LAUSR

OBJECTIVES to describe changes in relative and absolute inequalities in mortality by education level between 2001 and 2016 in the Emilian longitudinal study (SLEm) and to estimate the impact of these inequalities at population level. DESIGN closed cohort study based on record-linkage between municipal population registries, Census archives of 2001 and 2011, and the mortality register. SETTING AND PARTICIPANTS 2001- and 2011-Census respondents >=30 years old residing in Bologna, Modena, or Reggio Emilia followed up to the age of 75 years, death, emigration, or end of follow-up (December 2006 or December 2016). MAIN OUTCOME MEASURES premature mortality for all causes and for 16 groups of causes known to be associated with socioeconomic position. In order to capture various aspects of the inequalities, the association with the education level is assessed through summary regression-based indexes (Relative and Slope Index of Inequality) and the Attributable Population Fraction. RESULTS premature mortality declined across all educational level between 2001-2006 and 2011-2016; declines were greater among men than women. Among men, relative inequalities in mortality slightly increased (RII from 1.86 in 2001 to 2.13 in 2011), while absolute inequalities declined (SII from 382.3 to 360.6). Among women, both relative and absolute inequalities increased (RII from 1.23 to 1.65, SII from 73.7 to 137.4). Educational inequalities in lung cancer, respiratory and cerebrovascular diseases mortality decreased among men and increased among women. The proportion of the low educated shrank over time (men: from 40% to 36%; women: from 43% to 35%); nonetheless, the fraction of the deaths attributable to educational inequalities showed an upward tendency (from 18.5% to 21.9% in men and from 9.7% to 15.6% in women); the groups of causes that contribute most to this increase were malignant cancers, especially lung cancer, diseases of the circulatory and respiratory systems, and accidents. CONCLUSIONS relative inequalities slightly increased in both genders, while absolute inequalities only in women. A reduction in the population impact could be achieved by tackling educational inequalities in mortality due to lung cancer, diseases of the circulatory and respiratory systems, and accidents.