Covid-19 pandemic policy monitor (COV-PPM) - European level tracking data of non-pharmaceutical interventions

The Covid-19 Pandemic Policy Monitor (COV-PPM) dataset prospectively documents non-pharmaceutical interventions (NPIs) taken to contain SARS-Cov-2 transmission across countries in EU27, EEA and UK. In Germany, measures have also been recorded at the federal state and, partially, at the district levels. NPIs implemented since January 2020 have been retrieved and updated weekly from March 2020, from official governments webpages, Ministries of Health, National (Public) Health Institutes or Administrations. NPI categories collected refer to restrictions, closures or changes in functioning implemented in 13 domains: public events (gatherings in indoor or outdoor spaces); public institutions (kindergartens, schools, universities); public spaces (shops, bars, restaurants); public transport (trains, buses, trams, metro); citizens movement/mobility (e.g. pedestrians, cars, ships); border closures (air, land or sea, all incoming travels, from high-risk regions, only non-nationals); measures to improve the healthcare system (e.g. human resources or technical reinforcement, redistribution, material or infrastructural); measures for risk/vulnerable groups (e.g. elderly, chronically ill, pregnant); economic measures (e.g. lay-off rules establishment, actions to avoid job-loss, tax relaxation); testing policies (e.g. testing criteria changes); nose and mouth protection rules, vaccination and others/miscellaneous measures.


a b s t r a c t
The Covid-19 Pandemic Policy Monitor (COV-PPM) dataset prospectively documents non-pharmaceutical interventions (NPIs) taken to contain SARS-Cov-2 transmission across countries in EU27, EEA and UK. In Germany, measures have also been recorded at the federal state and, partially, at the district levels. NPIs implemented since January 2020 have been retrieved and updated weekly from March 2020, from official governments webpages, Ministries of Health, National (Public) Health Institutes or Administrations. NPI categories collected refer to restrictions, closures or changes in functioning implemented in 13 domains: public events (gatherings in indoor or outdoor spaces); public institutions (kindergartens, schools, universities); public spaces (shops, bars, restaurants); public transport (trains, buses, trams, metro); citizens movement/mobility (e.g. pedestrians, cars, ships); border closures (air, land or sea, all incoming travels, from high-risk regions, only non-nationals); measures to improve the healthcare system (e.g. human resources or techni-cal reinforcement, redistribution, material or infrastructural); measures for risk/vulnerable groups (e.g. elderly, chronically ill, pregnant); economic measures (e.g. lay-off rules establishment, actions to avoid job-loss, tax relaxation); testing policies (e.g. testing criteria changes); nose and mouth protection rules, vaccination and others/miscellaneous measures.
© 2021 The Author(s

Value of the Data
• COV-PPM dataset allows answering multiple research questions related to the impacts of specific measures on the SARS-Cov-2 pandemic across different geographical regions (32 European countries, including subnational data for Germany) or population subgroups and will also enable associations to be tested between higher-level characteristics (e.g., macroeconomic indicators) and NPI effectiveness, accounting for other contextually relevant characteristics of the population. • This dataset allows interested researchers to triangulate data, cross-validate NPI patterns and / or temporal sequences as collected by different initiatives and hence adds to the scientific process of consolidating knowledge on NPIs and their effectiveness in the European Region. • The raw data is freely available from open repository and further queries (data covering 2021, German-level data) can be made to the responsible researchers, so that multiple approaches can be taken to analyze NPIs effectiveness. This is particularly important as there is no "gold standard" method for measuring NPIs as complex, multi-layered social response to the pandemic.

Data Description
The non-pharmaceutical intervention (NPI) categories covered in this dataset, comprehensively cover several societal domains and regional resolutions implemented in countries of EU-27, EEA, UK and across the 16 German federal states, as detailed in Table 1 .
The proportion of days each measure was in place across the 32 European countries during the observation period from January 2020 until December 2020 are shown in Table 2 , allowing to track patterns and differences between countries. For the 16 German federal states differences in specific domains can be tracked also within the country (Table S1, Supplementary Material).
The analysis of textual elements and qualifiers (i.e., keywords identifying "restriction" or "withdrawal" added to each record) retrieved with each NPI description allows the creation of a subset of variables with disaggregation of broader selected categories: for example, in the case of public institutions, distinct variables were created to identify measures specifically aiming at (a) schools, (b) kindergartens, (c) universities/higher education, (d) gastronomy and (e) nightlife facilities. In the case of public events, this allows the creation of variables measuring changes to (a) public events indoors, (b) outdoors, and (c) in general.
A detailed codebook of the resulting variables and description, from the disaggregation of measures impacting public institutions in the 16 federal states in Germany, are provided in the Supplementary Material (Table S2). Fig. 1 shows the result of NPIs related to schools, for the  Table 1 Non-pharmaceutical interventions monitored by the COV-PPM in EU-27, EEA, UK and in the German federal states.

Domain of the policy measure Description of information collected
Sub-categories specifying scope of measures (or regional resolution) Public events Major restrictions implemented are recorded, considering limits in the number of persons allowed for different types of public events or activities, in indoor or outdoor spaces. This may include conferences, amateur and professional collective sports, concerts, festivals, etc.
Mandates to conduct health risk assessment for mass gatherings by authorities or collection of emergency contact details from participants was also recorded.

Public institutions
Major restrictions or closures are considered and recorded for schools, universities, public services, etc. Restrictions to these institutions may include institution-wide intermediary measures aimed at physical distancing practices or enhancing hygiene measures (e.g. increasing in cleaning and disinfecting practices, in school buildings, classrooms, water and sanitation facilities), adjustment to space or set up of infrastructures (e.g. spacing of tables in schools), home office options for some or all staff. Records are kept if any criteria for restriction or closure pertains to type of institution, size or specific location.
Not specified (code 2); Single cities (code 2a); State level (code 2b); National (code 2c); Public spaces Restrictions and closures in public spaces are recorded (including shops, bars, swimming halls, gyms, restaurants, etc.), and any intermediary systematic measure taken aiming at keeping physical distancing and reducing transmission risk (e.g. cleaning and disinfecting measures). Adjustments to space, set up or infrastructures are also recorded (e.g. number of tables in restaurants, limitations in number of simultaneous visitors inside stores).
Not specified (code 3; Single cities (code 3a); State level (code 3b); National (code 3c); Public transport Measures affecting the transportation services and mobility of the public within a country, region or single city, regardless of the service being privately or publicly owned and run, were recorded and specified. This includes trains, buses, trams, metro transport. Restrictions to be recorded may include intermediary systematic measures (e.g. to keep physical distancing and reduce transmission risk), such as reduction in maximum number of passengers capacity, enhancing of hygiene and disinfecting measures and frequency, reduction of routes/frequency, measures to protect drivers (e.g. not opening of front doors in buses, or restriction in the sales of tickets by drivers).
Not specified (code 4); Single cities (code 4a); State level (code 4b); National (code 4c); Movement/mobility Restrictions in the movement of individuals by any means other than a public transport (e.g. pedestrians, by car, plane, or boat/ship), are recorded and specified regarding its application to entire national territories or specific regions, or single cities.
Not specified (code 5); Pedestrians (code 5a); Private cars (code 5b); Aviation national travel code 5c; Others (ships, trains, etc.) code 5d; Travelling border closure All border closures are recorded, specifying if restrictions applied to travelling by air, land or sea, across national borders. Quarantine and mandatory testing are also recorded. A distinction was also made regarding specific locations and criteria for which the restriction applied, namely if it applies to all travelers, to non-nationals only, to non-nationals from specific "high-risk areas", or any other criteria used (or not specified).
Not specified (code 6); For non-nationals from high-risk regions (code 6a); for all non-nationals (code 6b); for all incoming travelers (code 6c); ( continued on next page ) Mask policies Mandatory or recommended use of mouth and nose protection (in specific settings, indoors or outdoors, etc), are recorded together with details of recommended context of utilization (e.g., age range, in public transports, schools, etc).

(code 12)
Vaccination Strategy officially documented for vaccination (e.g. risk groups prioritized, specific programs implemented, channeled through "normal" health system or other specificity); whenever available, name of vaccine(s), producer/company, type of vaccine(s) are also documented.

Table 2
Proportion of days in which NPIs were in place in 32 countries (EU/EEA/UK) during the observation period (January 2020 to December 2020).
Vaccination measures are not included because their implementation started after the period considered for this table.
16 German federal states. The same subset of variables can be generated for application at the EU-level.

Experimental Design, Materials and Methods
Since January 2020 and ongoing, the Covid-19 Pandemic Policy Monitor (COV-PPM) was established to record on a weekly basis, the non-pharmaceutical interventions (NPIs) taken in the EU27, EEA and UK (total of 32 countries). Additionally, data for Germany, where the research team is based, are being collected also at the federal state level (NUTS2) and partially at the district-level (NUTS3).
The categories of NPIs scoped were initially developed based on the knowledge available at the beginning of the pandemic with respect to measures being implemented in different countries and contexts, and complemented with the policies detailed in previous WHO reports and guidance (which are mainly based on the policies taken during the 2009 H1N1 influenza pandemic) [1] . The categories were progressively refined and adapted as the Covid-19 pandemic unfolded and different measures appeared (e.g. masks or mouth and nose protection, and vaccination rules and procedures).
The sources consulted for the recording of NPIs (recommended or enforced) were official websites of national health authorities (e.g. Ministries of Health, National Public Health Institutes and any specific websites managed by national governments to convey information about Covid-19), followed by national news websites. For German federal states and districts, information from health authorities of the corresponding geographical level was retrieved. Google translate was used for website translation whenever English versions of websites were not available. The team of researchers/assistants included German, French, Spanish and Portuguese speakers, that helped understand and document specific elements whenever translations provided dubious wording.
For each NPI category the team coded the exact dates of implementation and provided a textual description of the relevant information needed to contextualize each measure, saving the respective primary source (saved to pdf format, ensuring traceability of the collected information, since official health authorities websites, for example, are updated regularly).
The textual description added to each measure at the time of their coding, was preceded by selected keywords to allow qualification and recoding of NPIs according to their increasing or decreasing stringency. Specifically, for each point in time attached to the start, change or stop of each NPI category, a qualifier was added to identify and count if the coding refers to an additional "restriction", "withdrawal" or "recommendation". For the specific case of Masks utilization, two identifiers were added also to the categories where mouth and nose protection were additionally implemented as mandatory or recommended (e.g. enforcement of masks utilization in public transports, public institutions, public events).
To build our dataset, we recorded information in a long format with days as observation unit, coding each NPI from its starting to its ending date, as found in the corresponding sources. The text recorded following each keyword (i.e. after the start, ending or change of a measure indicated by the keywords "restriction", withdrawal" or "recommendation"), provides a summarized description retrieved from the original source, that enables identification of the scope of application (for e.g. a withdrawal in public events: "WITHDRAWAL: Wedding celebrations (after the actual wedding ceremony) are regarded as events without assigned and marked seats and can therefore take place with up to 100 people, regardless of whether they take place in a building or in the open-air area. From August 1st up to 200 people are allowed ." The team of researchers and assistants collecting the data, conducted frequent consensus meetings (weekly) and continuously updated the protocol for data collection according to changes also occurring throughout Europe, to encompass the nuances of increased or decreased stringency being implemented.
A Quality Assessment (QA) exercise was also conducted for the data gathered at the EU-level, and at the federal and district levels in Germany. The QA procedure consisted of a quantitative component and a qualitative one.
In the quantitative component, the team plotted the measures collected using interactive graphs, allowing a simple visualization of the aggregated data (example provided in supplementary material - Fig. S1). Based on these plots, the team searched for individual data points for each NPI to check for their accuracy and for potential non-plausible temporal sequences of NPIs (e.g. restrictions to public events or public transports coded for very short periods or with several interruptions in the time-series, measures coded before February 2020).
Following identification of potential coding mistakes visualized in the graphs, the individual country or federal state datasets were checked for the coding and a textual description recorded to assess whether an error was made in the documentation.
The qualitative component of the QA procedure consisted in analyzing the numerical coding and comments/textual elements saved for about 5-10% (around 2 weeks) of the total time sequence of each NPI and for each country, German federal state, and districts. The selection of the proportion of time to examine was based on the amount of comments/textual information registered (sequences with more frequent elements were selected but could vary for each NPI and for each geographical unit), and based on the available human resources. The numerical coding and textual information were then compared with the saved primary source to identify mismatches between the information provided in the sources and in the database. Two independent reviewers completed the QA process described for the German federal states, while most of the QA procedure at the EU level was conducted by a single reviewer. Any incongruencies found were then corrected.
The subnational dataset for Germany was recorded separately because of the federal nature of the country, which implies a high degree of autonomy for NPI implementation at subnational level. Although separate, the German subnational dataset was established following the same methodology and coding scheme used for European country-level, so that it can be directly merged into a single dataset.
A qualitative assessment of the text included for each subcategory is also being conducted, in order to establish even more fine-graded categories of measures that emerge across the different countries and German federal states, thus adding a layer of specificity to the analysis potential. Analysis of this qualitative information gathered, which is retrieved from countries' official governmental sources, allows disentangling the change that occurred to the nature of some NPIs, for example, from complete closures to conditions under which specific public or private spaces can operate or host events (and which would have been aggregated as the same measure, otherwise). This also provides the needed flexibility to compare the data gathered with other existing public policy trackers for Covid-19, allowing for cross-validation.
This dataset shares some of the approaches and rationales of other trackers (e.g. with respect to covered domains of NPIs), but also differs from other effort s in it s methodology [2][3][4][5] . It aims to monitor NPIs prospectively, as they are being implemented, across 32 European nations, and with a special focus in Germany, where NPIs are being documented at the federal state and partially at district level. The German data collected (at the level of the 16 federal states) continues to be updated as measures are implemented and can be of great use to understand the effectiveness of NPIs taken at the local level, and establish further comparisons with the effectiveness measurable at the national and European levels. The federal nature of the German government implies a degree of autonomy for NPIs implementation, whose effectiveness to mitigate the diseases transmission can be tested if measures are available at such disaggregated levels. While continuation of recording of data at level of 401 German districts was desirable, the task could not be continued due to resource constraints.

Ethics Statements
This project collects publicly available data, thus ethics approval was not needed.