Road traffic crashes, injury and fatality trends in Sri Lanka: 1938–2013

Abstract Objective To analyse trends in road traffic crashes, injuries and fatalities over 75 years in Sri Lanka. Methods Data on road traffic crashes, injuries and fatalities between 1938 and 2013 were obtained from the Police Statistics Unit. Rates per 100 000 population were calculated and trends were analysed using joinpoint regression analysis. Findings Road traffic crashes and injuries rose substantially between 1938 and 2013: annual crashes increased from 61.2 to 183.6 per 100 000 people; injuries, from 35.1 to 98.6 per 100 000; and fatalities, from 3.0 to 10.8 per 100 000 people per year. Joinpoint analysis showed large fluctuations in crashes and injuries over time but the fatalities rose almost continuously. These fluctuations paralleled the country’s political and economic development. In some years, better traffic law enforcement and improved public transportation may have been associated with reduced crashes and injuries, whereas rapid growth in vehicle numbers, especially two- and three-wheeled vehicles, may have contributed to increased crashes and injuries. In addition, insurance policies that did not require a police report to claim may have led to underreporting of crashes and allowed drivers to avoid prosecution. Conclusion Fluctuations over time in road traffic crashes and injuries in Sri Lanka are associated with changes in political, economic and traffic policy. There is potential for reducing road traffic crashes and injuries through better traffic law enforcement, restrictions on the importation of two- and three-wheeled vehicles and policies to improve road safety and prevent underreporting of crashes.


Introduction
Road traffic injuries are a major but neglected global public health problem. 1 Each year, road traffic crashes are responsible for over 1 million deaths and 20 to 50 million injuries worldwide. 1,2 Low-and middle-income countries are the most affected, because road traffic crashes and injuries are linked not only to the number of vehicles, road conditions and drivers' behaviour but also to the country's level of economic and social development. [1][2][3] In particular, poor road infrastructure, inappropriate mixing of vehicle types, inadequate traffic law enforcement and delayed implementation of road safety policies can increase road traffic crashes. 3 Sri Lanka is a lower-middle-income country in south Asia that has a substantial burden of road traffic injuries and fatalities. [4][5][6][7] Between 1938 and 1997, the absolute number of road traffic fatalities in the country increased 10-fold to reach 1835 deaths in 1997 in a population of around 18 million. 5,7 Despite the need for immediate action to reduce this growing burden, there is a paucity of coordinated government road safety strategies and road safety research in the country. 7 Sri Lanka has an interesting traffic history. Between 1815 and 1948, the island was governed by the United Kingdom of Great Britain and Northern Ireland. The British expanded the road infrastructure, 8 increased the number of vehicles and developed railways to transport goods. 9 In 1948, Sri Lanka gained its independence and the new government further improved the road infrastructure. However, because railways were neglected, road transport was used for goods, increasing the risk of crashes. Sri Lanka passed its first traffic act in 1934 and, from 1938 onwards, the police documented traffic accidents. 10,11 In 1951, the country mandated that all motor vehicles be registered with the Department of Motor Traffic, 8,12 which meant that vehicles had to be roadworthy. The law spe-cifically prohibited the registration of unlawfully fabricated motor vehicles, which were common at the time. In 1953, Sri Lanka established a separate traffic police division. 10,11 This increased the number of traffic police and improved traffic law enforcement.
In 1977, Sri Lanka introduced an open economic policy that promoted motor vehicle imports. This resulted in a massive influx of motorcycles and three-wheeled taxis, 8 which are prone to crashes because they are unstable and topple easily. Since two-and three-wheeled vehicles are less robust than other vehicles, occupants are more likely to be injured in a crash. [1][2][3] Finally, despite the large increase in motor vehicles, the road infrastructure did not develop at the same pace.
In Sri Lanka, a lack of road safety research and the limited availability of statistics on road traffic crashes and injuries made it difficult for policy-makers to propose interventions that would prevent road traffic crashes. The aims of this study were to describe the trends in road traffic crashes, injuries and fatalities in Sri Lanka from 1938 to 2013 and to identify factors associated with these trends.

Methods
Our analysis used statistics on road traffic crashes and road traffic injuries from the road traffic crash statistics' database maintained by the traffic police headquarters, which is the only comprehensive such database in Sri Lanka. 9 Permission to access these data was obtained from the police headquarters. By law, all road traffic crashes must be reported to the police within 24 hours. For the database, a road traffic crash was defined as a crash on a public highway or road that involved a vehicle and also involved personal injury or damage to property. 1,8 Crashes were classified as involving one of four types of injury: (i) fatal (i.e. a victim died due to injuries Objective To analyse trends in road traffic crashes, injuries and fatalities over 75 years in Sri Lanka. Methods Data on road traffic crashes, injuries and fatalities between 1938 and 2013 were obtained from the Police Statistics Unit. Rates per 100 000 population were calculated and trends were analysed using joinpoint regression analysis. Findings Road traffic crashes and injuries rose substantially between 1938 and 2013: annual crashes increased from 61.2 to 183.6 per 100 000 people; injuries, from 35.1 to 98.6 per 100 000; and fatalities, from 3.0 to 10.8 per 100 000 people per year. Joinpoint analysis showed large fluctuations in crashes and injuries over time but the fatalities rose almost continuously. These fluctuations paralleled the country's political and economic development. In some years, better traffic law enforcement and improved public transportation may have been associated with reduced crashes and injuries, whereas rapid growth in vehicle numbers, especially two-and three-wheeled vehicles, may have contributed to increased crashes and injuries. In addition, insurance policies that did not require a police report to claim may have led to underreporting of crashes and allowed drivers to avoid prosecution. Conclusion Fluctuations over time in road traffic crashes and injuries in Sri Lanka are associated with changes in political, economic and traffic policy. There is potential for reducing road traffic crashes and injuries through better traffic law enforcement, restrictions on the importation of two-and three-wheeled vehicles and policies to improve road safety and prevent underreporting of crashes.

Research
Road traffic injury trends, Sri Lanka Samath D Dharmaratne et al.
sustained in the crash, irrespective of the time interval between the crash and death); (ii) serious (i.e. the crash resulted in one or more kinds of severe injury, such as bone fractures, damage to internal organs, severe burns, permanent impairment of vision or hearing or serious disfigurement); (iii) minor; or (iv) none (i.e. the crash did not cause any injury and resulted in only damage to vehicles). 9 If more than one type of injury was present, the most serious type of injury was recorded.
Data were retrieved manually from the traffic police's database and entered into an Excel spreadsheet (Microsoft, Redmond, United States of America) for analysis. Information on the size of the population of Sri Lanka in the middle of each year was obtained from the Department of Census and Statistics 13 and used to calculate road traffic fatality and injury rates per 100 000 population. In addition, we obtained information on the total number of registered vehicles for the period 1938 to 2013 from the Department of Motor Traffic. 11 We plotted the number of road traffic crashes, the population and the number of registered vehicles over the study period using Excel and R (R Foundation for Statistical Computing, Vienna, Austria). We analysed long-term trends in road traffic crashes, injuries and fatalities using the Joinpoint Regression Program Version 4.0.4 (National Cancer Institute, Bethesda, USA). 14,15 Joinpoints are points in a time series at which statistically significant changes occur and joinpoint regression fits a series of joined straight lines between these joinpoints. The program starts with the simplest model of fit and tests several models, while increasing joinpoints, until a statistically significant fit is obtained. The program uses the Monte Carlo permutation method to test for significance. 14, 15 We reviewed the literature on Sri Lanka's traffic and transportation history to identify events that might have contributed to the significant changes observed in the joinpoint analysis. Finally, to supplement our findings, we conducted an additional analysis for the period 1977 to 2013, during which it was possible to separate data on serious and minor injuries.

Results
The incidence of road traffic crashes between 1938 and 2013 is shown in Fig. 1.
There were substantial changes in the trend: for the best fitting model, there were four joinpoints, in 1955, 1974, 2003 and 2007, respectively (Table 1). Overall, road traffic crashes increased markedly from 61.2 per 100 000 population in 1938 to 183.6 per 100 000 in 2013 -a threefold increase. However, the increase was not continuous. Road   Table 2). The road traffic injury rate increased until 1959 (to 92.5 per 100 000 population), decreased from 1959 to 1967 (to 58.1 per 100 000 population) and then increased again until 2013. However, during the latter half of the study period, annual rates deviated substantially from the trend line. Fig. 3 shows that road traffic fatalities increased markedly between 1938 and 2013, from 3.0 to 10.8 per 100 000 population. The joinpoint program indicated two join points, in 1944 (5.3 per 100 000 population) and 1947 (3.0 per 100 000 population), respectively (Table 3). Fatalities increased continuously from 1947 onwards.  Road traffic injury trends, Sri Lanka Samath D Dharmaratne et al. Fig. 5 show the absolute number of road traffic crashes and vehicles, and the population size over the study period. All three variables increased during the 75 years but the most marked increases in road traffic crashes and in the number of vehicles were observed in the second half of the study period. Fig. 6 shows the number of registered vehicles between 2003 and 2012, by vehicle type. Categorized data were not available before 2003. Fig. 7 shows the number of deaths, serious injuries and minor injuries due to road traffic crashes per 100 000 population that were reported to the police between 1977 and 2013. There was a marked increase in the number of serious injuries over the period, from 6.0 to 32.0 per 100 000. Although deaths increased steadily, the number of minor injuries fluctuated considerably.

Discussion
Our analyses cover 75 years of road traffic crashes, injuries and fatalities in Sri Lanka. Although the global status report on road safety produced by the World Health Organization (WHO) in 2013 included details of road traffic fatalities in Sri Lanka between 2001 and 2010, it covered only a short period and did not contain a detailed discussion of the possible reasons for any increase. 16 In addition, WHO's 2004 world report on road traffic injury prevention discussed road traffic fatality trends in Asia between 1987 and 1995 but did not explain trends in individual countries. 1 We found that road traffic crashes, injuries and fatalities in Sri Lanka all increased between 1938 and 2013. However, there were fluctuations over this time; and several factors are associated with these changes.
Road infrastructure developed rapidly between 1938 and 1955. During this period, the greater numbers of crashes may have been due to more frequent road use or to more vehicles on the roads. 1-3 Also, during this period the country had a poorly regulated bus service and drivers competed for passengers and violated traffic laws -this behaviour may have increased the number of bus crashes and made roads less safe for other users. Moreover, although traffic laws were in place, law enforcement was likely to have been poor because Sri Lanka did not have a traffic police unit during this period. Year 1935 1941 1947 1953 1959 1965 1971 1977 1983 1989 1995 2001 2007 2013   Year 1935 1941 1947 1953 1959 1965 1971 1977 1983 1989 1995 2001 2007 2013

Research
Road traffic injury trends, Sri Lanka Samath D Dharmaratne et al.
Monitoring traffic was one of the duties of general police units. 11 Three factors might have contributed to the reduction in road traffic crashes between 1955 and 1974. First, a new traffic law was introduced in 1951 that required all motor vehicles to be registered with the Department of Motor Traffic. A vehicle's roadworthiness had to be tested before registration. 12 Second, in 1953, Sri Lanka established a traffic police division and increased the number of traffic police leading to better enforcement of traffic laws. This division also assisted the Inspector General of Police in introducing and implementing new traffic policies. 11 Third, in 1958, the government nationalized the privately-owned bus service and improved the quality and availability of public transportation. 8 Better and more accessible public transportation might have encouraged more people to travel by bus and discouraged the use of private vehicles.
The introduction of an open economic policy in Sri Lanka could have contributed to the increase in road traffic crashes observed after 1977. This policy made it easier to import vehicles and led to an exponential growth in their numbers: 17 between 1977 and 2003 the number of vehicles increased by 2790%. The increase was greatest in motorized two-and three-wheeled vehicles, 11 which were involved in more crashes than other vehicle types. [1][2][3] During this period, there was no matching expansion of either the road infrastructure or the traffic police -this mismatch might have contributed to the marked increase in crashes observed between 1977 and 2003.
The reduction in road traffic crashes after 2003 might have been due to a smaller proportion of crashes being reported. Although Sri Lanka's traffic law requires all road traffic crashes to be reported to the police, many drivers do not report those that involve only minor injuries or damage. 7 In Sri Lanka, reporting crashes to the police can be exhausting: drivers have to spend many hours at the crash site or at a police station until the police complete the initial investigation. If a police report is required for insurance purposes, drivers may have to wait a few days.
To simplify procedures for drivers, Sri Lankan insurance companies introduced an on-site insurance payment policy in 2002. The so-called on-the-spot insurance scheme paid insurance claims without police reports. 18 Although drivers benefited, the new policy probably led to substantial underreporting of road traffic crashes. Our data on road traffic injuries indicate that underreporting probably occurred: no reduction in road traffic injuries was observed between 2003 and 2007, when road traffic crashes appeared to be falling. People injured in road traffic crashes usually present to hospital emergency departments and hospitals report these injuries to the police based in the hospital, who in turn report them to the Police Statistics Unit. 19 Therefore, although drivers might not have reported crashes to the police, associated injuries would have been reported through the hospitals.
It is important to note that the onthe-spot insurance scheme led not only to the underreporting of crashes but also enabled drivers involved in minor crashes to avoid police and legal procedures. These drivers could subsequently have been involved in further crashes, thereby making roads less safe for other drivers and pedestrians. To improve this situation, Sri Lanka should increase the traffic police workforce and require officers to reach accident sites quickly, to investigate and record all road traffic crashes and to provide on-site police reports for insurance purposes. Moreover, insurance companies should be required to obtain police reports before issuing insurance payments.
In 2007, a parliamentary committee was appointed to investigate reasons for the high number of road traffic injuries in Sri Lanka. 20 There was an increase in road traffic crashes reported after 2007.  1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014  The actions of this committee and the 2009 end of the civil war in the northeast may have both contributed to this increase in crashes reported. However, no amendments were made to the onthe-spot insurance scheme and the discrepancies between reported crashes and injuries continued. The parliamentary committee proposed several interventions to prevent road traffic crashes and a number have been implemented since 2010, including: computerizing driving licence registration; expanding the road network; displaying speed limits at the roadside; and training drivers. However, the available data do not allow us to test for any effects of these interventions.
The increasing trend in road traffic fatalities over the study period was much more consistent than those of road traffic crashes or injuries. This steady increase in road traffic fatalities has also been reported in other low-and middleincome countries, including Cambodia, India, Iraq, Myanmar, Nepal and Saudi Arabia. 16,21 As in Sri Lanka, these increases have been associated with a rise in the number of vehicles, poor traffic law enforcement and underdeveloped road infrastructure. [1][2][3][21][22][23] Our descriptive analysis of road traffic injuries between 1977 and 2013 showed that all types of injuries increased. However, there were greater fluctuations in minor injuries than in serious injuries or fatalities. The 1981 traffic laws that limited the number of passengers in motor vehicles and regularized the vehicle licencing process 10,11 and better traffic law enforcement in 1995 may have influenced these trends. However, passenger limits, licencing and law enforcement do not seem to have had a long-term effect in reducing minor injuries. Underreporting due to the introduction of the on-spot insurance scheme could be a plausible cause for the fluctuations seen after 2003 and this hypothesis needs further research.
One limitation of our study was that we used traffic-police data, which included only road traffic crashes and injuries either recorded by the police or reported to them. However, although crashes were underreported after 2003, we were still able to draw important conclusions from the data available. Another limitation was the paucity of scientific studies of the reasons underlying the fluctuations in road traffic crashes in Sri Lanka. A strength of our study was the use of joinpoint regression to analyse long-term trends in road traffic crashes, injuries and fatalities as has been done in high-income countries. 14,15,24,25 Sri Lanka's burden of road traffic injuries could be reduced by better enforcement of traffic laws, restrictions on the importation of two-and three-wheeled motor vehicles and the introduction of new policies to improve road safety. ■ Competing interests: None declared.  Note: The graphs show deaths, serious injuries and minor injuries due to road traffic crashes reported to the police.