Implications for Delivery
To effectively reduce serious road trauma, road safety professionals must move beyond surface-level crash causes and explore the deeper, systemic issues that contribute to them. Traditional crash investigations - especially those led by police - often focus on immediate actions leading to a crash (e.g. driver behaviour), but rarely uncover the broader, system-level factors such as poor infrastructure design, unclear responsibilities, or ineffective policy (known as "blunt-end" or latent failures). These underlying issues can significantly increase the risk or severity of collisions.
Practitioners should advocate for, and participate in, deeper crash investigations using systems-based methods like AcciMap or STAMP-CAST. These approaches help identify patterns across crashes and inform targeted, evidence-based interventions across infrastructure, policy, training, and vehicle design. Where possible, delivery teams should partner with local police, researchers, and transport authorities to support thematic investigations or sample-based studies. This could involve selecting high-risk crash types or locations for more detailed analysis. Building these insights into strategic plans and investment decisions can lead to safer systems and more effective prevention of road deaths and serious injuries.
Background
Learning from failure is one of the most important parts of implementing a system-based approach. The feedback from when things go wrong is a rich source of information on how the system can be improved.
The UK has one of the most comprehensive datasets for road crashes in the world (STATS19). However, the information in STATS19 provides a relatively shallow insight into what happened and very little information at all about why the crash happened.
For crashes where someone has died, a full, forensic police investigation is carried out and this goes into much more depth. The focus of the police investigation is usually around finding out if an offence has been committed and providing the court with evidence for either a criminal prosecution or to determine the cause of death. This can lead to a focus on what went wrong right before the crash - this is known as a sharp-end failure. These are the immediate actions or errors made by people on the road, like a driver running a red light or a pedestrian stepping into traffic. However, these actions often happen in a wider context shaped by what are called blunt-end or latent failures. These are the underlying issues built into the system - things like poor road design, unclear signage, unsafe speed limits, or a lack of safe crossing points. These system-level weaknesses can create the conditions where sharp-end failures are more likely to occur or have more severe consequences. To truly improve road safety, we need to address both: helping people make safe choices while also designing a road system that makes those choices easier and safer.
So, it is important that investigations go further to understand why crashes happen and identify latent failures as well as the immediate actions or errors involved. This is generally outside the scope of the police investigation so others, such as researchers or Highway Authority experts, will need to work alongside Police investigators to undertake additional enquiries.
In other transport modes in the UK this work is undertaken by an investigation branch, such as the Air Accident Investigation Branch or Rail Accident Investigation Branch. Some other countries have a similar body for road crash investigation, such as the American National Highway Traffic Safety Administration (NHTSA), The Dutch Safety Board (Onderzoeksraad voor Veiligheid, OVV), The Swedish Accident Investigation Authority (Statens haverikommission – SHK), and The Norwegian Safety Investigation Authority (Statens havarikommisjon – SHK). The UK has established a Road Safety Investigation Branch, but its scope is currently limited to investigating collisions involving Autonomous Vehicles.
Based on the number of road collisions, it may be more manageable to select a sample of collisions for more in-depth investigation. Using other data, such as STATS19 or the results of Fatal Accident Inquiries, themes could be identified so similar collision types are investigated in depth to understand if there are any patterns in the sharp- and blunt-end factors involved.
What the Research Shows
There are a number of projects that have undertaken in-depth investigations for a sample of collisions in the UK:
The Road Accident In-Depth Studies (RAIDS) Programme
The UK Department for Transport’s RAIDS programme investigates road collisions—not to assign blame, but to understand how injuries occur and how they can be prevented. It combines two types of detailed investigations:
On-scene studies, conducted soon after crashes, documenting vehicles, road conditions, and the crash environment.
Retrospective studies of more serious crashes, linking damaged vehicles and hospital injury data.
RAIDS collects a broad range of data including vehicle damage, road layout, and environmental conditions to injury patterns and restraint use—while protecting personal privacy. This rich dataset, which already includes over 1,200 collisions, helps identify how crashes happen and how design, vehicle design, and system interventions can reduce harm.
The Road Collision Investigation Project
In June 2018, the RAC Foundation launched the Road Collision Investigation Project (RCIP) with £480,000 in government funding—and match funding from National Highways—to pilot a more detailed, systems-based approach to investigating fatal and serious crashes in Dorset, Devon & Cornwall, Humberside, and West Midlands police force areas. Drawing inspiration from aviation, rail, and maritime accident inquiries, the RCIP aimed to determine whether investing more resources into road crash investigation would generate strong safety benefits for the public.
The project used Actor Maps and AcciMaps—a systems-based method—to capture deeper crash data including factors like standards, institutional responsibilities, and communication and coordination. The study found 60% of factors were blunt-end compared to 40% sharp-end.
The pilot demonstrated clear benefits: uncovering hidden causes, enabling no-blame thematic reviews, and guiding targeted safety interventions. This led to a strong business case for a permanent, independent Road Safety Investigation Branch (RSIB).
Using STAMP (System-Theoretic Accident Model and Processes)
In an alternative approach to the AcciMap method used in the Road Collision Investigation Project (RCIP) the STAMP-CAST method was used to examine ten fatal road crashes in Cambridgeshire and Peterborough. This found many latent failures and made recommendations across:
infrastructure design
speed management
vehicle safety
licensing and training
safety culture
maintenance
information and advice
policy and legislation
prioritisation and funding
This shows there are multiple methods that can be used to undertake more in-depth investigations.
Thematic investigations
While the projects above have taken samples across collision types to demonstrate the value of the approaches, others have taken a thematic approach. For example, Transport for London commissioned a study into fatal pedal cyclist collisions in London between 2007 and 2011 to identify common patterns and opportunities for prevention. The report highlighted the need for better street design to reduce conflict between cyclists and motor vehicles, especially at junctions. It also pointed to the importance of improving vehicle design (such as increased direct vision for drivers), training and awareness for all road users, and reviewing operational policies for large vehicle movements in urban areas. The research influenced development of the direct vision standards for large vehicles being used in London.
Sources
Jeavons, S., & Runacres, A. (2020). International review of road collision investigation approaches. RAC Foundation. https://www.racfoundation.org/wp-content/uploads/International_review_of-road_collision_investigation_approaches_Jeavons_Runacres_December_2020.pdf
RAC Foundation. (2022). Road Collision Investigation Project. https://www.racfoundation.org/collaborations/road-collision-investigation-project
Staton, M., Barnes, J., Morris, A., & Waterson, P. (2023). Learning not blaming: Investigating ten fatal road traffic collisions using STAMP-CAST. Transportation research interdisciplinary perspectives, 21, 100912. https://doi.org/10.1016/j.trip.2023.100912
Stanton, N. A., Box, E., Butler, M., Dale, M., Tomlinson, E. M., & Stanton, M. (2023). Using actor maps and AcciMaps for road safety investigations: development of taxonomies and meta-analyses. Safety science, 158, 105975. https://doi.org/10.1016/j.ssci.2022.105975
Talbot, R., Reed, S., Barnes, J., Thomas, P., & Christie, N. (2014). Pedal cyclist fatalities in London: Analysis of police collision files (2007–2011). Transport for London. https://content.tfl.gov.uk/pedal-cyclist-fatalities-in-london.pdf