key: cord-0750838-0v55vvfd authors: Islam, N.; Lopez, F. J. G.; Jdanov, D. A.; Royo-Bordonada, M. A.; Khunti, K.; Lewington, S.; Lacey, B.; White, M.; Morris, E. J.; Zunzunegui, M. V. title: Unequal impact of the COVID-19 pandemic on excess deaths, life expectancy, and premature mortality across Spanish regions in 2020 and 2021 date: 2021-11-29 journal: nan DOI: 10.1101/2021.11.29.21266617 sha: b5ae73eb25c96d5b2912480f9230279dcac61644 doc_id: 750838 cord_uid: 0v55vvfd Spain is one of the most heavily affected countries by the Covid-19 pandemic. In this study, we estimated the regional inequalities in excess deaths and premature mortality in Spain. Between January 2020 and June 2021, an estimated 89,200 (men: 48,000; women: 41,200) excess deaths occurred in the 17 Spanish regions with a substantial variability (highest in Madrid: 22,000, lowest in Canary Islands: -210). Highest reductions in life expectancy at birth (e_0) in 2020 were observed in Madrid (men: -3.48 years, women: -2.15), Castile La Mancha (men: -2.67, women: -2.30), and Castile and Leon (men: -2.00, women: -1.32). In the first six months of 2021, the highest reduction in e_0 was observed in Valencian Community (men: -2.04, women: -1.63), Madrid (men: -2.37), and Andalusia (men: -1.75; women: -1.43). In some Spanish regions, life expectancy at age 65 during the Covid-19 pandemic in 2020 was comparable to that observed as far back as 20 years ago. Since the emergence of Covid-19, countries and jurisdictions have employed a wide range of public health policy interventions with a view to minimising the impact of the pandemic. [1] [2] [3] [4] [5] [6] [7] These measures have affected many socioeconomic determinants of health, [8] [9] [10] [11] including the provision of healthcare services. [12] [13] [14] [15] Measuring the overall impact of the Covid-19 pandemic presents a multitude of challenges, including differences in the accuracy and completeness of reported deaths from 16] Therefore, 'excess deaths' (observed minus expected deaths) is widely considered the gold standard in estimating the overall impact of the pandemic since robust data on allcause mortality is less sensitive to misclassification in designating the cause of deaths. [15, [17] [18] [19] [20] [21] We have previously reported large differences between reported Covid-19 deaths and estimated excess deaths associated with the Covid-19 pandemic in 2020 at a national level, [22] but variations in excess deaths within countries at a regional level has not been well described. However, 'excess deaths' does have limitations as a measure of impact. In particular, it does not account for the age at death. [23] [24] [25] Analysis of life expectancy (LE) and Years of Life Lost (YLL) provides a more granular estimation of premature mortality. Life expectancy is an indication of the average number of years that people can expect to survive if the age-specific death rates of that year remain unchanged for the remainder of their life. [26] [27] [28] YLL takes into account the age distributions at death by giving larger weights to deaths at younger ages. [24] While LE is a widely used standardised measure based on a synthetic life table cohort, YLL is based on the numbers of deaths observed in real populations. An analysis of excess deaths, LE, and YLL provides a comprehensive examination on the effects of the Covid-19 pandemic. [22, 25] For example, we previously showed that, with a very similar excess death rates (per groups up to 90 years of age or older, sex and region from the Spanish National Statistics Office. Weekly mortality data from first week of 2019 and to week 24 of 2021 are publicly available at https://www.ine.es/jaxiT3/Tabla.htm?t=35179. Previous research showed that comparing the pandemic estimates with a single year in the prepandemic period ignores the recent trend in health improvements and seasonal variability. [25] We therefore obtained additional granular weekly data by region, age and sex for 2015-2018 from the National Statistics Offices upon request. We also obtained monthly mortality data by five-year age groups (including infant mortality), sex and region between January 2010 and December 2020, which enabled us to use a longer reference period for the calculation of changes in life expectancy and years of life lost. We obtained the annual population figures by age, sex, and region from the continuous census of the National Statistics Office (https://www.ine.es/jaxiT3/Tabla.htm?t=10262&L=1). All annual populations were 1 July estimates, except for 2021, when 1 January estimates were used. We used our previously developed validated methodology for the estimation of excess deaths, extensive details of which have been published elsewhere. [22, 33] In summary, observed weekly deaths in 2020 in each strata (by age, sex, and regions) were compared to the stratum-specific number of expected deaths. Excess death was the difference between the observed and the expected deaths. Expected death has been estimated based on the historical trends (2015-2019) using an over-dispersed Poisson model that accounts for temporal trends, seasonal and natural variability in mortality. [22, 33] Specifically, our mean model is: We applied this algorithm separately to each sex in each region. We have validated this approach extensively, [25] and found that the difference was <0.01 years for all ages below 80 (maximum absolute difference was 0.02). In this analysis, we have validated our approach for 2020 data, and found that the mean difference was 0.10 years (median: 0.10, maximum: 0.19). We used standard life is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 analysed in this study were fully anonymised and aggregated without any identifiable information. Relative to the trend observed between 2000 and 2019, the age-specific and agestandardised death rates increased sharply in 2020, with a steeper increase in the elderly. The increase was particularly higher in Madrid, Castile La Mancha, Castile and León, and Catalonia with a greater increase in men than women (Figure 1 We estimated 89,200 (95% CI: 87,600 to 90,800) excess deaths in Spain between January 2020 and June 2021. This estimate is 10% higher than the official reported Covid-19 deaths during the same period (Table 1) . However, we also show a wide variability in the estimated excess deaths with the highest in Madrid (22,000; 21,500 to 22,500) and lowest in the Canary Islands (-210; -530 to 100). Compared to the reported number of Covid-19 deaths, the estimated number of excess deaths was 42% higher in Madrid. On the other hand, estimated excess deaths were lower than the reported Covid-19 deaths in several regions including Cantabria, Murcia, and the Basque Country. The shape of the epidemic curve was also highly variable across the regions and age groups, with most of the excess deaths occurring in the elderly population. The shape of the epidemic in most regions revealed two waves. Madrid and Castile La Mancha had more than 100% excess deaths during the first wave, and nearly 50% excess deaths in the second wave. Other regions that almost reached 50% excess deaths at the peak of the second wave were Aragon, and Castile and León (Figure 2 Murcia, Galicia, Balearic Islands, Cantabria, and Canary Islands had the lowest agestandardised excess death rates below 50 per 100,000. Age-standardised excess . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10.1101/2021.11.29.21266617 doi: medRxiv preprint deaths were higher in men than women in most regions ( Figure . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 Madrid (2480; 759 to 4190), Castile and León (2330; 1900 to 2770 , and Aragon (1620; -465 to 3700) Generally, age-standardised excess death rate was highly correlated with excess YLL (correlation coefficient: 0.91). However, there were some notable differences. For example, the age-standardised excess death rate in men was slightly lower in La Rioja (166 per 100,000) than Extremadura (179). However, excess YLL (per 100,000) was 4.3 times higher in La Rioja (2434) than Extremadura (572). Similarly, in women, Catalonia (156) had a slightly higher excess death rate than Aragon (152), but the excess YLL was more than 1.6 times higher in Catalonia (2598) than in Aragon (1618). Based on data up to the 24 th week in 2021, in men and women, albeit at a lower level than men. Our study provides a comprehensive quantitative assessment of the impact of the . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10.1101/2021.11.29.21266617 doi: medRxiv preprint the hardest hit region in Europe, 44%, followed by Lombardia, Italy, 39%, and Castile La Mancha, 34%. [45] However, this report did not take into account the recent improvements in mortality and seasonal trends by age and sex in each European region. We Our findings have important policy implication. Reported Covid-19 deaths were lower than the number of excess deaths in nine of the regions included in our study. This suggests a possible underestimation and/or incomplete identification and reporting of the Covid-19 deaths, as well as the deaths that would have not happened had the Covid-19 pandemic not occurred, due to the health system malfunction and/or collapse in some regions. In contrast, the number of reported Covid-19 deaths were smaller than the number of excess deaths in the remaining eight regions. For instance, in the Basque Country, 3200 excess deaths were estimated by the model while 4535 confirmed Covid-19 deaths were registered. Therefore, there were 1335 deaths fewer than expected, despite the reported Covid-19 deaths. Those could have been the deaths that would have happened had the Covid-19 pandemic not occurred. This is an example of 'avoided mortality', which were prevented by behavioural and environmental changes due to the Covid-19 pandemic, such as wearing a facemask, or decreased air pollution, fewer occupational and traffic injuries, as was seen in 2008 economic crisis. [46] The regions that may have had fewer deaths than expected for causes other than Covid-19 were the Valencian Community, Basque Country, Galicia, Murcia, La Rioja, Balearic Islands and Cantabria and, to a lesser extent, Andalusia. In most of these regions, the . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. epidemic in Spring 2020 was less severe and health care services did not collapse, as was the case for regions with a strong epidemic in Spring 2020, particularly in Madrid and Castile La Mancha, the only two regions where weekly excess deaths reached more than 100% during the first wave. We hypothesize that efforts to prevent the spread of Covid-19 and associated environmental changes resulted in the prevention of deaths from respiratory diseases and occupational injuries that would have occurred had the Covid-19 epidemic not occurred. [47] It is worth noting that the observed number deaths in Canary Islands was lower than expected despite having 791 reported Covid-19 deaths. This might have been related to the fact that the Canary Islands quickly responded at the start of the pandemic. The Islands reported its first case of Covid-19, a German tourist at La Gomera, on 31 January 2020, which was arguably the first case of Covid-19 in Spain. The local authority responded quickly with isolation of the detected case and contact tracing, and thereby, avoided an early spread of the pandemic in the Canary Islands. After the second case was diagnosed, public health authorities in the Canary Islands applied strict quarantine measures to prevent SARS-CoV-2 from spreading. In Madrid, the region with the highest excess deaths, the first case was diagnosed on 25 February 2020, although retrospective investigation suggested that Madrid probably had the first Covid-19 case diagnosed in Spain, by early January 2020, [43] and the counts grow exponentially in late February and early Mach, according to the official information at https://cnecovid.isciii.es/covid19/. By the time public health authorities implemented measures in Madrid, in early March 2020, widespread community transmission of SARS-COV-2 was happening. CoV-2 spreading may have had accounted for much of the differences in excess deaths and reduction in life expectancy across the regions. In this respect, Madrid had an illprepared and underfunded public health system compared to most other regions. For example, 2018 data show that public spending on health, the percentage of public expenditure on primary care, and the rate of primary care nurses in Madrid were the lowest of all regions, 3.6%, 11.0% and 0.5 per 1000 inhabitants against an average of 5.5%, 13.9% and 0.7 per 1000 inhabitants, respectively. [48] Madrid was also the region . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10.1101/2021.11.29.21266617 doi: medRxiv preprint that applied the least stringent measures to contain the spread of infection. [49] [50] [51] The variability in excess deaths found in the regions of Spain during the first wave is consistent with the variability found in the first national seroprevalence study of SARS-CoV-2 (ENE-COVID Study) conducted in the community-dwelling population. [52] According to this survey conducted between April 27 and May 11 of 2020, Madrid, the neighbouring five provinces in Castile La Mancha and two close provinces of Castile and León had the highest prevalence of IgG antibodies to SARS-Cov-2. To a certain extent, excess deaths and high seroprevalence of SARS-Cov-2 antibodies in provinces close to Madrid could have been related to population mobility, in the absence of measures to contain the spread of the virus. According to a 2019 survey conducted by the National Statistics Institute, more than 160,000 people were travelling daily to work in Madrid from other provinces, and about 120,000 came from four neighbouring provinces in Castile-la Mancha and Castile and León. [53] In the latest round of the national survey, conducted in November 2020,[54] the region of Madrid continued to have the highest prevalence (12.4%) of SARS-CoV-2. Neighbouring four provinces of Castile La Mancha, and six provinces of Castile and León had prevalence rates higher than 13%. Only one additional province, Navarra, had such a high SARS-CoV-2 prevalence rate out of the remaining 41 provinces or autonomous cities of Spain. Although the ENE-COVID study was limited to the community-dwelling population and specifically excluded institutionalized populations, excess deaths highly reflected the prevalence rate of Covid-19 infection in the community. In Spain, as in many high-income countries, Covid-19 deaths have occurred primarily in nursing homes. Older adults living in nursing homes constitute less than 1% of the total population, and less than 4% of those aged 65. However, they have contributed 30,507 of the total 82,125 covid-19 deaths (37.1%) officially reported since the beginning of the pandemic until August 8, 2021. Some protocols were applied in several regions of Spain during the spring of 2020 regarding their healthcare. More specifically, the Madrid Government issued orders to prohibit the referral to public hospitals of nursing home . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10.1101/2021.11.29.21266617 doi: medRxiv preprint residents with severe disability or cognitive impairment. As a result, nursing home residents did not receive adequate medical management; many residents died alone in the nursing homes. [55, 56] This study confirms that, in Spain, the existing sex inequality in the annual standardised mortality rate, with higher rates in men than in women, was further widened in almost all regions during the study period, in agreement with what has been observed in a previous study on 29 high income countries during the Covid-19 pandemic. [22] The analyses presented here are based on newly developed methods that allow the examination of excess deaths and premature mortality by age groups and sex. The analysis also considers period and seasonal changes that have proved to be sensitive to environmental and societal changes. This methodology has advantages over alternatives, such as those used to assess the excess deaths associated with the macroeconomic changes in Spain during the Great Recession. [46, 57] The accuracy of the number of deaths and population by region, age group and sex provided by the National Statistics Office is robust and reliable, although the most recent figures may have some small errors. The estimates for the changes in life expectancy might be slightly biased because they are based on partial data for the first six months, which in normal circumstances have a slightly higher mortality than the second half of each year. Our study provides new evidence on the direct and indirect effects of the Covid-19 pandemic on excess and premature mortality. It underscores the importance of the availability of age and sex disaggregated data for more nuanced analysis, and estimation of the direct and indirect effects of the pandemic. The Canary Islands and Cantabria stood out as the two regions that had a lower-than-expected mortality across all the age groups, in both men and women, which could potentially be attributed to the country's prevention and control strategy early in the pandemic. Our findings also suggest that many regions had an underestimation or underreporting of Covid-19 deaths, a substantial increase in non-covid-19 deaths, or both. This study also highlights the potential value of more reliable and timely monitoring of excess deaths at both a national and regional levels, to inform public health policy to mitigate the impact of the pandemic on excess deaths and premature mortality. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10. 1101 /2021 Furthermore, such data would help to detect important social inequalities in the impact of the pandemic and inform more targeted interventions. A concerted effort to improve the sharing of data, and knowledge on the impact of public health policy, would improve the present response to the pandemic in Spain and its regions, and would improve the resilience of the country to future pandemics. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10.1101/2021.11.29.21266617 doi: medRxiv preprint 3 1 S á n c h e z -Ú b e d a E F , S á n c h e z -M a r t í n P , T o r r e g o -E l l a c u r í a M , e t a l . F l e x i b i l i t y a n d B e d M a r g i n s o f t h e C o m m u n i t y o f M a d r i d ' s H o s p i t a l s d u r i n g t h e F i r s t W a v e o f t h e S A R S -C o V -2 P a n d e m i c . I n t J E n v i r o n R e s P u b l i c H e a l t h 2 0 2 1 ; 1 8 : 3 5 1 0 . d o i : 1 0 . 3 3 9 0 / i j e r p h 1 8 0 7 3 5 1 0 3 2 L e g i d o -Q u i g l e y H , M a t e o s -G a r c í a J T , C a m p . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10.1101/2021.11.29.21266617 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10.1101/2021.11.29.21266617 doi: medRxiv preprint F i g u r e 1 : C r u d e a n d a g e -s t a n d a r d i s e d a n n u a l m o r t a l i t y r a t e i n S p a n i s h r e g i o n s , 2 0 1 0 -2 0 2 0 1 ( A ) : 7 0 -7 9 y e a r s . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10.1101/2021.11.29.21266617 doi: medRxiv preprint F i g u r e 2 : E x c e s s d e a t h s i n S p a n i s h r e g i o n s i n 2 0 2 0 -2 0 2 1 2 ( A ) : W e e k l y p e r c e n t c h a n g e i n e x c e s s d e a t h s , 2 0 2 0 -2 0 2 1 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10.1101/2021.11.29.21266617 doi: medRxiv preprint 2 ( B ) : M o n t h l y e x c e s s d e a t h s b y s e x . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10.1101/2021.11.29.21266617 doi: medRxiv preprint F i g u r e 3 : A g e -s t a n d a r d i s e d e x c e s s d e a t h r a t e s i n S p a n i s h r e g i o n s i n 2 0 2 0 -2 0 2 1 3 ( A ) : M e n . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10.1101/2021.11.29.21266617 doi: medRxiv preprint F i g u r e 4 : L i f e e x p e c t a n c y i n S p a n i s h r e g i o n s i n 2 0 1 0 -2 0 2 1 4 ( A ) : T r e n d s i n l i f e e x p e c t a n c y a t b i r t h . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10.1101/2021.11.29.21266617 doi: medRxiv preprint 4 ( B ) : T r e n d s i n l i f e e x p e c t a n c y a t a g e 6 5 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10.1101/2021.11.29.21266617 doi: medRxiv preprint 4 ( C ) : C h a n g e s i n l i f e e x p e c t a n c y i n 2 0 2 0 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 29, 2021. ; https://doi.org/10. 1101 /2021 300 to 10,100) Catalonia 14736 27