Association between Gross National Income per capita and COVID … – BMC Public Health

Data source and study design

Using an ecological study design, we analyzed publicly available data from the WHO pooled vaccination dataset (n=228 WHO members) merged with the World Banks GNI per capita data (n=215 members). The WHO vaccination data are extracted and compiled from members reports and third parties [24]. The dataset is updated weekly with vaccine introduction and administration by WHO members. We used data as of June 4, 2022. A detailed description of the WHO vaccination dataset and data collection can be found elsewhere [7, 25]. The World Bank GNI per capita data is routinely collected and compiled using reports published by its member countries national statistical authorities, the Organization for Economic Cooperation and Development (OECD), the International Monetary Fund, or directly from countries official data. We selected 2019 GNI per capita data as an indicator of the WHO members economy prior to the pandemic. A detailed description of the World Bank dataset and data collection can be found elsewhere [26, 27]. We excluded WHO members with missing vaccination rates, or GNI per capita; thus, our sample size in the merged dataset was 192 members. All datasets are publicly available and de-identified, therefore, a review from an Institutional Review Board was not required.

The primary outcome of interest was the cumulative number of persons fully vaccinated (i.e., receiving the required number of doses as per the vaccine guideline) per 100 population). Vaccination data were pooled from reports from members and WHO review of publicly available official data or data collected and published by third-party sites [28]. The population estimations for each country were extracted from the United Nations Department of Economic and Social Affairs, Eurostat, National Statistics Office Malta, The Government of the Pitcairn Islands, and Statistics Netherlands. A detailed description of the calculation can be found elsewhere [29, 30].

Cumulative persons vaccinated with at least one dose per 100population was the secondary outcome of interest.

GNI per capita was calculated by the Word bank in U.S. dollars using the Atlas conversion factor. We created a categorical variable to classify members into different income levels based on the World Banks 2019 Gross National Income thresholds. GNI was classified into low income (less than $1,026), lower-middle income ($1,026 to $3,995), upper-middle income ($3,996-$12,375) and high income (more than $12,375). A detailed description of the calculation can be found elsewhere [26, 27]. WHO regions were our secondary exposure. WHO regions are classified into the Africa, Americas, South-East Asia, Europe, Eastern Mediterranean, and Western Pacific regions based on members respective WHO regional offices [4].

The number of vaccine types in WHO members was the covariate. We include the number of vaccine types administered by WHO members as a covariate based on literature [20, 21] and a priori knowledge of possible confounding (availability of various vaccine choices). The number of COVID-19 vaccine types used by each WHO member was categorized into 1 to 3, 4 to 6, 7 to 9, and 10 to 12. We sourced the number of vaccine types data from the WHO data repository [28].

We performed descriptive statistical analyses to determine the cumulative meannumber of persons fully vaccinated per 100 populationstratified by GNI per capita,WHO region, and the number of vaccine types used by WHO members. We performed ANOVA tests to assess significant differences in COVID-19 vaccination by WHO members GNI per capita, region, and the number of vaccine type used. We also performed negative binomial regression to assess the association between the vaccination rates and GNI per capital, WHO region and the number of vaccine types using two models. We used the negative binomial regression model because the COVID-19 vaccination outcomes were rates normalized from count data (i.e., only non-negative integer), and it managed for overdispersion in the dataset.

The first model looked at the crude association between vaccination rates and GNI per capital, WHO region and the number of vaccine types respectively. The second model examined the association between vaccination rates and GNI per capital adjusted for WHO region and the number of vaccine types. To adjust for differential follow-up time in the reporting of vaccination data between countries, we offset in the regression the natural log of the time (in weeks) between April 9th, 2021, and the most recent date each country member reported/updated vaccination data to WHO. April 9th, 2021, was the oldest most recent date a member reported vaccination data to WHO as of June 4th, 2022, therefore, was selected as the baseline follow-up date. We used SAS version 9.4 and STATA 16.1 to perform the analyses.

More here:

Association between Gross National Income per capita and COVID ... - BMC Public Health