Open evidence-synthesis data system

We urgently need an open data system to streamline how we share and reuse evidence-synthesis data. We have drafted a ‘straw dogs’ report to engage our ‘community of communities’ about how the open data system should work.

This data system will underpin access to trustworthy, actionable insights on all of the big questions (or problems) of our time – from policy-scale, AI-enabled, living evidence syntheses – including how the insights vary by groups and contexts. Imagine every study (or evaluation) – from anywhere in the world – addressing the same question… all in one place, each with key data (e.g., about effect sizes) extracted and quality assessed, and with an overall synthesis about what was learned and with what equity considerations and confidence or caveats.

This data system will provide access to the synthesis- and AI-ready data underpinning these insights in case you want to contextualize the insights for a different group or context, or you want to adjust how studies were identified and selected or how evidence was extracted, appraised and synthesized. Imagine that we extract data once and conduct quality (e.g., risk-of-bias) assessments once, and use the data many times (and thereby avoid the extraordinary duplication happening in evidence synthesis every day).

We hope to come to agreement on and to begin implementing an agreement on an open data system sometime between April and June. We need the core elements of a workable system in place by September, by which time 11 sectoral hubs and two regional hubs will be dependent on it. We then need to continue to add elements quickly and in an initial priority order to be set by the sectoral and regional hubs. In the absence of an open data system, we will continue to waste time and money and will (likely irrevocably) cede evidence synthesis to tech solutions that do not get the balance right between rigour and speed.

We can learn from history: UK train-company owners spent much of the 1800s battling over, coming to agreement on, and slowly implementing an agreement on standardizing train track gauges. When incompatible track widths met, goods and passengers had to be moved from one train to the next, which wasted a great deal of time and money. We can and must do better with our open data system. We need to be able to share and reusing synthesis data much more efficiently.

Four key steps are important to understand in how we got here:

• a working group – operating as part of the broader ESIC planning process from January through June 2025 – articulated many of our collective aspirations for the sharing and reusing of evidence-synthesis data
• the Funder executive announced on 21 September 2025 that an open data system would be one of the five year-1 foundational investments
• the Funder executive shared on 11 October 2025 a memo outlining our proposed approach to technology-company engagement
• we convened a series of three touchpoints about the proposed open data system in October and November 2025.

The key documents prepared for or emerging from these steps can be found on the ‘how we got here’ webpage, with the most important of these documents including: 

• memo from the Funders executive that describes the next steps
• sample use cases
• existing and proposed capabilities for the open data system.

Four key processes are important to understand about our planning and engagement process in the first half of 2026:

• a seven-member planning group worked together – and with facilitation and business/legal input provided by the Wellcome Trust – from mid-January to mid-March 2026 to draft a ‘straw dogs’ report about how the open data system will work, which is now available for feedback from our ‘community of communities’
• the group hosted a community webinar on Monday 30 March and is offering bilateral and small-group discussions throughout April and May to answer additional questions, respond to additional feedback, or talk though potential implications for particular groups
• the planning group revised the report based on the very helpful feedback it received, including some very detailed and constructive feedback from several individuals, and now is posting this revised version in two formats (one with the additions/edits highlighted and a second without the highlights)
• in parallel to this cross-sectoral planning on an open data system, a minimum-viable product of a synthesis-ready evidence repository is being developed for the education sector – for launch in early May – and feeding lessons learned back into the cross-sectoral process.

Additional details, including a link to register for the webinar, can be found on the ‘planning and engagement’ webpage.