Northern Ireland collision blackspots: what the 2025 figures can — and cannot — tell us
Belfast recorded Northern Ireland’s highest number of people killed or seriously injured in 2025, but that does not make a particular Belfast junction the province’s “worst blackspot”. The full picture requires careful separation of fatality counts, serious injuries, geography and traffic exposure.
The numbers behind Northern Ireland’s road safety picture
PSNI figures for calendar year 2025 record 5,015 injury collisions and 7,879 casualties: 56 deaths, 1,066 serious injuries and 6,757 slight injuries. The 1,122 people killed or seriously injured (KSI) represented an 11.3% increase on 2024 even though fatalities alone fell from 69 to 56. The correct story is therefore not simply that the roads became safer. Fewer people died, but more people sustained serious injuries.
An injury collision is a reported incident, while a casualty is a person hurt in an incident. A collision involving several people can create several casualties. That is why the total number of casualties is higher than the collision count. Published figures relate to police-recorded injury incidents; they do not capture every near miss or all damage-only crashes.
The 2025 calendar-year figures must not be confused with a 2025/26 financial-year bulletin. Both sets are legitimate PSNI publications, but their dates differ. Combining totals from those two periods in one league table would mislead readers.
Belfast recorded the highest KSI total, while Mid Ulster had the most deaths
Belfast City recorded 180 KSI casualties in 2025, more than any other local government district. Mid Ulster recorded nine road fatalities, the highest fatality count among the 11 districts. Derry City and Strabane recorded no road deaths during that calendar year. Those are different measures: a council-area serious-casualty total cannot be compared directly with another area’s fatality count.
Large urban areas contain more intersections, pedestrians, cyclists, bus corridors and daily journeys. A raw count reflects exposure as well as potential danger. Without a reliable estimate of traffic using each section of road, it is impossible to establish whether an individual trip in Belfast is riskier than a similar trip somewhere else.
Even within one district, the map matters. A serious crash assigned to Belfast City does not identify whether it occurred at an arterial junction, on a residential street or in the city centre. Naming a particular junction as a blackspot would require location-level analysis that the summary release does not perform.
Young people, pedestrians and children: who appears in the figures?
People aged 16 to 24 accounted for 227 KSI casualties, or 20.2% of the Northern Ireland total. Drivers represented 41.8% of KSI casualties, while pedestrians represented 19.0%, despite accounting for 7.7% of all casualties. The mismatch illustrates how the severity mix varies by road user; it does not establish the probability of a crash for any individual pedestrian.
There were 115 children under the age of 16 killed or seriously injured in 2025, 22 more than in 2024. The PSNI road safety partnership describes that as a 24% rise. Children appear in the statistics as passengers, pedestrians, cyclists and other road users, so a single school-road explanation would be too simplistic.
A headline about “dangerous roads” should recognise both where collisions happen and who suffers the worst consequences. A busy street with repeated serious pedestrian casualties presents a different policy challenge from a rural route where high-energy collisions are less frequent but severe.
What the police record as leading causation factors
The most common recorded principal causation factor for KSI casualties was inattention or diverted attention, associated with 115 people killed or seriously injured. Wrong course or position was associated with 95, and excessive speed having regard to conditions with 86. Those classifications are based on police recording procedures; they are not a complete list of all influences on every crash.
“Excessive speed having regard to conditions” also has a specific meaning: it concerns a speed inappropriate for the road or conditions, not necessarily proof that the posted legal speed limit was exceeded. This distinction matters on bends, in poor visibility and where pedestrians or stationary vehicles change the driving environment.
Figures on principal factors should not be added together and described as the number of collisions caused by a particular behaviour. The published presentation relates to casualties associated with the police-coded principal factor, not a complete independent count of incidents attributable to each behaviour.
What would it take to identify a true collision blackspot?
The most credible next investigation would start with PSNI collision-location records and a transparent definition of the area being measured: for example, crashes within an explicitly defined distance of a named junction. Analysts would need consistent geographic coordinates, several years of observations, severity categories and checks for duplicate location labels.
A ranking based on all injury collisions could differ from one based on KSI casualties. Equally, a junction with ten slight-injury collisions and a junction with two fatal collisions cannot fairly be compared through one undifferentiated number. A weighted severity approach can help, but its assumptions must be published.
Traffic volume is the other missing denominator. Major routes can show high collision counts simply because more vehicles use them. A rate per vehicle movement, where reliable traffic counts exist, provides important context. Without it, a list is better described as “locations with the highest recorded totals” than “the most dangerous junctions”.
Changes in junction layouts, temporary works or enforcement can also make historical clusters less predictive of current conditions. A ranking must state its time period, methodology and publication date rather than implying that older collision concentrations describe the road exactly as it operates today.
What motorists should take away from the evidence
The data do not single out one universally hazardous Northern Ireland road. They do show that serious casualties rose during 2025, that pedestrians faced a substantial severity burden and that attention and road positioning featured prominently in recorded serious collisions.
For a reader preparing a journey, live roadwork restrictions, weather, visibility, junction approach speed and the presence of vulnerable users are all more immediate considerations than a sensationalised “worst road” label. Drivers should use the conditions they actually encounter rather than assume a route with fewer published crashes needs less care.
PureCar will treat this as a statistical analysis, not a verified junction-by-junction ranking. A separate blackspot ranking should only be published after a reproducible analysis of the detailed location records. That is a more credible approach than attributing invented collision totals to named roads.
Sources and methodology
The information above is based on the original official publications linked below. Figures relate to the stated reporting periods; proposed changes are distinguished from rules confirmed as operative.
