Bridge safety measures for suicide prevention

September 10, 2014

by Eva-Maria Bonin and David McDaid

In England alone, there were 12,479 suicides and an estimated 121,634 non-fatal suicide attempts in the three years from 2006 to 2008. The costs of suicide to society are high, in both human and financial terms; on average, for the whole population, these are estimated at £1.45m (at 2009 prices), including intangible costs (loss of life to the individual and the pain and suffering of relatives) as well as lost output (both waged and unwaged) and police time.

Jumping from a height accounts for around 3% of completed suicides. Given high fatality rates of over 50%, the lifetime costs of completed and attempted suicides by jumping account for more than £176m per year.

In 2011 we were asked by the Department of Health to identify and analyse the costs and economic payoffs of a range of interventions in the area of mental health promotion, prevention and early intervention, and to present this information in a way that would most helpfully support NHS and other commissioners in assessing the case for investment. As part of this report, we looked at the case for bridge safety measures for suicide prevention.


Bridges provide obvious jumping sites, and the construction of safety barriers has been shown successfully to reduce suicides on particular bridges (see Beautrais 2001, Reisch and Michel 2005, Skegg and Herbison 2009). It appears that these averted suicides are not simply displaced to other, unsecured jumping sites, but whether suicide occurs by another method is difficult to analyse.

The Clifton Suspension Bridge in Bristol is one such suicide ‘hot spot’. Following the installation of a safety barrier in 1998, at a cost of £300,000 (in 2009 prices), the number of suicides reduced from an average of 8.2 per annum in the five years before the barrier, to 4 per annum in the five years after it was installed.


Using the Clifton Suspension Bridge as a case study, our model estimated the savings (both tangible and intangible) to society of installing a safety barrier. It assumed that the barrier prevents around half of suicide attempts, but also considered the impact if these individuals instead attempt suicide using other methods. This displacement can still lead to a lower number of suicides, as the mortality of those who jump from this bridge is 95%, compared with around 9% for other suicide methods combined. The model included the probability of subsequent attempted and fatal suicides.

The cost savings were calculated first for a 1-year cohort of those attempting suicide from the bridge in a single year, and followed this group over a 10-year period. It then looked at aggregated savings from ten consecutive cohorts, assuming that the pattern of suicides would have recurred every year. The savings did not include the costs of bereavement support, or the impact on children losing a parent. It is assumed that barrier construction costs were incurred in the first year.

The results show that investment to prevent individuals from attempting suicide using high-fatality methods were likely to be cost saving, even if all the averted attempts are diverted to other suicide methods.

Pay-offs following installation of the Clifton Suspension Bridge safety barrier (2009 prices)
Pay-offs following installation of the Clifton Suspension Bridge safety barrier (2009 prices)

Key points

  • Investment in a barrier to prevent suicide jumping from a particular bridge can generate substantial financial benefits, even if suicides are displaced to other, less lethal, methods.
  • Such savings would potentially also apply to other suicide “hot spots”, including alternative jumping sites, and other high fatality suicide methods.

This post was first printed as a chapter in Knapp M, McDaid D, Parsonage M (eds) (2011) Mental Health Promotion and Mental Illness Prevention: The Economics Case, Department of Health, London.

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.