SOCRATES-CoMix: a platform for timely and open-source contact mixing data during and in between COVID-19 surges and interventions in over 20 European countries.
Free full text in Europe PMC
Abstract
Background
SARS-CoV-2 dynamics are driven by human behaviour. Social contact data are of utmost importance in the context of transmission models of close-contact infections.Methods
Using online representative panels of adults reporting on their own behaviour as well as parents reporting on the behaviour of one of their children, we collect contact mixing (CoMix) behaviour in various phases of the COVID-19 pandemic in over 20 European countries. We provide these timely, repeated observations using an online platform: SOCRATES-CoMix. In addition to providing cleaned datasets to researchers, the platform allows users to extract contact matrices that can be stratified by age, type of day, intensity of the contact and gender. These observations provide insights on the relative impact of recommended or imposed social distance measures on contacts and can inform mathematical models on epidemic spread.Conclusion
These data provide essential information for policymakers to balance non-pharmaceutical interventions, economic activity, mental health and wellbeing, during vaccine rollout.Free full text
SOCRATES-CoMix: a platform for timely and open-source contact mixing data during and in between COVID-19 surges and interventions in over 20 European countries
2 Sarah Vercruysse,2 Amy Gimma,3 Pietro Coletti,2 Jantien A. Backer,4 Kerry L. M. Wong,3 James Wambua,2 Kevin van Zandvoort,3 Lander Willem,1 Laurens Bogaardt,4 Christel Faes,2 Christopher I. Jarvis,3 Jacco Wallinga,4,5 W. John Edmunds,3 Philippe Beutels,1,6 and Niel Hens1,2Abstract
Background
SARS-CoV-2 dynamics are driven by human behaviour. Social contact data are of utmost importance in the context of transmission models of close-contact infections.
Methods
Using online representative panels of adults reporting on their own behaviour as well as parents reporting on the behaviour of one of their children, we collect contact mixing (CoMix) behaviour in various phases of the COVID-19 pandemic in over 20 European countries.
We provide these timely, repeated observations using an online platform: SOCRATES-CoMix. In addition to providing cleaned datasets to researchers, the platform allows users to extract contact matrices that can be stratified by age, type of day, intensity of the contact and gender. These observations provide insights on the relative impact of recommended or imposed social distance measures on contacts and can inform mathematical models on epidemic spread.
Conclusion
These data provide essential information for policymakers to balance non-pharmaceutical interventions, economic activity, mental health and wellbeing, during vaccine rollout.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12916-021-02133-y.
Background
Modelling a pandemic shaped by human behaviour
From the outbreak in December 2019 onwards, SARS-CoV-2 dynamics have been shaped by human behaviour [1]. For this reason, policymakers’ responses have been largely centred around social distancing measures to limit the burden of COVID-19 and to prevent healthcare systems from collapsing [2, 3]. Such measures—aimed at reducing the effective contact rate in society—will likely remain part of policymakers’ strategy until a substantial proportion of the population has been successfully vaccinated.
The importance of including social contact information in transmission models for close-contact infectious pathogens has been widely acknowledged in the literature, with the ‘social contact hypothesis’ [4] and the POLYMOD study [5] marking important milestones in the development and parameterization of such models. Over the past decades, social contact data have been increasingly used and collected in the context of transmission models of close-contact infections. A 2019 systematic review that retrieved 64 social contact studies reported common traits in terms of number of daily face-to-face conversational contacted persons (typically around 10 to 20) and general age-dependencies despite a variety of study designs [6].
Social contact pattern data have been indispensable for modelling SARS-CoV-2 transmission [7, 8]. Indeed, a number of SARS-CoV-2 modelling studies were capable of accurately and consistently predicting a variety of epidemiological parameters by relying on social contact data [9, 10]. Comparing reproduction numbers estimated from seroprevalence and virologic data to reproduction numbers estimated from social contact data in England, Davies et al. further validated the use of social contact data in the context of SARS-CoV-2 modelling [11]. Yet, adequate parameterization of such models requires country-specific social contact data collected under different policy interventions (e.g. lockdown versus no lockdown) and at different stages of the pandemic (e.g. in-between surges and during various ascending and descending stages of the pandemic).
Construction and content
CoMix: measuring behavioural change during the COVID-19 pandemic
CoMix is a longitudinal, multi-country social contact survey in representative panels of individuals in terms of age, gender, region of residence and—for most countries—either socio-economic status, occupation or educational attainment. The CoMix study started in March 2020, with survey data first being collected in the United Kingdom (UK), Belgium, and the Netherlands. It was set up to monitor awareness and behavioural changes during the pandemic. Each wave, panel members are invited to fill out the CoMix survey. On the survey day, participants retrospectively report all social contacts made from 5
am on the day preceding the survey up to 5am on the day of the survey. A contact is defined as an in-person conversation of at least a few words or a skin-contact [9]. For every first wave, the target quota is set at 1500 participants, while a drop-out rate of 5 to 10% is allowed for every subsequent wave (except for some UK panels that were replenished with newly recruited participants and had a higher quota of 2500 later in the survey). When a significant proportion of the panel is lost to follow-up—after sending three reminder invitations—additional panel members are recruited up to the point where the sample matches the target quota. A CoMix wave refers to one period of survey data collection, running from the point when the invitations are sent up to the point when the quota are met and the survey is closed. We refer to the works by Jarvis et al. and Coletti et al. for further methodological details [7, 9]. Apart from participants’ social contacts, the survey also records individuals’ risk perceptions, such as the perceived severity of COVID-19, perceived susceptibility to COVID-19 and the perceived effectiveness of social distancing measures. Due to its longitudinal nature, the survey is particularly suited to quantify how changes in non-pharmaceutical interventions (NPIs) and changes in perception influence NPI compliance and social contact behaviour over time [12–15]. While most data is collected on behaviour in adults, a proportion of the respondents report contacts on behalf of their children. This provides crucial information about social mixing behaviour in children (and adults) when circumstances change (e.g. schools open versus closed) [8].
In October 2020, several European countries were faced with a surge in COVID-19 cases and had to resort to a second lockdown. Given the diverse range of policy measures in place across Europe [16]—and the central role of social contact data in the parameterization of infectious disease models—the CoMix study was extended to another 17 European countries. In addition, we invited research teams in Norway and Germany (COVIMOD study) that adapted the original CoMix survey to join the initiative in order to set up a collaborative network [17, 18]. The map in Fig. Fig.11 shows the European countries that have been collecting social contact data within the context of—or similar to—the CoMix study. Figure Figure22 depicts an overview of the available and planned survey waves for all countries. Sample characteristics can be found in more detail in Additional file 1: Tables S1 to S19. We refer to the work by Coletti et al. [9] for sample characteristics for waves 1 to 8 in Belgium.
Map of Europe showing the geographical spread of CoMix and CoMix-like data collection
Overview of completed and scheduled CoMix data collection in Europe, as of 1 July 2021. Data collection in initial CoMix countries (as part of the EpiPose project) is depicted in blue, while extended CoMix data collection and data collection in collaboration with EpiPose partners are depicted in yellow and green, respectively. Colours correspond to the colours used in the map in Fig. Fig.1.1. NA: Not applicable due to sample statistics not yet being available. * Estimated number or estimated timing. ** The participation rate is defined as the number of participants that completed the entire survey relative to the number of participants that opened the survey link. ‡ Due to data management issues during the initial data collection phase, parts of the data for Finland, Switzerland and Lithuania in Q1 2021 was removed by Ipsos due to quality concerns. As a result of a limited size of valid CoMix data collected for the Q1 2021 period, additional data is now being collected from June 2021 onwards
The CoMix data proved valuable to quantify the impact of social distancing measures over the course of the COVID-19 epidemic in the UK, Belgium and the Netherlands [7, 9, 13, 14, 19], and results are in line with other ongoing studies into social contact patterns with a different study population [19]. The extension of the CoMix study thus provides the opportunity to evaluate policies more accurately within and across a further 17 European countries.
An open-source platform to extract SOcial Contact RATES (SOCRATES) from over 20 European countries
The typical CoMix data flow is reflected in Fig. Fig.3,3, yet deviations from this scheme are present. The data flow starts from the ‘master’ version of the CoMix questionnaire that was developed and implemented early in the pandemic in the UK (24 March 2020). In collaboration with local partners in each country, the questionnaire is adapted to countries’ circumstances and languages, after which the fieldwork is implemented by a market research company. The CoMix data are cleaned and validated according to a data management protocol, the details and code which can be found on a GitHub repository [20]. After data cleaning, the data is stored and prepared for sharing in the public Zenodo-based repository (accessible via: www.socialcontactdata.org/data) as well as on the CoMix-Socrates tool. Furthermore, we invited partners performing a CoMix-like survey (Fig. (Fig.1)1) to also store their data on the public repository. CoMix data are analysed to gain insights at the national and international level which are converted into advice for health policymakers. The CoMix study protocols and questionnaires were approved—or waivers were obtained—by local ethical committees, the details of which can be found in Additional file 2: Table S20.
Schematic overview of the different steps in the CoMix study. The figure reflects the typical data flow for most European countries, yet deviations from this scheme are present in some. Abbreviations: LSHTM, London School of Hygiene and Tropical Medicine; UHasselt, Hasselt University; EpiPose, Epidemic intelligence to minimize COVID-19’s public health, social and economic impact. Ipsos is a commercial market research company
Utility and discussion
Usefulness and limitations of the SOCRATES-CoMix platform
As described by Willem et al. [3], the SOCRATES tool allows users to extract contact matrices and contact rates by country and survey wave. While the initial tool provides contact rates from a variety of contact studies by country and year, the newly developed SOCRATES-CoMix tool focuses on contact rates collected during the SARS-CoV-2 pandemic. The SOCRATES-CoMix tool allows stratification by age (user-picked age groups), type of day (week versus weekend), intensity of the contact (physical versus non-physical) and gender (see Fig. Fig.4).4). Other features such as weighing by age and handling of missing data provides end-users the opportunity to match the social contact data extracted to their model requirements. We refer to the work of Willem et al. for further methodological details on how these data are summarised [3]. The platform is updated on a regular basis, adding social contact data from additional waves as they become available. The SOCRATES-CoMix tool can be found here: http://www.socialcontactdata.org/socrates-comix/.
A print screen of the SOCRATES-CoMix tool. This specific example shows a social contact matrix using data collected in wave 2 of the Belgian CoMix study with four age classes, for weekdays and physical contacts only
The SOCRATES platform brings together social contact data from over 20 European countries collected at different points in time throughout the SARS-CoV-2 pandemic. Furthermore, the SOCRATES tool provides timely, cleaned and protracted social contact rates that can directly be integrated into mathematical models, while allowing for easy and quick data stratification, for instance by age or type of day, by use of a drop-down selection menu (Fig. (Fig.4).4). The platform also directs researchers to the public datasets on Zenodo repositories. CoMix data will continue to be of utmost importance in assessing the future course of the SARS-CoV-2 pandemic and to design effective public health policies. For example, in light of optimal COVID-19 vaccine roll-out, testing strategies and gradually less stringent NPIs [12].
Nevertheless, the CoMix data and the SOCRATES-CoMix tool have limitations. That is, the CoMix data are being collected in an online environment such that participants can only take part when they have access to a digital device, e.g. a personal computer, laptop, tablet or smartphone, and an internet connection. That means that CoMix data for older age groups are likely more prone to selection bias [21]. Selection bias, nevertheless, remains a concern for other age groups as well. In addition, self-completed surveys during a pandemic might be prone to social desirability bias, given mandatory social distancing measures and policymakers explicitly relying on the public’s social responsibility to protect others. However, social desirability bias may be minimal in view of the anonymous data entry, without direct contact with an interviewer. Lastly, due to respondents retrospectively reporting social contacts, the CoMix data may be prone to recall bias. Yet, given that participants are reporting all contacts made between 5am the day preceding the survey and 5am of the day of the survey, we believe the scope for recall bias is very limited.
Conclusion
Epidemic modelling can be enhanced with data describing contact patterns of individuals. In order to understand, model and respond to the COVID-19 pandemic in a timely manner, there was a great need for social contact data. As the virus does not stay within a country’s borders, the data collection has been set up in many European countries.
Scientific and policy-related insights can be drawn for each country, but even more, cross-country analyses are feasible. Combined with mathematical models, these data provide insights for policymakers, balancing non-pharmaceutical interventions, economic activity, mental health and wellbeing, also during the vaccine rollout.
Supplementary Information
Acknowledgements
We gratefully acknowledge the efforts of all teams involved in the implementation of the CoMix study in their country. More specifically: in Austria: the team of Andrea Schmidt at the Austrian National Public Health Institute, in Croatia: the team of Goranka Petrović at the Croatian Institute of Public Health (CIPH), in Denmark: the team of Michael Bang Petersen at Aarhus University, in Estonia: the team of Sigrid Vorobjov at the National Institute for Health Development, in Finland: the team of Kari Auranen at the Finnish Institute for Health and Welfare (THL), in France: the team of Guillaume Béraud at the Centre Hospitalier Universitaire de Poitiers, in Greece: the team of Elpida Pavi at the University of West Attica (UniWA), in Hungary: the team of Gergely Röst at the University of Szeged, in Lithuania: the team of Marija Jakubauskiene at Vilnius University, in Malta: the team of Tanya Melillo at the Ministry for Health, in Poland: the team of Magdalena Rosinska at the National Institute of Public Health - National Institute of Hygiene, in Portugal: the team of Baltazar Nunes at the National Health Institute Doutor Ricardo Jorge (INSA), in Slovakia: the team of Henrieta Hudečková at Comenius University in Bratislava, in Slovenia: the team of Petra Klepac at the National Institute of Public Health, in Spain: the team of Cristina Vilaplana at the Institute for Health Science Research Germans Trias i Pujol (IGTP) that acknowledges support from the SMA-TB Project from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 847762, in Italy: the team of Daniela Paolotti at the ISI Foundation, in Switzerland: the team of Nicola Low at the University of Bern, in Germany: the team of André Karch and Veronika Jaeger at the Universität Münster, in Norway: the team of Bjarne Robberstad at the University of Bergen and at the National Institute of Health Norway.
We acknowledge support from the European Centre for Disease Prevention and Control (ECDC) in setting up the collaborations between the Epipose consortium, and universities and public health institutions in all other countries. We gratefully acknowledge the tremendous efforts put in all the steps by the EpiPose consortium, its collaborators and Ipsos. We thank Bieke Vanhoutte for research management support making CoMix possible. FV and PB acknowledge support from the Antwerp Study Centre for Infectious Diseases (ASCID) as well as the VAX-IDEA methusalem consortium. PC and NH acknowledge support from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement 682540 TransMID). LW acknowledges support from the Research Foundation – Flanders (FWO) (postdoctoral fellowship 1234620N).
Abbreviations
| CoMix | Contact mixing |
| UK | United Kingdom |
| NPI | Non-pharmaceutical intervention |
| SOCRATES | Social contact rates |
| NA | Not applicable |
| LSHTM | London School of Hygiene and Tropical Medicine |
| UHasselt | Hasselt University |
| EpiPose | Epidemic intelligence to minimize COVID-19’s public health, social and economic impact |
Authors’ contributions
WJE, JW, NH and PB initiated the study. CIJ, WJE and KvZ designed the initial CoMix survey in consultation with PB and NH. CIJ, AG, KLMW, KvZ and PC cleaned and prepared the data for the platform. LW and NH constructed the SOCRATES platform. FV wrote the first draft, with LH, JW, CF and SV contributing. SV created the manuscript figures. JAB and LB provided sample statistics and methodological details for the CoMix study in The Netherlands. FV, LH, SV and PC significantly revised the manuscript in response to two reviewer reports. All authors contributed and reviewed the manuscript and approved the final version for publication.
Funding
This work is funded by the EpiPose project from the European Union’s SC1-PHE-CORONAVIRUS-2020 programme, project number 101003688. We are grateful for additional funding to support CoMix or CoMix-like data collection in Belgium: Belgian Health Institute - Sciensano and Janssen Belgium; Germany: Institute of Epidemiology and Social Medicine at the University of Münster, Institute for Medical Epidemiology Biometry and Informatics at Martin Luther University Halle-Wittenberg, Robert-Koch-Institute Berlin, Helmholtz-Gemeinschaft Deutscher Forschungszentren e.V. via the HZEpiAdHoc “The Helmholtz Epidemiologic Response against the COVID-19 Pandemic” project, and the German Free State of Saxony via the SaxoCOV project; Norway: University of Bergen and the National Institute of Public Health Norway; the Netherlands: Ministry of Health, Welfare and Sport - the Netherlands; and the UK: Medical Research Council MRC (MC_PC_19065), National Institute of Health Research NIHR (CV220-088 - COMIX).
Availability of data and materials
The datasets generated and/or analysed during the current study are available in the Zenodo-based repository, www.socialcontactdata.org/data, as well as on the CoMix-Socrates tool, http://www.socialcontactdata.org/socrates-comix/.
Declarations
The CoMix study protocols and questionnaires were approved—or waivers were obtained—by local ethical committees, the details of which can be found in Additional file 2: Table S20.
Not applicable.
Not applicable.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
Articles from BMC Medicine are provided here courtesy of BMC
Full text links
Read article at publisher's site: https://doi.org/10.1186/s12916-021-02133-y
Read article for free, from open access legal sources, via Unpaywall:
https://bmcmedicine.biomedcentral.com/counter/pdf/10.1186/s12916-021-02133-y
Citations & impact
Impact metrics
Citations of article over time
Alternative metrics
Discover the attention surrounding your research
https://www.altmetric.com/details/114230560
Article citations
Vaccination strategies for different contact patterns: weighing epidemiological against economic outcomes.
Int J Health Econ Manag, 24 Sep 2024
Cited by: 0 articles | PMID: 39316345
Social contact patterns during the early COVID-19 pandemic in Norway: insights from a panel study, April to September 2020.
BMC Public Health, 24(1):1438, 29 May 2024
Cited by: 2 articles | PMID: 38811933 | PMCID: PMC11137890
The impact of quality-adjusted life years on evaluating COVID-19 mitigation strategies: lessons from age-specific vaccination roll-out and variants of concern in Belgium (2020-2022).
BMC Public Health, 24(1):1171, 26 Apr 2024
Cited by: 0 articles | PMID: 38671366 | PMCID: PMC11047051
Estimating household contact matrices structure from easily collectable metadata.
PLoS One, 19(3):e0296810, 14 Mar 2024
Cited by: 1 article | PMID: 38483886
Projecting COVID-19 intensive care admissions for policy advice, the Netherlands, February 2020 to January 2021.
Euro Surveill, 29(10), 01 Mar 2024
Cited by: 0 articles | PMID: 38456214 | PMCID: PMC10986673
Go to all (34) article citations
Data
Data behind the article
This data has been text mined from the article, or deposited into data resources.
BioStudies: supplemental material and supporting data
Similar Articles
To arrive at the top five similar articles we use a word-weighted algorithm to compare words from the Title and Abstract of each citation.
Social contact patterns during the early COVID-19 pandemic in Norway: insights from a panel study, April to September 2020.
BMC Public Health, 24(1):1438, 29 May 2024
Cited by: 2 articles | PMID: 38811933 | PMCID: PMC11137890
The influence of COVID-19 risk perception and vaccination status on the number of social contacts across Europe: insights from the CoMix study.
BMC Public Health, 23(1):1350, 13 Jul 2023
Cited by: 9 articles | PMID: 37442987
Social contact patterns following the COVID-19 pandemic: a snapshot of post-pandemic behaviour from the CoMix study.
Epidemics, 48:100778, 29 Jun 2024
Cited by: 1 article | PMID: 38964131 | PMCID: PMC11413520
Transmission of SARS-CoV-2 by children: a rapid review, 30 December 2019 to 10 August 2020.
Euro Surveill, 27(5), 01 Feb 2022
Cited by: 2 articles | PMID: 35115076 | PMCID: PMC8815097
Review
Free full text in Europe PMC
Funding
Funders who supported this work.
Horizon 2020 Framework Programme (1)
Grant ID: 101003688
Medical Research Council (1)
Understanding the dynamics and drivers of the COVID-2019 epidemic using real-time outbreak analytics
Professor Sir John Edmunds, London Sch of Hygiene & Tropic. Medicine
Grant ID: MC_PC_19065
