Vivli
Clinical trials data hold great potential to unlock new discovery for advancement of human health. Vivli will provide greater discoverability of data with a curated metadata catalog and a dynamic user portal. Vivli will also provide hosting services for investigators who need a way to share their clinical trials data. Over time, Vivli will adopt technologies to provide additional services, such as anonymization and data mapping, to expand types of data hosted.
Vivli will utilize natural language processing technology to identify information about participant Population, Intervention, Comparator and Outcome (PICO) elements, and will transform that information into computable, searchable fields. This enables researchers to build precise questions in the Vivli search engine and to discover the data most useful to their research.
Vivli will also provide a secure environment for data requestors to combine anonymized Individual Patient Data (IPD) from different hosts and data generators, including industry, academia, and biotech. In this controlled-access environment, strict data security models and export controls can be applied according to the specifications of each data contributor. A wide array of analytical tools will be available through the platform.
Vivli Scope:
- Host data for stakeholders that do not have the ability to do so
- Enable interoperability of data from multiple sources
- Coordinate and integrate existing data-sharing initiatives, policies and processes as appropriate
- Promote reasoned solutions to the challenges of data sharing.
Share your data on the Vivli platform
We have partnered with some of the biggest supporters of clinical trials to architect the Vivli platform to increase the impact of the trial participants. When you contribute your data on the Vivli platform, you are opening the doors for new discovery while reducing risks to human participants. If you need data hosting or are committed to sharing your collected data for others to examine, Vivli offers the most complete IPD data platform to date. Vivli will be agnostic to condition, geographic region, and generator when accepting anonymized IPD from clinical trials.
We strongly encourage, but will not require, that IPD be submitted in a standardized format (e.g. CDISC Study Data Tabulation Model [SDTM]). When you submit the anonymized IPD to Vivli, you will be asked to also submit the final study protocol, final amendments, and the statistical analysis plan. These study documents provide the building blocks of the Vivli metadata curation process. If you are submitting data in response to a regulatory, publication, or sponsor requirement, Vivli will accept both the Post-Publication data packages and the Fully Analyzable data package.
Vivli will mint a Digital Object Identifier (DOI) for submitted datasets. A DOI is a persistent identifier of your work that enables citation. By sharing your data on the Vivli platform, you are gaining visibility for your work, increasing the impact of the data derived from the trial participants, creating possibilities for new science, and generating solutions for both patients and medicine.
Contributors of data to Vivli could be individuals or individuals submitting on behalf of an entity. A data contributor agreement will be required in order to submit data to the platform. These agreements are for the protection of trial participants and data generators.
Connect with others
A major gap in data staring is the ability to combine data about conditions or solutions across many contributors. Stakeholders in clinical trials data span many audiences: industry generators, academic researchers, biotech, foundation sponsors, and individual participants. There has never been a platform that offers a solution across audiences the way that Vivli aims to do. When you share your data on the Vivli platform or make your data discoverable through the Vivli metadata curation engine, you are connecting to multiple populations who will find new value in the contributions of clinical trial participants.
Connecting platforms is only the first step. The true utility of data lies in a researcher’s ability to combine datasets from disparate generators. Vivli aspires to make data not only discoverable, but also requestable through the Vivli portal. Users will be able to file a request for multiple datasets that reside on the Vivli platform or on other platforms. Requests for data will go through one or more Independent Review Panel processes (IRP) that will result in approval or non-approval of specific access to requested data. Vivli aspires to harmonize request templates, data use agreements, and IRP processes for a seamless user experience that respects the processes of data generators who wish to retain data on their own platform. If generators agree to allow their data to move to Vivli for the purpose of research and data interdigitation, Vivli will provide a secure work environment with robust analytical tool offerings.
Vivli will also participate in and, when needed, launch initiatives to harmonize secondary use of clinical trials data. The goals of these initiatives include the creation of educational materials for researchers, the harmonization of agreements, and the development of new services. Topics for these initiatives include the following: Informed Consent Language for Secondary Data Use, Common Data Elements, Anonymization Best Practices, Risk Assessments, and academic credit for data authors/generators. Vivli will collaborate with a multi-stakeholder audience and subject matter experts to create and disseminate these guides, materials, tools, and services.
As Vivli matures, we will launch community verticals and will explore innovative differential privacy tools that will enable work on rare disease data while lowering the risk of re-identification for small populations.
Search and Discover
Vivli offers a major improvement in clinical trial data discoverability. Partnering with existing technologies, Vivli has designed an enhanced search engine to index clinical trial metadata with machine-assisted curation. The metadata for studies will be captured via application interfaces with trial registries and other repositories. However, the existing catalog metadata items that can be extracted from trial registration records do not support highly precise search and browse functionality. The final study protocol will also be uploaded to Vivli, where it will enter the secure extraction pipeline. The PICO elements from these documents will be structured into computable, searchable elements via machine extraction and natural language processing. Vivli is in development of a semi-automated extraction and annotation pipeline that leverages medical annotation, text mining, and machine learning to perform the curation. This pipeline will be combined with both human curation and a QA routine.
Using the data extraction and enhanced computability of PICO terms (described above), we have designed a dynamic query-building tool to assist researchers. Datasets will be discoverable by trial number or through text entry of a scientific question. PICO terms in the query will be identified and navigable, so as to assist the requestor in creating a more precise question. The search return visualizations will update in real time. They will show dataset attributes such as the age distribution of participants, the number of participants, the geographic region, the location at which the IPD are hosted, among others. This will facilitate the discovery of data relevant to a research question, regardless of where data are hosted.
The search functionalities will be open to the public, who may browse, discover or conduct meta-analysis. To request data, a user must create a profile and log in to Vivli to complete the request forms and data use agreements.
