Abstract
BACKGROUND: Measures to contain the COVID-19 pandemic led to healthcare services underutilisation during year 2020, with potential harmful consequences for people with specific needs. Our objective was to assess non-COVID-19 mortality (referred to as "non-covid") during year 2020 in a population of older people or with chronic conditions in France. METHODS: A nationally representative sample of adults aged 70 or more or with chronic conditions was followed from the beginning of the first lockdown to 31/12/2020. We used causes of death data to measure non-covid mortality during the entire study period and three additional periods: March 17 to May 10 (first lockdown), May 11 to October 29 (between the two lockdowns), October 30 to December 14 (second lockdown). We computed Standardized Mortality Ratios (SMRs) for the entire study population and for groups of interest defined upon age, sex, socioeconomic position, prevalent conditions and region of residence, using a 2016-2017 population as reference. RESULTS: Overall, non-covid deaths among the 1,594,222 individuals included (mean age 68·4 years, 54% women) were 2% lower than expected (SMR 0·98 [0·97 - 0·99]). Non-covid mortality was lower only during first lockdown for men (SMR 0·94 [0·91 - 0·98]) and during second lockdown for women (SMR 0·93 [0·90 - 0·96]). Excess non-covid mortality was observed for people with dementia (SMR 1·08 [1·03 - 1·14]), Ile-de-France residents (SMR 1·07 [1·01-1·15]) and Occitanie residents (SMR 1·14 [1·06 - 1·23]) during the first lockdown; people with end-stage renal disease (SMR 1·15 [1·03 - 1·28]) and Corse residents (SMR 1·34 [1·13 - 1·60]) between lockdowns; residents of the most deprived areas (SMR 1·05 [1·00-1·11] during second lockdown. Analyses stratified by region of residence confirmed the results for people with dementia and also revealed excess non-covid mortality for people with diabetes and cardiovascular diseases. CONCLUSIONS: Non-covid mortality was lower than expected during year 2020 among older people or with chronic conditions in France. Analyses during defined periods and analyses stratified by regions found significant excess non-covid mortality among specific groups of interest. Extending follow-up to years 2021 and 2022 is necessary to confirm non-covid excess mortality and assess the potential effect of healthcare underutilisation.