Association of education and non-partnered relationship with all-cause and cardiovascular mortality before and during the COVID-19 pandemic in a prospective population-based cohort study

Background There is considerable literature suggesting that health inequalities increased during the COVID-19 pandemic, including conditions other than those immediately caused by COVID-19. Much of the evidence comes from routine unlinked data. Studies examining health inequalities during the pandemic in individual-level data remain sparse, particularly in Central and Eastern Europe. Purpose To compare, in a prospective population-based cohort, the educational inequalities in mortality from all causes and from cardiovascular diseases (CVD) and coronary heart disease (CHD) in the Czech Republic before and during the COVID-19 pandemic. We also examined the role of partnered relationship. Methods We used data from the Czech arm of the multi-centre HAPIEE cohort, which recruited 8,857 men and women aged 45-69 years selected from population register at baseline in 2002-2005. The cohort was linked with national mortality register until 2022. Association between education (classified as primary, vocational, secondary and tertiary) and mortality was analysed (1) over the full follow up 2002-2022, (2) pre-pandemic period 2017-2019, and (3) pandemic period 2020-21. Participants marital status was classified into partnered (married/cohabiting) vs. non-partnered (single/divorced/widowed) categories. Age- and sex-adjusted odds ratios were estimated for each outcome. Results Over 2002-22, the numbers of deaths from all causes, CVD and CHD were 2771, 1020 and 488; in pre-pandemic period 516, 210 and 105; and in pandemic period and 512, 163 and 74, respectively. Over 2002-2022, there was a pronounced educational gradient in all mortality indicators; the age-sex adjusted hazard ratios for lowest vs. highest educational categories were 2.2 (1.9-2.5) for all causes, 2.8 (2.1-3.65) for CVD and 3.7 (2.6-5.41 for CHD. Comparing pre- vs. pandemic periods, the adjusted odds ratios were 2.6 (1.8-3.8) vs. 1.8 (1.2-2.6) for all-causes, 4.0 (2.1-7.7) vs. 1.2 (0.7-2.3) for CVD and 3.2 (1.4-7.2) vs. 2.6 (1.0-7.0) for CHD (Figure 1). The pattern was similar in men and women. In analyses of mortality in partnered vs. non-partnered subjects, the protective effect of living with a partner disappeared during the pandemic period; for all causes, the odds ratio changed from pre-pandemic 1.35 (1.08-1.68) to pandemic 0.85 (0.67-1.14) and for CVD mortality it changed from 1.71 (1.23-2.37) in pre-pandemic period to 0.75 (0.50-1.14) in pandemic period; again, the pattern was similar in men and women. Conclusions The relative educational gradient in mortality from all causes, CVD and CHD remained relatively stable over the pandemic period, in contrary to the original hypothesis that inequalities were widening. We observed a reduction or reversal of the protective effect of living with a partner; the interpretation of this finding is unclear. Some of the inconsistencies in the findings may be partly due to the smaller numbers of deaths in pre- and pandemic periods.