for a wide series of causes is equivalent to a life shortening among the exposed. A method to produce reasonable estimates of life-shortening is proposed as the effect of exposure at individual level. This approach is simple and do not require sophisticated statistical tools. It appears a promising approach for other settings.
Keywords: individual life shortening; survival analysis; coal fired power plant.