Overview

Researchers are utilizing de-identified NHS data from 57 million individuals in England to develop an AI model named Foresight. This initiative aims to predict future health outcomes and enhance personalized care.

Key Objectives of the Foresight Model

• Facilitate earlier interventions in healthcare.
• Improve care planning and management.
• Identify health inequalities across different demographic groups.
• Support population-level risk analysis.

Research Insights

Lead researcher Chris Tomlinson from UCL Institute of Health Informatics emphasized the importance of quality data for AI models:

“AI models are only as good as the data on which they’re trained. Using national-scale data allows us to represent the diverse population of England, particularly for minority groups and rare diseases.”

How Foresight Works

Foresight operates similarly to large language models like ChatGPT, but instead of generating text, it analyzes:

• Hospital admissions
• Covid-19 vaccination records
• A&E attendances

It forecasts health events such as hospitalizations and new diagnoses.

Current Pilot and Future Aspirations

The pilot is currently focused on Covid-19 research, utilizing datasets from November 2018 to the end of 2023. Richard Dobson, a lead researcher, noted:

“We’d like to include richer sources of information, such as clinicians’ notes and results from blood tests and scans.”

Data Privacy and Security

Foresight operates within the NHS England Secure Data Environment (SDE), ensuring that de-identified health data is accessed securely while maintaining NHS control over the data and AI model.

Statements from Health Officials

Vin Diwakar, national director of transformation at NHS England, stated:

“AI has the potential to change how we prevent and treat diseases, especially when trained on large datasets.”

Wes Streeting, health secretary, expressed his commitment to leveraging this technology to:

• Reduce unnecessary hospital visits
• Accelerate diagnosis times
• Free up staff resources

Collaborative Support

The pilot is supported by various partners, including:

• Health Data Research UK
• The National Institute for Health and Care Research
• The British Heart Foundation Data Science Centre