New Study Reveals NHS Struggles To Implement AI Diagnostic Tools

The £21 million AI chest diagnostics program from NHS England faced major delays, with procurement, IT challenges, and staff concerns slowing its rollout significantly.

A new study reported that NHS England’s AI chest diagnostic program has proven far more complex and time-consuming than expected.

In July 2023, the Artificial Intelligence Diagnostic Fund (AIDF) of NHS England distributed £21 million to 12 imaging networks, which served 66 NHS Trusts. The program implemented AI technology to enhance chest diagnostic capabilities while reducing patient wait times and decreasing staff workload.

The research team from University College London and their partners conducted a fast evaluation between March and September 2024 to study the acquisition and deployment readiness of AI tools. The researchers conducted 51 interviews, observed 57 planning meetings, and analyzed 166 documents to understand the entire process.

The study found that the procurement process extended beyond expectations as networks needed to form evaluation panels, conduct tender reviews, and manage supplier complaints about non-selection.

Contracts, initially expected in November 2023, were only signed months later.

The deployment also faced seatbacks as different healthcare facilities operated with incompatible IT systems, followed distinct governance protocols, and maintained varying levels of data quality.

Staff training was another hurdle. Some clinicians worried about accountability if AI missed a diagnosis. While training made clear that “AI should only act as decision support,” not replace human judgment, concerns lingered. Patient communication also varied: some networks planned posters or leaflets, while others had no clear strategy.

The research identified three key success factors, which included effective leadership, dedicated project managers, and clinical champions who supported the initiative. The networks achieved better results through their mutual exchange of professional knowledge.

The authors concluded: “Implementing AI involved complex social and technical processes, requiring significant resources.” They warned that, while AI may improve diagnostics, it will not fix NHS pressures as easily as policymakers hope.

The project was funded by the National Institute for Health and Care Research (NIHR).