Is the real risk of a bridge zero compared to the current crossing?
Residents have challenged Network Rail’s assumption that if they closed The Hatches level crossing, and replaced it with a bridge at Spencer Close, the risk would be zero.
Network Rail say they do not consider the risk of replacement bridges in such assessments.
I was curious to see if I could make my own estimate, so asked Network Rail if they would share their Strategic Risk Model, or SRM. They did send it to me, and I used it to estimate the risk as explained below:
The risks associated with bridges
The Safety Risk Model (SRM), from which we can derive a quite granular breakdown of the risks of passenger slip, trip or fall in a station (which totals 26.7 FWI/year), and other risks that could be associated with a bridge within or outside a station:
| HE Code | HE Description | FWI/year |
| HEN-14B | Passenger slip, trip or fall (stairs) | 10.000 |
| HEN-14G | Passenger slip, trip or fall (ramps) | 0.569 |
| HEN-15 | Passenger fall from overbridge at a station | 0.063 |
| HEN-36 | MOP (adult trespasser) slip, trip or fall while on railway infrastructure | 0.776 |
| HEN-40 | MOP (child trespasser) slip, trip or fall while on railway infrastructure | 0.059 |
| HEN-45 | MOP (non-trespasser) fall from height onto railway infrastructure | 1.078 |
| HEN-58 | MOP (non-trespasser) injury from objects/structures (eg bridges, steps, subways) other than in stations whilst on railway infrastructure | 0.020 |
| HEN-71 | MOP (adult trespasser) fall/jump from height on railway infrastructure | 2.684 |
| HEN-72 | MOP (child trespasser) fall/jump from height on railway infrastructure | 0.157 |
| HEN-78 YDS | MOP (non-trespasser) slip, trip, or fall on YD&S site | 0.028 |
| HEN-83 | Passenger awkward movement injury | 0.170 |
| 15.61 |
Estimate of mean risk per bridge
Network Rail owns 2,400 passenger footbridges[1]. To estimate the risk per bridge would require data/estimate of the number of stairs and ramps within the 2,576 Network Rail stations that are not associated with bridges.
Then we could estimate the average risk of an average footbridge = Risks associated with stairs and ramps/Total number of footbridges and other structures with stairs and ramps + risks not associated with stairs and ramps that could be associated with footbridges/ Total number of footbridges.
This would be an over-estimate, as some of the risks that could be associated with footbridges may not be attributable to footbridges, but the relative numbers are very small.
For example, if we assumed that on average, every Network Rail station has one bridge or ramp not associated with a bridge, the mean risk for slips, falls and trips on bridges would on any single bridge would be = (10.0 + 0.569)/(2400 + 2576) + 5.04/2400 = 0.0021 + 0.0021 = 0.0042.
The following table shows how sensitive this figure is to a range of assumptions:
| Number of non-bridge stairs & ramps assumed on average at each station | HEN-14 Risk | Other Risks | Total Risk
FWI/year |
| None | 10.57/2400 | 5.04/2400 | 0.0065 |
| 1 | 10.57/(2400 + 2576) | Ditto | 0.0042 |
| 2 | 10.57/(2400 + 5152) | Ditto | 0.0035 |
| 3 | 10.57/(2400 + 7728) | Ditto | 0.0031 |
| 10 | 10.57/(2400 + 25760) | Ditto | 0.0024 |
Estimate of risk of Spencer Close bridge taking into account useage levels
The above analysis divides the total risk equally across every bridge. This would be valid if all bridges had an equal level of useage. In reality, some bridges will be much more busy than others.
1.2bn rail journeys were made in the 12 months to 30/9/22[2]. If we assume every journey-maker uses one bridge, the average number of crossings per bridge = 1.2bn/2,400 = 500,000. This compares with The Hatches census of 143,000 per annum.
Therefore it would be reasonable to weight the risk accordingly, i.e. The Hatches risk = 143,000/500,000 = 29% of the average bridge. Applying this to the table above would result in:
| Number of non-bridge stairs & ramps assumed on average at each station | Average Bridge Total Risk
FWI/year |
Risk for Spencer Close bridge @ 29% average | Current Hatches crossing risk rating | Bridge Risk as% Current Risk |
| None (as above) | 0.0065 | 0.0019 | 0.0239 | 8% |
| 1 | 0.0042 | 0.0012 | 0.0239 | 5% |
| 2 | 0.0035 | 0.0011 | 0.0239 | 5% |
| 3 | 0.0031 | 0.0010 | 0.0239 | 4% |
| 10 | 0.0024 | 0.0007 | 0.0239 | 3% |
This would lead to the conclusion that the proposed bridge would have a risk of c. 4% – 5% of the current The Hatches crossing.
Railway footbridges and suicide
It is the case that people do commit suicide from railway footbridges. Should such a risk be included in the estimate?
I have not done so.
I think it is important to distinguish between correlation and causation, and I do not believe there is any evidence that railway footbridges cause suicide.
Weaknesses in estimate
The above estimates make a number of basic assumptions. It is possible these assumptions are flawed, which why I shared this analysis with Network Rail.
What do Network rail say about this?
Network Rail have have commented that my methodology is sound, however I have over-estimated the risk by misunderstanding the SRM definitions. I’ve asked for details.
Network Rail have also said they have previously conducted analysis[3] that suggests annual risk per footbridge is 0.000268FWI, or 1% of the risk of The Hatches current crossing. They haven’t shared the details of this analysis.
I hope this is useful.
Stuart Black, February 2023
[1] https://www.ribacompetitions.com/networkrailfootbridge/requirements.html
[2] https://dataportal.orr.gov.uk/media/2167/passenger-rail-usage-jul-sep-2022.pdf
[3] Email Sam Pead (NR) – Black (SHBC) 27/1/23
Stay In Touch