Social Finance has partnered with the Wellcome Trust to deliver the Wellcome Data Prize in Mental Health.
This initiative will support collaborative approaches to research into anxiety and depression in young people. Teams based in the UK and South Africa will explore existing data to find new insights and build digital tools that enable future research.
At a glance
Applications are open from 4 April to 5 June 2022.
There are three phases to the prize, each six months long and with different levels of funding and support on offer:
- Discovery phase: 10 teams will be selected to receive £40,000 of funding
- Prototyping phase: five teams will be selected to receive £100,000 of funding
- Sustainability phase: £500,000 will be allocated across three winning teams
Visit the funding page on Wellcome’s website for more details on how to apply.
To stay updated about developments of the prize, sign up to our mailing list.
If you are interested in joining a team and wish to engage with other potential applicants join our Wellcome Data Prize Slack Group.
Is there a special rational for focusing on the UK and South Africa?
The data prize is a new funding model for Wellcome where instead of using data globally, we decided to focus on a smaller subsection of countries as a starting point. In terms of scope, we wanted to include data sources from both the Global North and the Global South. Social Finance and Wellcome conducted initial scoping around the quality and depth of existing data sets relevant to the focus of the prize, as well as the wider feasibility of a data prize in various geographies, which resulted in the decision to focus on the UK and South Africa.
Are you expecting the analysis to be in both countries or is it possible to only focus on one country?
Teams should focus on working on a data source from one specific region (the UK or South Africa) and formulate their research question based on that.
In cases where the data is free to access via the UK Data service, for example Next Steps, is any proof of access still required for the application?
In the UK, a lot of the data for data sets like Next Steps are available under the end user license. If the research can be carried out without access to any of the fields that sit in the more secure areas of the UK data service, then all that would need to be demonstrated in the proposal is that the publicly available data is sufficient to address the research question and then upload the end user license.
If the team brings their own data for the Wellcome Data Prize, what level of data sharing is required?
We understand a lot of these data sets might include personal and/or sensitive personal data that cannot be shared more widely. As a benchmark, we would expect sufficient data sharing to enable people to replicate the analysis carried out in the research or tooling. This could involve pseudonymizing or reducing the level of data while maintaining the structure of the data set, for example.
Is it possible to only apply to the discovery phase if a team has the required skills for the discovery phase but not for the prototyping phase?
The ambition of the prize is to support multi-disciplinary teams that have the potential to take a project from discovery through to prototyping as two integrated parts. We will therefore be looking for teams to apply with a view to being able to do both, linking their research to the digital tool they would produce.
Within the application form, we are asking teams for their plans for the discovery phase, alongside their prospective view of what they will focus on during the prototyping phase. During the discovery phase, we anticipate more of a focus on the groundwork in terms of the research and data analysis and teams will also be starting to think about ideating for the tool development. There will be a selection process at the end of the discovery phase which will assess what research has been done at that point and the outputs of teams’ ideation for their tool. Based on this, we will pick the 5 teams with the most promising potential to progress to prototyping.
We recognise that the skillsets necessary for both phases might not sit in one organisation, therefore we are encouraging applications from multi-disciplinary teams and collaborations across organisations.
What do you expect when referring to involving lived experience, considering that primary research is out of scope?
Engaging with lived experience might for example relate to interpretation of findings that can be nuanced with lived experience, having support from young people in planning the approach of the project or having a youth advisory board to guide on decision making. These are only examples and depending on the focus on the project, the ways teams might involve people with lived experience will vary depending on what would be most relevant and impactful for the project.
How are you proposing that the active ingredients are implemented in the first instance in order to get the right data to analyse and incorporate that in the tooling?
This is up to the researcher to define the active ingredients they propose to focus on and link these to datasets in the way that they see fit. In the datasets we are proposing for the prize, it should be possible to identify data connected to a wide range of the active ingredients that Wellcome has focused on previously. However, it is not necessary to focus on ingredients that have been highlighted previously by Wellcome and teams can propose new ingredients if these are clearly conceptually defined and in accordance with Wellcome’s overall definition of active ingredients. The way the active ingredients will relate to the tool will depend on the use case chosen by the team (here ‘use case’ refers to the combination of the research question and the digital output). You can hear about examples of use cases in the webinar recording.
Is it possible to use older British birth cohorts – e.g. 1958-1970?
We are interested in research focused on ages up to 30 years old. Whilst it might be possible to look at data on young people historically, you would need to make the case that the results would still be generalizable to today.
Is it possible to look at a range of active ingredients as an example or do they need to be defined in advance or during the discovery stage?
This would depend on the research question. We envisage that for some research questions, for example when looking at identifying a specific combination of active ingredients, teams might not know upfront the exact list of active ingredients they are going to be looking at in the data. In this case, teams would have to be able to demonstrate that the data set being proposed has sufficient breadth for their analysis and provide some suggestions of the active ingredients that they think are represented well in the longitudinal study they will be focusing on.