The aim of this Vivli AMR Surveillance Open Data Re-use Data Challenge, funded by Wellcome, is to promote utilization of the Vivli AMR Register to more researchers and drive advances in the AMR field. We believe that important questions in surveillance would be answered, or tools developed through this Data Challenge.
Center for Global Research Data
Data Challenge Overview
The WHO has declared that antimicrobial resistance (AMR) is one of the top 10 global public health threats facing humanity, and antimicrobial resistant infections may become the leading cause of death globally by 2050.
The Vivli AMR Register was successfully launched in June 2022 and is sharing surveillance data from GSK, Johnson & Johnson, Merck, Pfizer, Shionogi, Paratek and Venatorx, and welcomes data from other potential data contributors in pharmaceutical, diagnostic, biotech and generics companies who generate surveillance data.
The open sharing of these industry data through a single platform enables researchers to:
- Detect trends in multi-drug resistance over time
- Inform national and international policy, and antibiotic stewardship
- Allow modelling of future resistance trends
Vivli AMR Surveillance Open Data Re-use Challenge, funded by Wellcome
Our hope is that the research projects enabled by this platform will further advance understanding, inform decision-making and drive policy and behavioral changes in the medical community, by global organizations and wider society.
The key goal of this Vivli AMR Surveillance Open Data Re-use Data Challenge, funded by Wellcome, is to promote utilization of the Vivli AMR Register to more researchers and drive advances in the AMR field. We believe that important questions in surveillance would be answered, or tools developed through this Data Challenge.
This Challenge is in direct alignment with Wellcome and Vivli’s open science goals in that through this project it will:
- Maximize the reach and discoverability of research
- Maximize the potential for research to be used and have real world impact
- Facilitate collaboration, engagement and dialogue around research
- Expand the user-base of the AMR surveillance data available on the AMR Register
Here are some potential examples of how the data sets can be utilised:
- Data sets can be used to (I) model global spatiotemporal trends of AMR using susceptibility data for pathogen-antibiotic combinations provided (II) to estimate the long term economical and public health implications.
- AMR industry surveillance data can be used to (I) estimate the burden of AMR in low-middle income countries (LMICs) and (II) to compare how it relates to global trends of AMR or (III) to integrate the data sets with any national surveillance data available to get a better insight into capacity and capability gaps in LMICs.
AMR surveillance is the monitoring of changes in populations of microbes to help understand evolving patterns of resistance to anti-infectives.
Antimicrobial resistance surveillance entails the collection of clinical samples from patients with infections in hospital or in the community. These samples are then tested to determine what microbes or isolates they contain. The isolates are then exposed to a bank of different antibiotics to determine whether the microbe is susceptible to the antibiotic, i.e., the infection would be cured by the antibiotic, or whether they are resistant to the antibiotic, i.e., another antibiotic would have to be identified to treat the infection.
The pharmaceutical industry routinely collects surveillance data that could be highly valuable to support the collective global efforts to curb AMR. These industry programmes monitor the susceptibility of clinical isolates to marketed antibiotics and record pre-launch resistance levels of new products compared to antibiotic generic drugs to resistance trends. While there is no requirement for industry to run surveillance on marketed antibiotics, regulatory bodies need surveillance data before launching a new antibiotic as part of approval requirements. Research using this data, shared through the AMR Register, could guide appropriate antibiotic prescription, set up breakpoints for antibiotics, define strategies for new drug discovery and development, reveal unmet needs and allow the modelling of future trends in resistance.
This Data Challenge invites researchers from all fields, including those working in AMR, data scientists, AI/ML researchers etc., to put forward their ideas initially as an Expression of Interest (EOI).
The Data Challenge is in the form of an open question. Participants are challenged to come up with innovative insights or uses for the AMR surveillance data. Participants might combine the data available on the AMR Register with their own open datasets to address questions related to AMR. Success will be determined by innovative and original ideas backed up with robust methods.
The data are available in the form of Excel spreadsheets and contain information on minimal inhibitory concentrations (MICs), plus country of collection, infection & specimen type, year, microorganisms and antimicrobial used, plus most datasets include basic demographics such as patient’s age range and gender. The available data was collected in 85 countries over 17 years and contains nearly one million isolates.
Details of the data available can be reviewed at https://env-amrvivliorg-dcstaging.kinsta.cloud/members/research-programs/
Expressions of Interest for the Data Challenge should be submitted on the AMR Register at https://env-amrvivliorg-dcstaging.kinsta.cloud. For details of how to take part, see How to Participate.
For any questions, please read the FAQs, or ask them via the Slack channel.
- Expressions of Interest must be submitted by midnight Eastern Time on 07-May-2023
- Teams must comprise of at least 2 and a maximum of 5 team members
- Expressions of Interest will be reviewed by Vivli to ensure they meet basic high-level criteria
- Data is available in the AMR register
- Have a minimum of 2 and maximum of 5 team members
- Pose a scientific question related to AMR
Individuals who lack teammates and have interesting AMR questions but lack analytical skills or vice versa can contact Vivli by emailing amr@vivli.org and we will attempt to find a suitable team match.
- Datasets will be downloadable, and teams will have a 30-day submission period to download and work with the data, and submit their written submission document.
- Researchers may bring their own data to the challenge to supplement the data available.
- We are able to provide access to the front-end AMR Register code base to verified researchers upon request within GitHub during the data challenge. This process would enable interested researchers access to the code base and provide input to us in a secure manner. Please email us at amr@vivli.org for access.
- Teams will present their concept as a written submission document. Guidelines for the submission document may be found here. These guidelines must be followed and final results submitted within the 30-day submission period to be considered a valid submission.
- Each submission document will be scored by a judging panel.
- Finalists will be chosen for the “pitch” phase involving a live Zoom presentation. This pitch phase will enable the judges to have an interactive Q&A session and allow finalists to bring a demonstration dimension and more depth to their solutions.
- Awardees will be decided by the highest scores – if there is a tie, the judges’ decision will be final.
- At least two of the selected teams will be from LMICs. We will judge applications equitably by separating applications from HICs and LMICs, and judging within these groups separately.
Each item will be weighted equally and will be scored from 1 lowest-5 highest
- Innovation – how creative is this idea, technique or solution?
- Generalizability – how generalizable is the proposed research or solution?
- Impact – how would you rate the impact of this solution or tool on the field if successfully implemented?
- A short list of 8-10 finalists will be chosen for the “pitch” phase involving a live Zoom presentation of 5 minutes each. This pitch phase will enable the judges to have an interactive Q&A session and allow finalists to bring a demonstration dimension and more depth to their solutions.
- Awardees will be decided by the highest scores – if there is a tie, the judges’ decision will be final.
Final Awards
A total of 5 awards will be provided that are associated with monetary awards/travel
- Grand Prize Award – $20,000 (or equivalent in GBP)
- 4 awards – $10,000 (each) (or equivalent in GBP) – 2 in Innovation and 2 in Impact
- The 5 winning teams will be invited to submit an abstract of their project to ECCMID 2024. If the abstract is accepted, the successful teams will be provided with $5,000 (or equivalent in GBP) funding to be used towards travel, registration, and other related expenses for ECCMID 2024 to present their abstract.
The significant outputs of this data challenge may be tools, applications, reports, publications or software. As strong proponents of Open Science, to ensure that the outcomes of the Data Challenge can be used for the greatest impact, participating teams are expected to make their outputs openly and publicly accessible.
All final solutions would be published in an open access journal, platform or repository, such as OSF or GitHub. Teams would be best placed to identify the most appropriate repository for their outputs and would be expected to share them in line with the timeline outlined in the FAQs on the data challenge webpage. Outputs would be accompanied by documentation or metadata to allow the solutions to be interpreted and used by others.
Links would be posted on the AMR website at www.env-amrvivliorg-dcstaging.kinsta.cloud, and shared by Vivli on social media. Winning research teams would be invited to present at ECCMID 2024 if an abstract is accepted. Any researcher solutions that are created remain the intellectual property of the research team(s). Delays or restrictions on data or software sharing may be appropriate in limited cases to protect intellectual property.