Background Mortality statistics are fundamental to public health decision making. Mortality varies by time and location, and its measurement is affected by well known biases that have been exacerbated during the COVID-19 pandemic. This paper aims to estimate excess mortality from the COVID-19 pandemic in 191 countries and territories, and 252 subnational units for selected countries, from Jan 1, 2020, to Dec 31, 2021. Methods All-cause mortality reports were collected for 74 countries and territories and 266 subnational locations (including 31 locations in low-income and middle-income countries) that had reported either weekly or monthly deaths from all causes during the pandemic in 2020 and 2021, and for up to 11 year previously. In addition, we obtained excess mortality data for 12 states in India. Excess mortality over time was calculated as observed mortality, after excluding data from periods affected by late registration and anomalies such as heat waves, minus expected mortality. Six models were used to estimate expected mortality; final estimates of expected mortality were based on an ensemble of these models. Ensemble weights were based on root mean squared errors derived from an out-of-sample predictive validity test. As mortality records are incomplete worldwide, we built a statistical model that predicted the excess mortality rate for locations and periods where all-cause mortality data were not available. We used least absolute shrinkage and selection operator (LASSO) regression as a variable selection mechanism and selected 15 covariates, including both covariates pertaining to the COVID-19 pandemic, such as seroprevalence, and to background population health metrics, such as the Healthcare Access and Quality Index, with direction of effects on excess mortality concordant with a meta-analysis by the US Centers for Disease Control and Prevention. With the selected best model, we ran a prediction process using 100 draws for each covariate and 100 draws of estimated coefficients and residuals, estimated from the regressions run at the draw level using draw-level input data on both excess mortality and covariates. Mean values and 95% uncertainty intervals were then generated at national, regional, and global levels. Out-of-sample predictive validity testing was done on the basis of our final model specification. Findings Although reported COVID-19 deaths between Jan 1, 2020, and Dec 31, 2021, totalled 5 center dot 94 million worldwide, we estimate that 18 center dot 2 million (95% uncertainty interval 17 center dot 1-19 center dot 6) people died worldwide because of the COVID-19 pandemic (as measured by excess mortality) over that period. The global all-age rate of excess mortality due to the COVID-19 pandemic was 120 center dot 3 deaths (113 center dot 1-129 center dot 3) per 100 000 of the population, and excess mortality rate exceeded 300 deaths per 100 000 of the population in 21 countries. The number of excess deaths due to COVID-19 was largest in the regions of south Asia, north Africa and the Middle East, and eastern Europe. At the country level, the highest numbers of cumulative excess deaths due to COVID-19 were estimated in India (4 center dot 07 million [3 center dot 71-4 center dot 36]), the USA (1 center dot 13 million [1 center dot 08-1 center dot 18]), Russia (1 center dot 07 million [1 center dot 06-1 center dot 08]), Mexico (798 000 [741 000-867 000]), Brazil (792 000 [730 000-847 000]), Indonesia (736 000 [594 000-955 000]), and Pakistan (664 000 [498 000-847 000]). Among these countries, the excess mortality rate was highest in Russia (374 center dot 6 deaths [369 center dot 7-378 center dot 4] per 100 000) and Mexico (325 center dot 1 [301 center dot 6-353 center dot 3] per 100 000), and was similar in Brazil (186 center dot 9 [172 center dot 2-199 center dot 8] per 100 000) and the USA (179 center dot 3 [170 center dot 7-187 center dot 5] per 100 000). Interpretation The full impact of the pandemic has been much greater than what is indicated by reported deaths due to COVID-19 alone. Strengthening death registration systems around the world, long understood to be crucial to global public health strategy, is necessary for improved monitoring of this pandemic and future pandemics. In addition, further research is warranted to help distinguish the proportion of excess mortality that was directly caused by SARS-CoV-2 infection and the changes in causes of death as an indirect consequence of the pandemic. Funding Bill & Melinda Gates Foundation, J Stanton, T Gillespie, and J and E Nordstrom Copyright (c) 2022 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license.

Background Accurate, comprehensive, cause-specific mortality estimates are crucial for informing public health decision making worldwide. Incorrectly or vaguely assigned deaths, defined as garbage-coded deaths, mask the true cause distribution. The Global Burden of Disease (GBD) study has developed methods to create comparable, timely, cause-specific mortality estimates; an impactful data processing method is the reallocation of garbage-coded deaths to a plausible underlying cause of death. We identify the pattern of garbage-coded deaths in the world and present the methods used to determine their redistribution to generate more plausible cause of death assignments. Methods We describe the methods developed for the GBD 2019 study and subsequent iterations to redistribute garbage-coded deaths in vital registration data to plausible underlying causes. These methods include analysis of multiple cause data, negative correlation, impairment, and proportional redistribution. We classify garbage codes into classes according to the level of specificity of the reported cause of death (CoD) and capture trends in the global pattern of proportion of garbage-coded deaths, disaggregated by these classes, and the relationship between this proportion and the Socio-Demographic Index. We examine the relative importance of the top four garbage codes by age and sex and demonstrate the impact of redistribution on the annual GBD CoD rankings. Results The proportion of least-specific (class 1 and 2) garbage-coded deaths ranged from 3.7% of all vital registration deaths to 67.3% in 2015, and the age-standardized proportion had an overall negative association with the Socio-Demographic Index. When broken down by age and sex, the category for unspecified lower respiratory infections was responsible for nearly 30% of garbage-coded deaths in those under 1 year of age for both sexes, representing the largest proportion of garbage codes for that age group. We show how the cause distribution by number of deaths changes before and after redistribution for four countries: Brazil, the United States, Japan, and France, highlighting the necessity of accounting for garbage-coded deaths in the GBD. Conclusions We provide a detailed description of redistribution methods developed for CoD data in the GBD; these methods represent an overall improvement in empiricism compared to past reliance on a priori knowledge.

Background The burden of fatal police violence is an urgent public health crisis in the USA. Mounting evidence shows that deaths at the hands of the police disproportionately impact people of certain races and ethnicities, pointing to systemic racism in policing. Recent high-profile killings by police in the USA have prompted calls for more extensive and public data reporting on police violence. This study examines the presence and extent of under-reporting of police violence in US Government-run vital registration data, offers a method for correcting under-reporting in these datasets, and presents revised estimates of deaths due to police violence in the USA. Methods We compared data from the USA National Vital Statistics System (NVSS) to three non-governmental, opensource databases on police violence: Fatal Encounters, Mapping Police Violence, and The Counted. We extracted and standardised the age, sex, US state of death registration, year of death, and race and ethnicity (non-Hispanic White, non-Hispanic Black, non-Hispanic of other races, and Hispanic of any race) of each decedent for all data sources and used a network meta-regression to quantify the rate of under-reporting within the NVSS. Using these rates to inform correction factors, we provide adjusted estimates of deaths due to police violence for all states, ages, sexes, and racial and ethnic groups from 1980 to 2019 across the USA. Findings Across all races and states in the USA, we estimate 30 800 deaths (95% uncertainty interval [UI] 30 300-31 300) from police violence between 1980 and 2018; this represents 17 100 more deaths (16 600-17 600) than reported by the NVSS. Over this time period, the age-standardised mortality rate due to police violence was highest in non-Hispanic Black people (0.69 [95% UI 0.67-0.71] per 100 000), followed by Hispanic people of any race (0.35 [0.34-0.36]), nonHispanic White people (0.20 [0.19-0.20]), and non-Hispanic people of other races (0.15 [0.14- 0.16]). This variation is further affected by the decedent's sex and shows large discrepancies between states. Between 1980 and 2018, the NVSS did not report 55.5% (54.8-56.2) of all deaths attributable to police violence. When aggregating all races, the age-standardised mortality rate due to police violence was 0.25 (0.24-0.26) per 100 000 in the 1980s and 0.34 (0.34-0.35) per 100 000 in the 2010s, an increase of 38.4% (32.4-45.1) over the period of study. Interpretation We found that more than half of all deaths due to police violence that we estimated in the USA from 1980 to 2018 were unreported in the NVSS. Compounding this, we found substantial differences in the agestandardised mortality rate due to police violence over time and by racial and ethnic groups within the USA. Proven public health intervention strategies are needed to address these systematic biases. State-level estimates allow for appropriate targeting of these strategies to address police violence and improve its reporting. Copyright (C) 2021 The Author(s). Published by Elsevier Ltd.