Article Text
PDF
Abstract
Background To perform a cost–benefit analysis of the installation of speed cameras on the beltways of Barcelona.
Methods The analysis was performed from the society perspective over a 2-year period using a controlled before-and-after study design. The net benefit was calculated using, as benefits, the willingness to pay for the estimated number of people who avoided injury as a consequence of the intervention, subtracting costs and savings. Sensitivity analyses were performed using the maximum and minimum estimated number of people who avoided injury, assuming that there was a minimum of one death avoided, using the maximum value of a statistical life, assigning all implementation costs to the first year and assuming there was no time lost due to speed reduction.
Results Base case results showed a net benefit of €6.8 million. Sensitivity analyses suggested that net benefits could range from €5.6 to €23.1 million over 2 years.
Conclusions The use of speed cameras in urban areas has a favourable economic impact even when assessed using conservative assumptions.
- Cost benefit analyses
- traffic crashes
- public health
- economics
Statistics from Altmetric.com
Introduction
Injury is the worldwide leading cause of death in people between the ages of 1 to 45 years and accounts for more years of potential life lost before the age of 75 years than either cancer or heart disease. Motor vehicle injury is the largest single component of these losses.1 Currently, more than 1.3 million people die each year and between 20 and 50 million people have permanent disabilities as a result of road traffic accidents.2 If these trends continue, by 2030 motor vehicle traffic accidents will represent the fifth leading cause of death and the fourth largest contributor to the global burden of disease and injury.2 3 A large number of fatal and very severe injuries take place on non-urban roads. However, although motor vehicle traffic accidents in urban settings result in less severe injuries, they account for the greater burden of disease.4 In Barcelona, with an extensive urban area, there are over 10 000 motor vehicle accidents annually with more than 12 000 people injured every year and less than 50 road traffic deaths.5 In economic terms, the cost of traffic accidents in Barcelona represents more than €300 million.6
Control of speed has proved to be one of the most efficient interventions to reduce road traffic injuries, both in urban and in non-urban areas.1 7 One of the speed enforcement tools commonly used in different countries around the world is speed cameras.8 A recent systematic review found that, in the vicinity of camera sites, pre/post reductions ranged from 14% to 72% for all accidents, 8% to 46% for accidents causing injury and 40% to 45% for accidents resulting in fatalities or serious injuries. Compared with controls, the relative improvement in pre/post accident numbers resulting in any type of injury ranged from 5% to 36%.9
During the last decade, under the Municipal Road Safety Plan of Barcelona, several initiatives have been developed including public campaigns and speed, seatbelt, helmet and alcohol reduction enforcement. In particular, eight speed cameras came into operation on the beltways of Barcelona on 26 March 2003 with the aim of reducing the number of road accidents and their consequences. Evaluation of their effectiveness showed that they had a protective effect in reducing the number of crashes, people injured and vehicles involved in crashes.10 The RR of a road collision occurring on the beltway after the installation of speed cameras versus before was 0.73 (95% CI 0.63 to 0.85). This protective effect was greater during weekend periods. No differences were observed for arterial roads where no speed cameras were installed (comparison group) (RR 0.99; 95% CI 0.90 to 1.10). Attributable fraction estimates for the 2 years of the study intervention showed that 364 accidents were prevented, 507 fewer people were injured and 789 fewer vehicles were involved in crashes.10 A reanalysis with data for 4 years after the intervention confirmed these results.11
To our knowledge, only a small number of economic evaluation studies have been performed assessing the economic impact of speed cameras, including two cost–benefit analyses, and none have been performed in Spain.12–16 Our study aims to perform a cost–benefit analysis of the installation of speed cameras on the beltways of Barcelona in the period from April 2003 to March 2005 now that its effectiveness has been demonstrated.
Methods
General settings
Speed cameras were introduced on the beltways of Barcelona in 2003 and their effectiveness was studied over a 2-year period from April 2003 to March 2005. A cost–benefit analysis was performed from the society perspective over the same time period. We calculated the net cost–benefit of the intervention, taking into account the willingness to pay (WTP) for the reduction in injuries as benefits and subtracting the difference between costs and savings.
Costs included all costs related to installation and operation of the speed cameras—implementation costs, maintenance costs, police costs, photograph/ticket processing and delivery costs. We also included an estimation of the cost of time lost due to speed reduction. Savings included medical costs corresponding to injured people avoided, private property damages avoided and productivity losses avoided. Costs were expressed as a value in Euros in 2008 and updated when necessary using the corresponding consumer price index.17
Injury information sources
As mentioned above, a previous controlled before-and-after study was conducted to assess the effectiveness of the speed cameras installed on the beltways of Barcelona. Methodological details are described elsewhere.10 The RRs were derived from the adjusted models and the attributable fraction was calculated to estimate the number of accidents prevented and the reductions in the number of people injured and vehicles involved. During the first 2 years of speed camera operation it was estimated that there would be 364 fewer accidents and 507 people would have avoided injury (95% CI 179 to 902). These data were derived from the local police accident database.
The Road Injury Emergency Surveillance System provided information on the distribution of the nature of injuries of people who attend an emergency department after experiencing a motor vehicle accident.5 As it is not possible to identify the exact location of the accident from these data, we carried out a probabilistic record linkage between the police accident database and the Road Injury Emergency Surveillance System for the same years of the study period to identify those accidents which occurred on the beltways of Barcelona.18
The overall injury severity for each individual was assessed based on the Injury Severity Score (ISS, range 1–75) from the hospital emergency diagnosis19 and categorised as follows: slight injuries (ISS 1–3), moderate (ISS 4–8), severe (ISS 9–75). This procedure was carried out using the ICDMAP-90 software.20
WTP for the reduction in injuries
The record linkage showed that, in 2002, the distribution of the injuries profile of the accidents which occurred on the beltways of Barcelona was: 90.2% slight injuries, 9.4% moderate injuries and 0.4% severe injuries. We applied this distribution to the estimated number of injuries avoided and applied WTP values based on data from the Pan-European Coordinated Accident and Injury Databases (Pendant).21 In this project a UK injury cost model was adapted to the EU to evaluate the costs associated with injuries. Methodological issues on derivation of WTP values for fatalities and injury states are beyond the scope of this paper and are described in detail elsewhere.22 We set an equivalence between the injury states described in the Pendant project and the injury profile classification used in Barcelona, the minimum relative value of injuries over the value of statistical life estimated in the Pendant project (table 1) and the value of a statistical life in the road safety setting estimated for Spain updated to 2008.23
Equivalence between the injury states described in the Pendant project and the injury profile classification used in Barcelona
Costs
Implementation costs included the purchase and installation of photo radar equipment. According to the Department of Mobility of the Barcelona City Council, the total expenditure for installation of the speed cameras was €1 388 554.9 in 2003. This value was amortised over a 10-year period, corresponding to the estimated economic life of the equipment, using a 3% discount rate. An estimate for the cost over 2 years was derived by multiplying this figure by a factor of two and then updated to 2008 (table 2).
Unit costs and values used in the analysis
Maintenance costs included regular checks and repairs of the speed cameras and related equipment. According to the Department of Mobility of the Barcelona City Council, the annual costs were €107 465.2 in 2003. An estimate for the cost over 2 years was derived by multiplying this figure by a factor of two and then updated to 2008 (table 2).
Police costs referred to five police members who were dedicated to the monitoring and control of the photo radar units (personal communication). The annual police cost for operating the speed cameras was €138 904.2 in 2003 (personal communication). An estimate for the cost over 2 years was derived by multiplying this figure by a factor of two and then updated to 2008 (table 2).
The City Institute of Treasury dealt with photograph processing, data management and analysis, ticket generating and mailing expenses. The reported costs were €122 041.0 for 2003 and €424 389.5 for 2004, updated to 2008 (table 2).
The installation of speed cameras may have led to a reduction in traffic speed and an increase in travel time. To evaluate this effect we considered the number of vehicles travelling on the beltways of Barcelona over the 2 years after installation of the speed cameras (549 434 vehicles/working day), the occupancy rate of passenger vehicles in Barcelona (1.2 people/vehicle), an estimated mean traffic speed reduction of 1.8 km/h and a unit value of travel time of €22.26 updated to 2008.13 24–26
Savings
We assumed that the estimated 507 people who avoided injury would have followed a similar pattern of resource consumption (ie, ambulances, emergencies and hospitalisations) to that observed in people injured by vehicle crashes in Barcelona city in the years 2003 and 2004 registered in the Road Emergencies Surveillance System. We assumed that people who avoided injury would have been taken in an ambulance to an emergency service. In 2002, 7.8% of those who attended an emergency room because of road traffic injuries were subsequently hospitalised. We assumed the same pattern for those persons who would have had the crash on the beltway. The distribution of these hospitalisations by hospital complexity and length of stay emulated that actually observed in people injured in vehicle crashes in the stated period. To calculate the cost of acute care service we used the reimbursement tariffs for ambulances, emergencies attended and in-hospital stays of the framework agreement between insurance companies and ambulances/hospitals in Barcelona in 2008 (table 2).27
Data on private property damage were obtained from insurance companies and corresponded to €1200 per crash (table 2).28
Short-term productivity losses were calculated by dividing the regional annual average wage (€31 949.66 in 2008) obtained from the National Statistics Institute by 365 days and multiplying this value by the number of days off work on sick leave.29 Sick leave days corresponded to all in-hospital days of people who avoided injury, plus 2 days for persons who avoided injury who would not have been hospitalised (table 2).30
Sensitivity analysis
Following reporting conventions, the main results are presented in a base case.31 Additionally, sensitivity analyses were performed using alternative values: (1) the limits of the CI of the estimated number of people who avoided injury; (2) a minimum of one death avoided; (3) the maximum value of a statistical life estimated for Spain; (4) assigning all implementation costs to the first year; and (5) assuming there was no time lost due to speed reduction.
Cost–effectiveness analysis
In order to make our results comparable with those of other interventions, we also calculated the incremental cost–effectiveness ratio of implementing the intervention. We took into account the direct costs incurred over the 2 years and divided the value by the number of people who avoided injury. Also for the cost–effectiveness analysis, in the sensitivity analysis we took into account the limits of the CI of the estimated number of people who avoided injury.
Results
The results of the cost–benefit analysis showed a net benefit of €6.8 million in the base case (table 3). The cost of the intervention was around €14.5 million, which was offset by the €20.5 million of benefits and the €831 802 of savings. The biggest share of the savings was damage to property avoided and medical savings.
Base case and sensitivity analysis results
The use of different scenarios in the sensitivity analysis suggested that the net benefit could be up to €23.1 million, assuming the maximum number of injuries avoided. The least favourable scenario showed a net loss of €6.6 million corresponding to the minimum number of injuries avoided (table 3).
When a minimum of one death avoided was taken into account, the results showed a net benefit of €9.6 million. When assigning all implementation costs to the first year, the results still showed a net benefit of €5.6 million. When the maximum value of a statistical life estimated for Spain was taken into account, the results showed a net benefit of €12.9 million. Finally, assuming there was no time lost due to speed reduction, the results showed a net benefit of €19.7 million (table 3).
The results of the cost–effectiveness analysis are also shown in table 3. These estimates suggest that the cost per person who avoided injury was €28 552. The sensitivity analysis showed that this value ranged between €16 048 and €80 870.
Discussion
Our results show that the installation of speed cameras on the beltways of Barcelona has net benefits for society even under fairly conservative assumptions. The intervention was successful in reducing motor vehicle traffic accidents and the number of people who avoided injury, with net economic benefits up to €6.8 million over 2 years.
Our results are in line with previous similar studies. A Canadian study showed net benefits of $C114 million in 2001, while a UK study showed that the benefits were five times higher than the costs 1 year after the installation of speed cameras.13 15 Injury prevention interventions in general—and road safety measures in particular—usually show net benefits due to their relatively low cost and high effectiveness.32 In terms of the cost–effectiveness ratio, the results obtained (€28 552 per person who avoided injury) are comparable to those of another preventive intervention recently assessed in Barcelona.33
The maximum speed allowed on the beltways is 80 km/h (50 mph), with some stretches limited to 60 km/h (37.3 mph), and there are no traffic lights or intersections. This speed is higher than the average in the rest of the city and intuitively one might think that the injuries profile on the beltways is more complex than on other city roads. However, a detailed characterisation of injuries on the beltways of Barcelona shows a pattern similar to that reported in other studies where slight injuries (whiplash, sprains, scratches) are those with the highest share of costs because of the frequency and the long-term morbidity they represent.34–36 Nevertheless, we assumed a shorter sick leave and hospital length of stay derived from this morbidity compared with other studies.37
Furthermore, the study assessing the effectiveness of the installation of speed cameras did not predict the number of fatalities avoided because of the small number observed in the period 2001–2.10 For this reason, in the sensitivity analysis we tested this feature assuming a minimum of one death. If this number was larger, the benefits of the intervention would be greater.
In order to make our results comparable with other studies, we set an equivalence between the injury states described in a recent EU research project and the injury profile classification used in Barcelona.21 Our assumptions are fairly conservative, since the relative value of injuries over the value of statistical life is smaller than that in the original Pendant study. In fact, the values used are similar to those reported in other studies assessing the value of non-fatal injuries in relation to a fatal casualty.38
A controversial topic is the cost of lost travel time which law offenders achieve by speeding.39 To be comparable with previous studies and as conservative as possible, we decided to take this cost into account in the analysis.13 However, the yearly vehicle-distance of travel on the Barcelona beltways has increased from 2 513 482 vehicle-km in 2001 to 2 788 089 vehicle-km in 2004, showing that there could be a smoother traffic flow after the intervention.24 The sensitivity analysis performed, assuming there was no time lost due to speed reduction, showed a net benefit of €19.7 million; the results would be even more favourable if travel time is taken into account as a benefit for society as a whole rather than as a cost, according to the actual data on vehicle-km in the Barcelona beltways. Policies to avoid the costs of road congestion resulting from accidents are increasingly receiving the attention of EU members.40
Potential benefits related to fines are not included in our cost–benefit analysis because they are a transfer from the offender to the government and their inclusion could overestimate the benefits of the intervention.22 We did not calculate other benefits derived from savings in motor fuel or clean-up costs avoided.41
We were also unable to quantify the possible reduction in levels of air pollution because the locations of the traffic/urban air quality monitoring stations do not properly reflect the emissions from the beltways. However, a recent study of particulate matter <2.5 and 10 μm in diameter (PM2.5 and PM10, respectively) using data from an urban background monitoring station located at the Institute of Environmental Assessment and Water Research about 1 km from the northern beltway showed that the mean proportion of PM10 levels due to traffic sources (sum of vehicle exhaust, road dust and 70% of secondary nitrates) decreased from 44% in 2003 to 37% in 2005 while the proportion of PM2.5 levels due to traffic sources remained around 45%.42
The analysis presented in this paper shows that the installation of speed cameras generates significant economic net benefits to society. It supports the process of evidence-based policy and decision-making, undertaking economic evaluations of public health interventions in the area of traffic injuries, a field that has not attracted the attention of analysts despite its enormous impact on the population.43
What is already known on this subject
Road traffic accidents are a leading cause of death and an important contributor to the global burden of disease and injury.
Control of speed is one of the most efficient interventions to reduce road traffic injuries, both in urban and non-urban areas. The installation of speed cameras on the beltways of the city of Barcelona has been effective in reducing the numbers of accidents, people injured and vehicles involved in accidents.
What this study adds
Base case results showed a net benefit of €6.8 million, ranging from €5.6 to €23.1 million, in the sensitivity analyses.
The installation of speed cameras on the beltways of Barcelona has net benefits for society even under fairly conservative assumptions.
Acknowledgments
We are grateful for the comments of colleagues at these presentations and especially those provided by Ramon Luengo-Fernández and David Whitehurst. The authors would like to thank Mercè Navarro (Direcció de Serveis de Mobilitat, Ajuntament de Barcelona, Department of Mobility, Barcelona City Council); Antoni Rodríguez Sivera (Institut Municipal d'Hisenda, City Institute of Treasury); Elvira Torné (Consorci Sanitari de Barcelona) and Francesc Cots (Servei d'Avaluació i Epidemiologia Clínica, Hospital del Mar) for the information they supplied; Peter Lovell (King's College London) for reviewing the English version of this manuscript; and Paolo Rungo (University of A Coruña) for his insights in some of the calculations.
References
Footnotes
An earlier version of this paper was presented at the 69th Health Economics Study Group Meeting. The discussion from this meeting had a considerable impact on the final manuscript.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.