Using individual level survival prediction to estimate years of life lost due to Covid-19 and their socio-economic distribution

Several studies suggest that mortality due to Covid-19 is concentrated among people with a low socioeconomic status. The extent to which this socioeconomic inequality in mortality translates into an inequality in years of life lost is largely unknown because years of life lost are highly dependent on underlying health conditions. Using administrative data, we estimated individual-level survival probabilities for the Dutch population, using predictors such as medication use, hospital and nursing home admissions, disposable income, gender and age. Using the estimated model, we predicted years of life lost for those who died from Covid-19. Results from our analyses show that there is a huge variation in years of life lost across the population, suggesting that not only the vulnerable and sick died from Covid-19. Differences in the burden of disease due to Covid-19 by income are caused by differences in the number of deaths, and not by the number of years of life lost per death.