Editorial: Thinking and organising in systems: reframing the long problem of learning from incidents (BMJ, 9 December 2025)

Summary

Learning from safety incidents is one of the most common and widespread improvement strategies in healthcare. It is also one of the most problematic. Healthcare systems around the world expend enormous time and effort investigating large numbers of incidents, writing reports and issuing recommendations and a wide array of policies, frameworks, tools and methods surround and support these efforts.

To take just two examples: the English NHS now collects around 3 million patient safety incident reports each year; and between just 2020 and 2023, England’s Maternity and Newborn Safety Investigation body conducted around 3000 incident investigations and issued over 4620 recommendations. The remarkable scale of these activities is increasingly matched by growing frustrations at the limited return on these investigative investments: patients continue to be harmed by the same types of incidents in the same ways, while investigations find the same problems and often repeatedly issue the same recommendations.

Why is all this effort producing so little improvement? Why is learning from incidents so hard?