COVID-19: Seroprevalence data and policymaking

COVID-19 - A survey of seroprevalence studies

Epidemiologists and decision-makers need reliable methods to measure the extent of the COVID-19 in specific populations. But clinical testing often misses cases of mild or asymptomatic infection and provides little information about populations that have not been adequately tested. By contrast, seroprevalence studies systematically measure the presence of antibodies to the COVID-19 virus in random population samples - providing an objective estimate of the proportion of the population that has been infected. In the last year, there have been hundreds of these studies, each offering invaluable information.  Here Frontiers Evidence Snapshots brings this information together in a unique data resource, summarizing results from studies all over the world.

This resource includes:

• Expert commentary from Dr. Milo Puhan, Swiss School of Public Health

• 301 studies (links to original papers provided)

• 600 separate population samples

• 69 countries in 5 continents

• Data for special populations (health workers, hospital patients, school children, the elderly, other vulnerable groups)

• New studies added periodically

To generate this dataset, the authors searched the scientific literature for journal articles, technical reports, and preprints referring to studies of the seroprevalence of the SARS-CoV-2 (COVID) virus.  The search, conducted between 21 October 2020 and 14 January 2021,  used the query (SARS-Cov-2 OR Covid) AND (seroprevalence OR serological OR seroconversion). The team downloaded all articles that met the following inclusion criteria: the article reported a seroprevalence study performed by the authors of the article, the article specified the location where the study was conducted, the characteristics of the study population, the start and end dates for the study, the size of the sample, and the crude prevalence rate for the study population. Each downloaded article was analyzed by two members of the team working independently. In cases where the information from the two members did not match, the data collected were verified against the original article. These pages present a summary of the data.

The survey of seroprevalence studies was conceived and led by Richard Walker, with the collaboration of Cho Enriquez and Victoria Marquez (team leads), Patricia Ilagan, Joseph Rasalan, and Nathan Afalla

Corona pandemic as big boost for seroepidemiological studies – considerations for optimally informing policy-making 

Milo A. Puhan, University of Zurich, Department of Epidemiology & Corona Immunitas, Swiss School of Public Health 

We are currently experiencing an unprecedented boom of studies that use antibody tests to determine the development of (humoral) immunity against SARS-CoV-2 in the general population or specific subpopulations. Seroepidemiological studies capture the spread of a virus like SARS-CoV-2 much more completely than virus testing (by PCR or rapid antigen tests), with a median ratio of seroprevalence estimates to the corresponding cumulative incidence of SARS-CoV-2 infections of 14.5 across 125 studies included in a systematic review (Bobrovitz et al., 2020). Seroepidemiological studies provide estimates of seroprevalence for entire populations or specific subgroups but also serve as reference estimates for other research efforts including modelling studies.  

Seroepidemiological studies have great potential to inform policy makers on many different levels, from specific work, educational or social settings to entire countries (Altmann et al., 2020). Estimates of seroprevalence provide a compass through a pandemic: They allow for comparisons of seroprevalence estimates between the general population and specific subpopulation (e.g. particularly vulnerable or exposed groups), determine the spread of a virus like SARS-CoV-2 geographically and over time, and help determine the impact of public health measures and protective strategies in specific settings like schools or nursing homes. Currently, seroprevalence estimates, ideally paired with evidence on neutralizing activity, are of great importance to inform the planning or surveillance of vaccine campaigns.  

The collection of 258 studies presented by Frontiers shows the diversity of seroepidemiological studies performed in countries around the globe and predominantly in the first half of 2021. They include populations of high interest to decision makers and often have appropriate sample sizes. About half of them included samples from the general populations whereas the other half focused on specific settings or subgroups. A recently updated systematic review of 338 studies shows a similar picture (Bobrovitz et al., 2020).  

These two overviews highlight a number of limitations of the current evidence on seroprevalence. The majority of studies are at moderate or high risk of bias because of their sampling schemes, selection of test, lack of adjustment for test inaccuracy or because of other reasons. Also, approaches for large-scale testing like using remaining samples from routine health care may, unlike in previous epidemics give biased estimates of seroprevalence because the strict public health measures influence health care seeking behavior. Even more importantly, the seroprevalence estimates are mostly not comparable within and across countries or across time, which may limit their usefulness for policy makers.  

The SARS-CoV-2 pandemic provides an unprecedented opportunity for seroepidemiological studies to inform the current management of the pandemic but also provides many lessons for future pandemics. The following considerations may help to increase the usefulness of seroepidemiological studies for policy makers:  

1. A collaborative approach of researchers and policy makers sharpens the goal of a seroepidemiological study and increases its relevance for decision makers.  

1. The precise definition of the target population and a sampling scheme that yields a study population representative for that target population is key.  

1. The relevance of a seroepidemiological study increases by creating comparisons. For example, seroprevalence in a vulnerable or exposed population is only interpretable if there is a reference to the general population. The same applies to comparison across regions or countries. Also, a comparison over time is more informative for observing the spread of a virus than cross-sectional snap shots.  

1. Seroprevalence estimates should be as timely as possible but not at the price of validity. Population-based studies with random sampling of participants takes time and it is not simple to short cut study procedures. In contrast, setting specific studies like in schools can be done within a short period of time so that results can be communicated quickly as the Ciao Corona study of 55 randomly selected schools in Switzerland shows (Ulyte et al., 2020).   

1. Study protocols should be harmonized as much as possible. The WHO protocol for seroepidemiological studies is available since the beginning of the pandemic, but harmonization should go much beyond that. If sampling procedures, study designs, questionnaires, antibody tests and statistical analyses as key components of seroepidemiological studies are harmonized as much as possible comparisons as those mentioned above become possible and the resulting evidence will be even more relevant for policy makers than it currently is the case. International initiatives and collaborations are key to foster such harmonization. 

It is very understandable that these considerations for seroepidemiological studies happen to a limited extent in this current pandemic only. Each country is so concerned with managing the pandemic in its country that collaboration and coordination happens at best within single countries like the Corona Immunitas initiative in Switzerland.(West et al., 2020) But the learning curve with seroepidemiological studies and how they can be as informative for policy makers as possible is very steep so that the current and for sure future pandemics will benefit from an constantly improving evidence base on seroprevalence in the general population and specific groups and settings.  

Altmann, D. M., Douek, D. C., and Boyton, R. J. (2020). What policy makers need to know about COVID-19 protective immunity. Lancet 395, 1527–1529. doi:10.1016/S0140-6736(20)30985-5. 

Bobrovitz, N., Arora, R. K., Cao, C., Boucher, E., Liu, M., Rahim, H., et al. (2020). Global seroprevalence of SARS-CoV-2 antibodies: a systematic review and meta-analysis. medRxiv, 2020.11.17.20233460. doi:10.1101/2020.11.17.20233460. 

Ulyte, A., Radtke, T., Abela, I. A., Haile, S. R., Berger, C., Huber, M., et al. (2020). Clustering and longitudinal change in SARS-CoV-2 seroprevalence in school-children: prospective cohort study of 55 schools in Switzerland. medRxiv, 2020.12.19.20248513. doi:10.1101/2020.12.19.20248513. www.ciao-corona.ch 

West, E. A., Anker, D., Amati, R., Richard, A., Wisniak, A., Butty, A., et al. (2020). Corona Immunitas: study protocol of a nationwide program of SARS-CoV-2 seroprevalence and seroepidemiologic studies in Switzerland. Int. J. Public Health 65. doi:10.1007/s00038-020-01494-0. www.corona-immunitas.ch