Estimating Global and Country-Specific Excess Mortality During the COVID-19 Pandemic

Victoria Knutson,Serge Aleshin-Guendel,Ariel Karlinsky,William Msemburi,Jon Wakefield

DOI: https://doi.org/10.48550/arXiv.2205.09081

2022-05-19

Abstract:Estimating the true mortality burden of COVID-19 for every country in the world is a difficult, but crucial, public health endeavor. Attributing deaths, direct or indirect, to COVID-19 is problematic. A more attainable target is the "excess deaths", the number of deaths in a particular period, relative to that expected during "normal times", and we estimate this for all countries on a monthly time scale for 2020 and 2021. The excess mortality requires two numbers, the total deaths and the expected deaths, but the former is unavailable for many countries, and so modeling is required for these countries. The expected deaths are based on historic data and we develop a model for producing expected estimates for all countries and we allow for uncertainty in the modeled expected numbers when calculating the excess. We describe the methods that were developed to produce the World Health Organization (WHO) excess death estimates. To achieve both interpretability and transparency we developed a relatively simple overdispersed Poisson count framework, within which the various data types can be modeled. We use data from countries with national monthly data to build a predictive log-linear regression model with time-varying coefficients for countries without data. For a number of countries, subnational data only are available, and we construct a multinomial model for such data, based on the assumption that the fractions of deaths in sub-regions remain approximately constant over time. Based on our modeling, the point estimate for global excess mortality, over 2020-2021, is 14.9 million, with a 95% credible interval of (13.3, 16.6) million. This leads to a point estimate of the ratio of excess deaths to reported COVID-19 deaths of 2.75, which is a huge discrepancy.

Applications,Methodology

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to estimate the excess mortality globally during the period from 2020 to 2021 caused by the COVID - 19 pandemic. Specifically, the paper focuses on how to estimate the excess mortality in each country through modeling methods in the case of incomplete or unavailable data. Excess mortality is defined as the difference between the actual number of deaths in a specific period and the expected number of deaths in a normal period, that is:

\[
\delta_{c,t} = Y_{c,t} - E_{c,t}
\]

where \(\delta_{c,t}\) represents the excess mortality in country \(c\) in month \(t\), \(Y_{c,t}\) is the actual all - cause mortality in country \(c\) in month \(t\), and \(E_{c,t}\) is the expected mortality in country \(c\) in month \(t\) predicted based on historical data.

The main challenges of the paper are:
1. **Data Incompleteness**: Many countries lack high - quality monthly all - cause mortality data, especially in Africa, the Eastern Mediterranean and Southeast Asia.
2. **Data Delay**: Even if some countries can provide data, these data are often delayed and are usually reported one year or more after the death occurs.
3. **Data Quality Variation**: There are significant differences in data collection and reporting capabilities among different countries, resulting in uneven data quality.

To overcome these challenges, the authors developed a series of statistical models, including:
- **For countries with monthly data**: Use the negative binomial distribution model to fit historical data and predict the expected mortality in 2020 and 2021.
- **For countries with only annual data**: First use the negative binomial distribution model to predict the total annual mortality, and then allocate the total annual mortality to each month through the temperature model.
- **For countries without data**: Use the over - dispersed Poisson log - linear regression model in the Bayesian framework, combined with covariates (such as COVID - 19 mortality, test - positive rate, socioeconomic index, etc.) to predict the all - cause mortality per month.

Finally, the paper provides a point estimate of global excess mortality of 14.9 million, with a 95% confidence interval of 13.3 million to 16.6 million, and this result reveals a huge difference between the reported COVID - 19 deaths and the actual excess mortality.