How NHS is scaling data upload in Bigpicture

“It really does require input across a wide range of teams,” explains Dr Chloe Kirkby, Research Delivery Manager at the NHS’ National Pathology Imaging Co-operative (NPIC). “From AI scientists and pathologists to legal and information governance teams, everyone plays a role.”

A workflow shaped by complexity

This reflects a broader reality across the consortium. Uploading data into Bigpicture is not a single workflow, but a sequence of interconnected steps: ethical approvals, dataset identification, opt-out checks, data extraction, de-identification, quality control, clinical validation, and finally technical standardization and ingestion.

Looking back, the early stages of the Bigpicture project were defined by uncertainty. Standards were still evolving, and many processes had to be built from the ground up. “It took years of background work to get to that first upload,” says Dr Alyn Cratchley, consultant liver pathologist in the NHS and clinical lead for NPIC. “I don’t think we quite expected how long it would take.”

But that investment is now paying off. “We’ve completed a number of uploads now… we’ve found our feet,” Chloe notes. “With each upload, we’ve become more efficient and things run more smoothly.”

Collaboration as the key enabler

A key driver behind this progress is the combination of local expertise and consortium-wide collaboration. Technical teams have developed scripts to streamline workflows. Regular work package meetings enable partners to share challenges and solutions. Issues that once slowed progress can now be resolved in days. “We were all in it together, learning as we went along,” Alyn reflects. 

From operational progress to strategic impact

This collaborative model is not just operationally effective, it is foundational to Bigpicture’s long-term ambition. By contributing high-quality, well-curated datasets, partners like the NHS are helping build a resource that goes beyond individual institutions. The goal is clear: enable large-scale AI development that can improve diagnostic practice and patient outcomes. “I think it’s a really great opportunity to be part of that,” Chloe says. “Ultimately, it’s about helping to improve patient outcomes.”

For future data contributors, the advice is simple and consistent: “Just start uploading,” Chloe says. “It really does get easier.” 

As Bigpicture moves further into its operational phase, these experiences highlight a shift already underway, from building the infrastructure to actively using it. And as seen across the consortium, momentum is growing.