AI-Augmented Mapping of the Ageing Research Landscape - Dunhill Medical Trust

About the client

Dunhill Medical Trust (DMT) is dedicated to funding research to improve the health and wellbeing of older people. They contracted me to identify ageing grants within an openly available dataset of health research grants in the UK to assist in a mapping exercise of the ageing research funding landscape.

Challenge

Due to the difficulty of clearly defining what constitutes “ageing research”, identifying relevant grants from a large open dataset required a combination of traditional search methods and innovative AI techniques. DMT needed a systematic, data-driven approach that could handle the inherent ambiguity of the field while maintaining classification accuracy.

Approach

I worked closely with the DMT team throughout the project:

• Co-developed a broad search string alongside the DMT team to retrieve ageing research grants from the open dataset, ensuring broad coverage of the field.

• Developed an automated system using a large language model to scan through the grants retrieved by the search string, reducing the false positive rate. This system achieved high agreement with expert manual classification done by the DMT team.

• Performed various analyses on the curated subset of ageing-related grants to identify key funders, organisations, and areas of work within the field.

• Utilised AI to categorise grants against DMT’s strategic priorities, providing insights into how the broader research community’s efforts align with DMT’s focus areas.

Outcomes

The results of this analysis provided DMT with a data-driven overview of the ageing research funding landscape in the UK, as a starting point to inform their strategic planning and identify potential gaps and opportunities in the field.

This project showcased the power of combining open data sources with advanced data analyses to speed up what could be very time-consuming work. The high agreement rate between the automated system and expert classifications underscores the potential of AI in streamlining research analysis processes.