NHS England Image Reading Design Sprint: Two Weeks Inside Breast Screening Reporting

Key Takeaways

• Two‑week sprint mapped breast screening image‑reading reporting workflow.
• Created unified data model supporting reader, director, SQAS KPIs.
• Prototype Streamlit app proved data model viability and visualisation.
• Identified need for integrated reporting within new digital service.
• Cross‑disciplinary collaboration accelerated insight, reducing months of work.

Summary

dxw conducted a two‑week design sprint with NHS England to overhaul reporting for breast‑screening image readers. The team mapped the end‑to‑end workflow, built a unified data model, and defined key performance indicators for readers, unit directors, and the Screening Quality Assurance Service. A rapid prototype using Streamlit demonstrated that the new model could generate complex visual reports, prompting discussions on embedding reporting within the upcoming digital service. The sprint delivered a validated data foundation, KPI framework, and design concepts, highlighting reporting’s role in clinical quality and patient safety.

Pulse Analysis

Breast‑screening programmes rely on precise image interpretation to catch cancer early, yet the reporting infrastructure behind radiologists remains anchored in the legacy Breast Screening Information Service (BSIS). That system forces separate logins, offers limited visualisation, and hampers the ability of individual readers, unit directors, and the national Screening Quality Assurance Service (SQAS) to compare performance consistently. As NHS England pushes toward a fully digital service, the need for a modern, interoperable reporting layer has become a strategic priority for patient safety and regulatory compliance.

During the two‑week sprint, dxw’s multidisciplinary team combined service design, data science, and clinical expertise to re‑engineer the reporting ecosystem. They produced a comprehensive data model that captures every step—from image capture to arbitration—and aligned it with a KPI matrix that satisfies the distinct needs of readers, directors, and SQAS. A lightweight Streamlit prototype, built in just two days, proved the model could drive complex charts and dashboards, validating both the technical feasibility and the visual language required for actionable insights. Early stakeholder testing ensured the concepts resonated with end‑users, avoiding the pitfalls of fragmented, back‑office‑only solutions.

The sprint’s outcomes signal a shift in how the NHS will monitor breast‑screening quality. By embedding reporting directly into the forthcoming digital service, clinicians gain real‑time, meaningful feedback, while regulators obtain standardized, comparable data across regions. This approach not only streamlines workflows but also reinforces a culture of professional reflection and accountability—key drivers of diagnostic accuracy. Other health systems facing similar legacy constraints can look to this sprint as a blueprint for rapid, user‑centred transformation of critical quality‑assurance processes.