Have any questions?
Got a question? We’re here to answer! If you don’t see your question here, drop us a line on our Contact Page.
FAQs are updated as of 20 June 2024.
- I want to know what data is available on TRUST To browse TRUST data catalogue (detailed data elements, formats, period etc.), researchers have to register to be a TRUST Member on TRUST website.
- What is required to be TRUST member? Who will approve? To be TRUST member, the individual must meet all of the following:
- Employees of institutions that have signed the Data Request Agreement* with TRUST.
- Bona fide researcher (verify through pubmed ref or ORCID ID, CV or institution profile page).
- Verified institution email account.
TRUST will verify the information, approve and provide member account details within 5 working days.
*Data Request Agreement – Overarching agreement between TRUST and Institutions to allow its researchers to request access to TRUST Data. Prevents the need for multiple agreements to be signed. Once signed, TRUST Member Account will be generated for each request to view the detailed data catalogue.
For more information, please see link
- I want to request for access to TRUST data. Only TRUST members can submit a data request. TRUST members can retrieve the data request form on the TRUST website and submit the completed application to TRUST DAC secretariat. All requests are subjected to TRUST DAC discretion and approval.
- I am not familiar with TRUST data and need help to identify the right data for my research. TRUST Data Concierge will support TRUST users throughout their journey. Do reach out to the TRUST Data Concierge for assistance. They will be able to advise you on the data elements. TRUST Data Concierge contact will be available in the TRUST member portal.
- What do I need to take note of when submitting data request?
Do ensure the following:
- All fields in the data request form are completed.
- IRB approval has been obtained for the research proposed and included use of the data requested.
- TRUST has noted that for IRB approvals, some institutions have adopted a “Review Not Required” process. This could be an institution endorsed form indicating that IRB is not required, or institutional IRBs issuing this as an outcome from IRB review.
From 26 January 2024, TRUST will accept both 1) institution-endorsed IRB “Review Not Required” form or 2) “Review Not Required” as an outcome from IRB review. Do submit the relevant endorsed documentation in your submission.
Researchers are encouraged to approach the institutional research office and / or IRB for further clarification about the utility and implementation of this practice.
- If you are bringing in your own cohort data for linkage to TRUST data (e.g. clinical data), consent must be obtained from the participants to access and link their data to the requested datasets.
- If you are requesting for TRUST real-world data (e.g. clinical data), articulate the public benefit of your research either directly to the individuals or indirectly through the interest of the Singapore public sector.
- All Data Requests will require endorsement by their institutions.
- If you are collaborating with Government Entities (GEs) for the proposed research, please seek their endorsement.
- Personal undertaking has been signed by all analysts who will be accessing the data when request is approved.
- After I submit the request, when can I get access to the data? All data requests will be consolidated on the last Friday of each month for review in the following month. Any requests received after this date will be reviewed the month thereafter. Requests will be assessed and reviewed by TRUST DAC within 4-6 weeks.
- 4 weeks for request review.
- 2 weeks for data preparation & provision as well as user account provision.
Data would be provisioned on TRUST for access 6-10 weeks post approval. Note that data provisioning could be delayed by complexity of the data requested and the data availability from the data contributor.
- I want to bring my own codes and libraries to analyse the data. After data request has been approved, TRUST data concierge will inform you and ask for library requirements. Please inform TRUST data concierge of your request. All codes and libraries are subjected to TRUST approval.
- I want to export the insights after my analysis on TRUST platform. TRUST data users will only be able to export aggregated research insights from TRUST. All exports of insights are subjected to TRUST Data Concierge’s approval.
Please refer to the Guide to quality TDR submissions for more details. The Guide is available under the ‘Request TRUST Data’ tab in the TRUST Member’s Portal.
When a dataset has been approved for contribution to the TRUST platform, TRUST will –
- Provide a script to de-identify and anonymise data at the data contributor’s source environment before transferring data to TRUST.
- Combine the required datasets according to the scope of the approved project.
- Enable the use and access to the datasets within TRUST’s terms of use.
- Be accountable for the protection of datasets that TRUST processes when the data is within TRUST platform.
Note that TRUST is not a data repository, and TRUST will not store datasets perpetually. All datasets remain in its source environment (i.e. with the Data Contributor) and will only be made available to a data user upon approval of a project by the TRUST Data Access Committee (DAC).
For more information, please see link
The overview of the agreements for users to access TRUST is shown in Table 1.
Table 1: User agreements overview
Type | Purpose | |||
---|---|---|---|---|
Data Request Agreement (DRA) | Overarching agreement between TRUST and Institutions to allow its researchers to request access to “members’ area” on TRUST website view the detailed data catalogue and make data request and eventual access to TRUST platform for research analysis. | |||
Documents within the DRA |
|
TRUST is a data intermediary, which aims to facilitate secure data linkages for health analytics. As such, TRUST makes available the datasets based on approved data requests. Depending on the request and needs, data could include population-related datasets (e.g. demographic data), lifestyle (e.g. wearable data), chronic disease screening data, clinical data (e.g. diagnosis data, medication data, and healthcare financing data). In addition, TRUST is also seeking to facilitate access to strategic population and disease research datasets (e.g. genomic data, phenotypic data). We welcome feedback on datasets that can be beneficial to TRUST users.
Note that certain more sensitive datasets can only be accessed via a Micro-Access Laboratory (MAL) that is certified by MOH.
To browse TRUST data catalogue, researchers have to first register to be a TRUST Member on TRUST website. To facilitate the discovery of TRUST data, TRUST has a searchable inventory along with a data dictionary of databases that are available for access. There are data concierges and analytics support to assist approved researchers on their TRUST user journey.
TRUST does not seek to own any Contributor’s datasets, nor will it seek a share of FIP arising from the research projects. Under the agreement with the Data Contributor, all rights, interest (including IP) over the Contributor’s datasets shall be owned or remained owned by the Data Contributor. Fused datasets shall be owned by TRUST, which TRUST is accountable for when data is within TRUST environment and control.
TRUST only shares what is permissible as stated in the declaration form submitted by the Data Contributor. TRUST does not review the collaboration agreement / separate agreement of the Data Contributor and Data Requestor.
If the request involves a public institution, where necessary, TRUST will obtain opinion(s) or input(s) from Data Contributors in connection with the Data Requestor’s Research Project. If the request involves a private organisation, TRUST will direct the Data Requestor to the Data Contributor, for the purposes of establishing the necessary collaboration agreement.
The purpose of the exclusivity period is to allow the Data Contributors to have exclusive access to their datasets first, before allowing them to be used for sharing with other requestors (other than the Data Contributor’s collaborators) via TRUST. After the exclusivity period(s) expires, the Data Contributor grants a licence to TRUST to use and share the datasets (with other requestors) for the Permitted Purposes. The months of exclusivity periods are aligned with NMRC provisions but could be subject to variation depending on the nature of funded project / datasets. There is no exclusivity period for population-related data from public agencies. Exclusivity period only applies to research datasets.
TRUST values the close partnership with the research community. Should a Data Contributor wish to terminate the agreement and stop sharing their datasets with the wider research community, TRUST will work with the Contributor to understand their concerns. The reasons for withdrawal must be reasonable. Any on-going / approved projects on TRUST that involve that particular dataset would not be affected by any termination of agreement with the Contributor.
During exclusivity period, intellectual property would be determined by the agreement between Data Contributor and Data Requestor.
After the exclusivity period ends and if the Data Contributor is a collaborator in the Research Project, it would be attributed in accordance with the terms agreed by the Data Contributor and the Data Requestor in its own collaboration agreement. After exclusivity period ends and if the Data Contributor is not a collaborator in the Research Project, foreground intellectual property would be vested in the Data Requestor.
Data Users must provide TRUST data concierge with a copy of the proposed publication (such as, news items, articles, publications, advertisements, prepared speeches) which relate to derived output / insights, and/or which are in connection with the research purpose(s) ; and give TRUST twenty-one (21) working days to respond before the public release of any information.
Datasets on TRUST are shared based on permission granted and anonymised in compliance with the PDPA and Government IM8 measures. TRUST will not seek consent directly from participants and will only collect, use or disclose data with consent that has been obtained. Data Contributors must be able to provide to the TRUST platform the relevant IRB approval (or relevant waiver), including the scope and the purposes under which the permission was provided.
To enable the clinical data on TRUST to be accessible and usable, TRUST uses the OMOP common data model (CDM) as the data standard. OMOP CDM is commonly adopted both locally and internationally and for the following reasons –
- Data Protection: The data will be de-identified to protect privacy.
- Standardised Vocabularies: Contains all necessary and appropriate corresponding standard healthcare concepts. However, focus is on clinical data, with limited coverage for financial and ops data.
- Open source: All the source codes are available online and customisation can be done on top of the source codes.
- Technology neutrality: The CDM does not require a specific technology.
- Scalability: The CDM is optimise for data processing and computational analysis to accommodate data sources that vary in size, including databases with up to hundreds of millions of persons and billions of clinical observations.
- Backwards compatibility: Older versions of the CDM can be easily created from the current version, and no information is lost that was present previously.
- Adoption: OMOP is widely adopted and it will reduce effort for fusing data across contributors who are already using OMOP.
For genomic data, TRUST will review the best format for maximum utility.
Currently, TRUST will only support CSV, Hail/Parquet (genomic). The reason for limiting the file formats at this point is that the de-identification tool has been developed using these widely used formats. If there are other formats that needs to be catered for, data contributors can highlight this to TRUST for assessment and work together towards a feasible timeline for enhancement if necessary.
Each dataset file size is recommended to be <50GB (if necessary, Users are requested to split the dataset to meet this requirement). Data sharing is based on the research project and not shared in real-time.
Analytical tools that would be available on TRUST include Apache Spark, Hail, Jupyter notebook and selected GWAS tools. Although TRUST would not be able to accommodate bespoke tools on the platform, users can share their requirements and justification with TRUST. TRUST would then consider any security and infrastructure capacity in its review of the request.
We are reviewing the charging model for TRUST and will make the information available when it is ready.