COVID-evidence is a continuously updated database of the worldwide available evidence on interventions for COVID-19.
We provide information about worldwide planned, ongoing, and completed randomized controlled trials on any intervention to treat or prevent SARS-CoV-2-infections.
This repository provides the source code to setup and update the COVID-evidence database.
- COVID-evidence
- Table of contents
- Project
- Setup COVID-evidence Basic
- Publications and Projects
- Trials assessing non-pharmaceutical interventions (NPIs) to prevent COVID-19
- Analysis of the COVID-19 research agenda in Germany
- Hydroxychloroquine/chloroquine collaborative review
- Convalescent plasma collaborative review
- The first 100 days of the clinical research response to COVID-19
- The fate of randomized clinical trials registered in the first 100 days
- Frequently Asked Questions
- What is COVID-evidence BASIC?
- What is the data flow?
- How reliable are filters?
- What are “Expansion Modules”?
- Do you share your data?
- How accurate is the data?
- Do you work with other initiatives?
- What makes COVID-evidence different from other initiatives?
- Where do I find more details on COVID-evidence?
- Who is paying?
- How can I help?
- License
With this freely available continuously updated database, we aim to concentrate all available trial evidence on benefits and harms of interventions for SARS-CoV-2 infection in an easy-accessible manner.
COVID-evidence is a non-profit initiative of the Department of Clinical Research at University of Basel and the Meta-Research Innovation Center at Stanford with other international partners and collaborators.
We aim to provide a reliable starting point for those who need to make treatment decisions, plan new trials, or develop clinical practice guidelines and systematic reviews.
COVID-evidence collects up-to-date trial information as well as links to original sources such as study protocol documents, registry entries, and publications. The COVID-evidence database is used to foster collaboration and evidence generation on COVID-19 such as our rapid reviews on mortality outcomes with hydroxychloroquine and chloroquine or on convalescent plasma.
We consider reports, registry entries, and manuscripts of trials with various designs. Eligible are trials testing any interventions actively allocated to humans to treat or prevent SARS-CoV-2-infections. Interventions include drug and non-drug treatments, diagnostic procedures, and decision-algorithms.
We include randomized controlled trials (RCTs) with no restriction on geographical regions or settings.
Current data sources:
- For trial registry entries: Systematic searches in ClinicalTrials.gov and the WHO's International Clinical Trials Registry Platform (ICTRP).
- For publications and preprints: Systematic searches in the L·OVE platform of all randomized controlled trials reporting results. The platform covers PubMed, medRxiv, bioRxiv and many other data sources.
Trial information that is not included in these sources may also be added to COVID-evidence as 'handsearch results'. We will continuously refine the screening processes and transparently document the underlying methods in the Open Science Framework (OSF; DOI 10.17605/OSF.IO/GEHFX)
Full protocol on OSF (V5.0; 1 February 2022).
The COVID-evidence database now focuses on reporting key characteristics of randomized controlled trials such as status, design features, the clinical question (i.e. the population, compared interventions, key outcomes), and setting. We are using the COVID-evidence database as a collaborative platform to foster evidence generation (Publications/Projects).
The source code of the update utility is written in GO. To build the COVID-evidence update tool an installed GO compiler is required. A bash script is available for building the binary and a run script template is provided to start an update cicle. The update utility is only tested on macOS. The data of COVID-evidence is stored in a MySQL database (≥5.7).
An easy way to run a MySQL database on a local machine is to use Docker.
# create a directory to persist data
mkdir data
# run mysql docker container publishing standard port 3306
docker run -v "$PWD/data":/var/lib/mysql \
--name mysql \
-p 3306:3306 \
-e MYSQL_ROOT_PASSWORD=secret_root_password \
-e MYSQL_DATABASE=covid_evidence \
-e MYSQL_USER=covid_evidence_admin \
-e MYSQL_PASSWORD=secret_admin_password \
-e MYSQL_PASSWORD=secret_admin_password \
-d mysql:5.7 \
--character-set-server=utf8mb4 \
--collation-server=utf8mb4_unicode_ci
The project provides a build script to compile the COVID-evidence update tool.
# change to the build directory
cd build
# compile the COVID-evidence updater
./build.sh
All tables that are used to import the data from the different sources, are provided as SQL scripts in the setup folder.
# copy the setup scripts into the mysql container
docker cp setup mysql:/tmp/
# run each setup script to create the basic database tables
docker exec mysql /bin/sh 'mysql -u covid_evidence_admin --ppassword < /tmp/setup/01_cove_basic.sql'
docker exec mysql /bin/sh 'mysql -u covid_evidence_admin --ppassword < /tmp/setup/02_screening_ictrp.sql'
docker exec mysql /bin/sh 'mysql -u covid_evidence_admin --ppassword < /tmp/setup/03_screening_clinicaltrials.sql'
docker exec mysql /bin/sh 'mysql -u covid_evidence_admin --ppassword < /tmp/setup/04_screening_love.sql'
docker exec mysql /bin/sh 'mysql -u covid_evidence_admin --ppassword < /tmp/setup/05_screening_johnshopkins.sql'
docker exec mysql /bin/sh 'mysql -u covid_evidence_admin --ppassword < /tmp/setup/06_screening_projects.sql'
docker exec mysql /bin/sh 'mysql -u covid_evidence_admin --ppassword < /tmp/setup/06_screening_topics.sql'
To import the registry entries from the WHO ICTRP you need to download the provided CSV file from the ICTRP Website. For the import of the L·OVE platform, you need a CSV file with the following fields:
reviewer_name, results_pub, title, abstract, author, publication_title, doi, registry, date_added, randomized
It is possible to update the data of all sources together or to update a single data source e.g. WHO ICTRP.
./build/bin/cove-updater -mysql="covid_evidence_admin:secret_admin_password@(127.0.0.1:3306)/covid_evidence" -tmp=./exports -ictrp=./imports/COVID19-web.csv -love=./imports/love.csv -ct -jh
| argument | required | type | description |
|---|---|---|---|
| -mysql | Yes | string | MySQL connection string: user:password@(host:port)/database_name |
| -tmp | Yes | string | Path to a temporary folder |
| -ictrp | No | string | Path to the WHO ICTRP CSV file |
| -love | No | string | Path to the love CSV file |
| -ct | No | boolean | If the argument is set, the entries of clinicaltrials.org, wil be updated |
| -jh | No | boolean | If the argument is set, the number of covid cases per country will be updted |
A systematic scoping review of registered randomized trials assessing non-pharmaceutical interventions (NPIs) to prevent COVID-19 as of August 2021 summarizing key trial characteristics and presenting results:
Hirt J, Janiaud P, Hemkens LG. Randomised trials on non-pharmaceutical interventions for COVID-19 as of August 2021: a scoping review. BMJ Evid Based Med 2022; http://dx.doi.org/10.1136/bmjebm-2021-111825
Additionally, a blog spotlight has been published: Hirt, J., Janiaud, P. & Hemkens, L. G. (2022). Why we urgently need a long-term research agenda on non-pharmaceutical interventions to guide policies and practices in the current and future public health emergencies. Blog entry written on: Randomized trials on non-pharmaceutical interventions for COVID-19: a scoping review, (bmjebm-2021-111825).
A living overview on registered trials assessing NPIs to prevent COVID-19 was launched as part of COVID-evidence. The protocol is published on Open Science Framework.
@article{hirt2022randomised,
title={Randomised trials on non-pharmaceutical interventions for COVID-19 as of August 2021: a scoping review},
author={Hirt, Julian and Janiaud, Perrine and Hemkens, Lars G},
journal={BMJ Evidence-Based Medicine},
year={2022},
publisher={Royal Society of Medicine}
}
A systematic analysis of randomized clinical trials assessing interventions to treat or prevent COVID-19 that were registered in 2020 and recruited or planned to recruit participants in Germany. Additionally, we requested recruitment accrual information from trial investigators as of April 2021.
Hirt J, Rasadurai A, Briel M, Düblin P, Janiaud P, Hemkens LG. Clinical trial research on COVID-19 in Germany – a systematic analysis. F1000Res 2021;10:913. https://doi.org/10.12688/f1000research.55541.1
Data is available on Open Science Framework.
@article{hirt2021clinical,
title={Clinical trial research on COVID-19 in Germany--a systematic analysis},
author={Hirt, Julian and Rasadurai, Abeelan and Briel, Matthias and D{\"u}blin, Pascal and Janiaud, Perrine and Hemkens, Lars G},
journal={F1000Research},
volume={10},
number={913},
pages={913},
year={2021},
publisher={F1000 Research Limited}
}
An international collaborative meta-analysis on the effects of hydroxychloroquine and chloroquine on mortality outcomes in patients with COVID-19 from all available RCTs by October 16, 2020. In addition to published data, investigators from all ongoing, completed, or discontinued RCTs were invited to share their unpublished mortality data.
Axfors C, Schmitt AM, Janiaud P, van’t Hooft J, Abd-Elsalam S, Abdo EF, et al. Mortality outcomes with hydroxychloroquine and chloroquine in COVID-19 from an international collaborative meta-analysis of randomized trials. Nature Communications. 2021 Apr 15;12(1):2349. doi: 10.1038/s41467-021-22446-z
Axfors C, Schmitt AM, Janiaud P, et al. Mortality outcomes with hydroxychloroquine and chloroquine in COVID-19: an international collaborative meta-analysis of randomized trials. medRxiv. Published online October 22, 2020:2020.09.16.20194571. doi: 10.1101/2020.09.16.20194571
Data and protocol available on the Open Science Framework.
@article{axfors2021mortality,
title={Mortality outcomes with hydroxychloroquine and chloroquine in COVID-19 from an international collaborative meta-analysis of randomized trials},
author={Axfors, Cathrine and Schmitt, Andreas M and Janiaud, Perrine and van’t Hooft, Janneke and Abd-Elsalam, Sherief and Abdo, Ehab F and Abella, Benjamin S and Akram, Javed and Amaravadi, Ravi K and Angus, Derek C and others},
journal={Nature communications},
volume={12},
number={1},
pages={1--13},
year={2021},
publisher={Nature Publishing Group}
}
An international collaborative meta-analysis on the effects of convalescent plasma on mortality outcomes in patients with COVID-19 from all available RCTs. In addition to published data, investigators from all ongoing, completed, or discontinued RCTs are invited to share their unpublished mortality data.
Protocol available on the Open Science Framework.
@article{axfors2021association,
title={Association between convalescent plasma treatment and mortality in COVID-19: a collaborative systematic review and meta-analysis of randomized clinical trials},
author={Axfors, Cathrine and Janiaud, Perrine and Schmitt, Andreas M and van’t Hooft, Janneke and Smith, Emily R and Haber, Noah A and Abayomi, Akin and Abduljalil, Manal and Abdulrahman, Abdulkarim and Acosta-Ampudia, Yeny and others},
journal={BMC infectious diseases},
volume={21},
number={1},
pages={1--23},
year={2021},
publisher={Springer}
}
Descriptive analysis of planned, ongoing or completed trials by April 9, 2020 testing any intervention to treat or prevent COVID-19, systematically identified in trial registries, preprint servers, and literature databases. A survey was conducted of all trials to assess their recruitment status up to July 6, 2020.
Janiaud P, Axfors C, van’t Hooft J, et al. The worldwide clinical trial research response to the COVID-19 pandemic - the first 100 days. F1000Research. 2020;9:1193. doi: 10.12688/f1000research.26707.2
Data and protocol available on the Open Science Framework.
@article{janiaud2020worldwide,
title={The worldwide clinical trial research response to the COVID-19 pandemic-the first 100 days},
author={Janiaud, Perrine and Axfors, Cathrine and Van't Hooft, Janneke and Saccilotto, Ramon and Agarwal, Arnav and Appenzeller-Herzog, Christian and Contopoulos-Ioannidis, Despina G and Danchev, Valentin and Dirnagl, Ulrich and Ewald, Hannah and others},
journal={F1000Research},
volume={9},
year={2020},
publisher={Faculty of 1000 Ltd}
}
Between October 5 and 15, 2020, all trial groups of randomized clinical trials registered in the first 100 days were invited to share information on recruitment status, enrollment accrual, reasons for trial termination or discontinuation (if applicable), and any results reporting. Additional publications of RCTS reporting results were systematically searched using the Living Overview of Evidence (LOVE) platform.
Janiaud P, Axfors C, Ioannidis JPA, Hemkens LG. Recruitment and Results Reporting of COVID-19 Randomized Clinical Trials Registered in the First 100 Days of the Pandemic. JAMA Netw Open. 2021 Mar 1;4(3):e210330. doi: 10.1001/jamanetworkopen.2021.0330
@article{janiaud2021recruitment,
title={Recruitment and results reporting of COVID-19 randomized clinical trials registered in the first 100 days of the pandemic},
author={Janiaud, Perrine and Axfors, Cathrine and Ioannidis, John PA and Hemkens, Lars G},
journal={JAMA network open},
volume={4},
number={3},
pages={e210330--e210330},
year={2021},
publisher={American Medical Association}
}
It is a comprehensive dataset with basic information on all randomized clinical trials to treat or prevent COVID-19 included in the database.
The WHO ICTRP and ClinicalTrials.gov are searched and imported automatically whereas the L·OVE platform is searched manually. After conversion to a standard format, all entries are assigned to a screening collection (i.e. dataset). The collections ensure transparency and reproducibility of each steps of the data flow.
All entries are screened for eligibility (automatically for the WHO ICTRP and ClinicalTrials.gov collections and manually for the L·OVE platform collection).
Following automatic de-duplication and data extraction, eligible entries are indexed in the Cove BASIC collection. The variables of entries retrieved from the WHO ICTRP and ClinicalTrials.gov are automatically updated on a weekly basis, when applicable.
Entries that are related to the same trial are linked and stored in the linked trials collection.
Entries included in COVID-evidence BASIC collection are further manually screened, verifying their eligibility. Manual extraction and data verification of automatic data extraction are done for “Expansion Modules” that go beyond the collection of basic information of trials.
In the COVID-evidence database, entries and all related information are made publicly available.
Filters provide a fast access related to users’ specific interests and in large majority are automatically populated based on information available in the database. The filters are designed to focus on trial characteristics and important clinical questions (topics). Filters on trial characteristics are based on automatic data extraction and processing.
Filters on “Vaccine” and “Post-acute COVID” topics are based on keyword searches in title and intervention field of a trial entry. Trials retrieved by the topic “Preventive non-pharmaceutical interventions (NPI)” are manually identified in our data set and restricted to registry entries. Details on filters are provided in our protocol on Open Science Framework.
Expansion modules aim to address more detailed questions that go beyond the collection of basic information of trials.
The modular concept allows targeting emerging needs for example by extracting in-depth information for only a subset of trials (e.g. trials assessing hydroxychloroquine or convalescent plasma), or for a certain population (e.g. children only), or country of origin. International collaborations and close interactions with various trial teams are the cornerstone of our expansion modules aiming at providing rapid answers to important clinical questions.
All data can be downloaded for free on our database page. Our protocol and other relevant projects can be accessed on Open Science Framework.
We aim to provide data of highest standard through duplicate extractions and by keeping our methods transparent. We focus on automatic extractions with continuously evolving filters due to the unprecedented growth of accumulating trial information. We present the current status of the data in the “status of review” column, marking automatic extractions (“automatic”), ongoing manual extraction (“in manual extraction”), and fulfilled manual extraction (“manual extraction completed”. Specific information on topics covered by expansion modules is available via specific filters.
Yes, we are in contact with several initiatives whose aims overlap with ours. For example, we have partnered with the Living OVerview of Evidence (L·OVE) platform for COVID-19 that facilitates our tracking of publications and preprints of randomized controlled trials assessing an intervention to treat or prevent COVID-19.
Examples of other efforts to summarize the literature on COVID-19 are presented here. Some overlap exists, mainly for COVID-evidence BASIC. We aim to make use of them in order to increase data reliability and work efficiency. We would like to emphasize the following COVID-evidence characteristics:
- We apply highest-standard systematic review methods.
- Beyond tracking trial registries, we integrate manual data extraction by experienced meta-researchers for specific fields with special importance.
- We have a broad definition of trials included – not only drug treatments.
- Trials are included all across their lifespan – from planned to published.
- We integrate information from different sources, including fulltexts, and our reviewer team has native speakers of several languages (e.g., English, Chinese, French, German, and Dutch).
- We have created strong international collaborations with trial teams.
For more details on our work, please visit our full protocol on Open Science Framework (OSF; DOI 10.17605/OSF.IO/GEHFX). We aim to continuously upload our updated methods and processes.
This is a non-profit initiative. The project is funded by the Swiss National Science Foundation (project 31CA30_196190). The members of the COVID-evidence core team are supported by our institutions at the University of Basel (DKF) and Stanford University.
If you are interested in contributing to this project, please send an email to lars.hemkens@usb.ch.
The source code of covid-evidence updater is MIT-licensed.