Collision analysis in a transport assessment, section by section
The collision chapter of a Transport Assessment has to show whether a safety problem exists and whether the development worsens it. Here is how to obtain the data, define the study area and read severity figures without tripping over the 2016 reporting change.
The collision analysis in a Transport Assessment answers one question: is there an existing highway safety problem at or near the site access, and would the proposed development make it worse. Everything else in the section is evidence for that answer. Officers and highway engineers read it looking for a pattern, not a list, and a chapter that tabulates every recorded injury over five years without interpreting it is the most common version of this section to attract a holding objection.
Recorded injury collision data comes from police records collected under the national STATS19 system. You obtain it from the local highway authority's road safety team, or from the Department for Transport's published road safety statistics, and you agree the extent and the period with the highway authority at scoping. Get that agreement in writing before the analysis starts.
Obtaining the data and the period covered
Five full calendar years is the working convention, with the most recent year added if the data has been validated. Shorter periods hide patterns. Longer periods pull in road layouts that no longer exist, which is the more damaging error, because a highway engineer will know the junction was rebuilt in the intervening time and will discount the whole table.
Two points to settle with the authority before you pay for the extract:
- the study area, expressed as named junctions and link extents rather than a radius
- whether the authority wants raw records, or its own analysis, or both
Authorities differ on charging. Some supply extracts free to applicants, some charge an administrative fee, and some direct you to published datasets. Ask early, because a two week wait for data is a two week delay to the whole Transport Assessment.
Defining the study area defensibly
The study area follows the development's traffic, not a circle drawn on a plan. In practice it covers the site access and its immediate approaches, every junction where assigned development traffic materially changes a turning movement, and the links between them. The convention many authorities still apply is that a link enters the assessment where development traffic adds more than a small percentage of two-way flows, with a lower trigger inside urban areas and near schools. That percentage is a scoping agreement, not a rule in national policy, so record it as such.
Walking and cycling routes to local facilities belong in the study area too. Where a scheme relies on an active travel route to justify its trip rates, the collision record on that route is relevant and its omission is noticeable.
Patterns rather than a list of incidents
The analytical step is grouping. Take the records and sort them by location, then by manoeuvre, then by casualty class, then by conditions. What you are looking for is repetition: three or more collisions of the same type at the same point, over the period, is a cluster worth explaining.
Useful groupings in practice:
- Location clusters, mapped as a collision plan with a numbered key.
- Manoeuvre type, for example right turn into a side road, rear end on the approach, loss of control on a bend.
- Casualty class, separating pedestrians, pedal cyclists and powered two wheelers from car occupants.
- Time and light conditions, which often explain a cluster better than geometry does.
- Contributory factors recorded by the attending officer, treated as indicative rather than conclusive.
Where there is no pattern, say so plainly and show the working. A sparse, scattered record across five years on a busy distributor road is a normal background rate, and stating that with the data behind it is stronger than silence.
Severity classifications and a trap in the numbers
STATS19 records severity as fatal, serious or slight, based on the most seriously injured casualty. There is a reporting discontinuity to handle. From 2016 onwards police forces progressively adopted injury based severity reporting systems, which changed how injuries are classified and increased the number recorded as serious. The Department for Transport publishes adjusted figures for this reason.
If your five year window straddles the change for that police force, do not present a trend line showing serious collisions rising. Either use the adjusted series, or split the analysis at the changeover and say why. A highway engineer who spots an unadjusted trend claim loses confidence in the rest of the chapter, and rightly.
Damage only collisions are not in STATS19 at all. Residents will raise them at consultation, and the honest answer is that they are not recorded nationally, so the assessment cannot test them. Where the authority holds local complaint records, ask for them.
Linking the record to the proposed access
This is the part that decides the section. Set the pattern you found against the change the development makes.
A cluster of right turn collisions at the junction your traffic will use for 70 per cent of its trips is directly relevant, and the scheme has to respond. A cluster of loss of control collisions on a bend 400 metres away, unaffected by the development's turning movements, is context. Say which is which.
Then test visibility. Measure the achievable sight lines at the access against the stopping sight distances in Manual for Streets or the design standard the authority applies, using recorded speeds rather than the posted limit. A speed survey costs little compared with an access refused on visibility grounds. Where sight lines fall short, the options are relocating the access, removing the obstruction by agreement, or reducing approach speeds through works secured by a highway agreement under section 278 of the Highways Act 1980.
Mitigation that answers the pattern
Mitigation has to be specific to the collision type you identified. Generic offers of a contribution to road safety education read as an admission that the analysis found nothing to fix.
Responses that work, matched to pattern:
- right turn collisions: a ghost island or a right turn lane, or access relocation
- pedestrian collisions near a crossing desire line: a formal crossing, dropped kerbs, tactile paving
- cyclist collisions at a side road: continuous footway treatment or a set back give way
- rear end collisions on the approach: improved forward visibility, signing, surfacing
Larger accesses attract a Stage 1 Road Safety Audit under the relevant national design standard before the highway authority will agree the layout. That audit is a separate exercise carried out by an independent team and is not one of PlanScope's report types, so budget for it separately and programme it before the application rather than after.
Presenting it to the highway authority
Give the section a fixed shape: data source and period, study area with a plan, the tabulated schedule, the collision plan, the pattern analysis, the conclusion on existing safety, the effect of the development, and the mitigation. Anonymise everything. Records supplied by an authority can contain enough detail to identify individuals, so strip dates of birth, exact addresses and any narrative before the report is published on the planning register.
The policy hook is the long standing test carried through successive editions of the National Planning Policy Framework, including the current August 2026 edition, that development should only be prevented or refused on highways grounds where there would be an unacceptable impact on highway safety, or where the residual cumulative impacts on the road network would be severe. Your collision chapter exists to show which side of that line the scheme sits on.
Where the scheme is small enough that a full assessment is not proportionate, the same analysis appears in shorter form inside a Transport Statement, and the trip generation and mode share work carries through into a Travel Plan if one is required by condition. Construction phase movements are handled separately in a Construction Traffic Management Plan.
Draft the structure of your Transport Assessment on PlanScope, then hand it to a transport planner with the collision extract and the speed survey to complete and sign off.
PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval. For specialist technical reports, qualified professional input may be required.
Frequently asked questions
How many years of collision data should a Transport Assessment cover?
Five full calendar years is the working convention, agreed with the highway authority at scoping. Shorter periods hide clusters. Longer periods drag in junction layouts that have since been rebuilt, which undermines the whole table once an engineer recognises the change. Record the agreed period and study area in writing before commissioning the extract.
Where does personal injury collision data come from?
Police records collected under the national STATS19 reporting system. Applicants usually obtain an extract from the local highway authority's road safety team, and the Department for Transport publishes the national road safety statistics built from the same returns. Charging practice varies by authority, so ask early because the wait can delay the assessment by weeks.
Why did recorded serious injuries rise after 2016?
Police forces progressively adopted injury based severity reporting systems from 2016, which changed how injuries are classified and raised the number recorded as serious. The Department for Transport publishes adjusted series for comparison. If your study period straddles the change for that force, use the adjusted figures or split the analysis, and explain why.
Are damage only collisions included?
No. STATS19 records injury collisions only, so knocks with no casualty never enter the dataset. Residents often raise them at consultation. The honest response is that they are not nationally recorded and cannot be tested, while asking the authority whether it holds local complaint or near miss records that can be considered alongside the formal data.
Does a collision analysis replace a Road Safety Audit?
No. The collision analysis is planning evidence about existing safety and the effect of the development. A Stage 1 Road Safety Audit is an independent check of a proposed highway layout against national design standards, carried out by a separate audit team. Larger accesses need both, and the audit is not one of PlanScope's report types.
Sources and further reading
This article is general guidance about the planning system in England and, where stated, the rest of the UK. It is not legal or professional advice. Check your local planning authority’s validation requirements and take professional advice for your specific site.