Factors affecting of road traffic injuries in the lower southern region of Thailand: a content analysis of road traffic investigation reports, 2006-2019

Background – Thailand has the second highest rates of road traffic mortality globally. Detailed information on the combination of human, vehicle and environmental risks giving rise to each incident is important for addressing risk factors holistically. This paper presents the result of forensic road traffic investigation reports in Thailand and determines risk factor patterns for road traffic injuries. Methods – Detailed forensic reports were extracted for 25 serious traffic accident events. For each report, accidents were characterized by the number of vehicles, number and type of injuries and deaths, road user types, involvement of roadside hazards, medical care provided, and human, vehicular and environmental influences on the incident. The Haddon matrix was used to analyze risk factors in three phases (pre-event, event, post-event) stratified by four agents (human, vehicle, physical and socio-economic environment). Results - The 25 events analyzed involved 407 victims and 47 vehicles. 65.8% of victims were injured, including 22.0% who died. Median age of those injured was 27 years, and 28 years for fatalities. Vehicles crashing with fixed objects, and pickup-related crashes, accounted for the majority of accidents and deaths. Head or neck trauma was the main cause of death. The majority (61%) of deaths occurred at the scene. Human error-related factors included speeding, and drowsiness from night and long-distance driving. Passenger risks included not using seat belt, sitting in cargo area and the cab of pickups. Not having anyone trained in first aid on the scene, first aid being provided by bystanders, and delayed calls to Emergency Medical Services increased injury risk. Overloaded vehicles, unsafe car modifications, no occupant safety equipment, and having unfixed seats were vehicular risks. Environmental risks included fixed objects on the roadside, no traffic lights, no guard rails, no traffic sign, road accident black spots, and hazardous objects roadside. Conclusions - Thailand must address all three temporal phases of the Haddon model and all three factors – human, vehicle and environment. At present traffic accidents cause much avoidable severe injury and death. The Haddon matrix is useful to structure road traffic investigations, revealing multi-level factors common on Thai roads.


Introduction
Road traffic injury is a significant cause of morbidity and mortality, and is in the top 10 causes of death globally, leading to approximately 1.4 million deaths each year. Among populations aged 15-29 years, road traffic injuries are the top cause of death [1]. Almost 75% of road traffic accident deaths affect males aged under 25 years [2]. As well as mortality, road traffic injuries lead to significant morbidity, with global estimates of more than 20 million nonfatal road traffic injuries occurring each year. Disability and economic impacts of fatal and non-fatal injuries are substantial. As well as the age distribution of road traffic injury, socioeconomic status also plays a role in injury risk. Low-and middle-income countries experience 90% of road traffic accident mortality [2].
Southeast Asia is a particular hot spot for road traffic injury, accounting for 25% of global mortality, or more than 300,000 deaths annually [3] Within this region, Thailand has the highest rate of mortality from road traffic injury, and in 2015 had the second highest rate globally, following Libya, with an estimated fatality rate of around 36.2 per 100,000 population [4]. The incidence of traffic injuries rose from 449.0 per 100,000 in 2012 to 524.9 cases per 100,000 in 2016 [5].
As in many other countries Thai road traffic injury and road traffic mortality are connected to a number of identified and preventable risk factors. Firstly, motorcycles are associated with nearly 40% of all road traffic accidents. Male sex and being aged less than 40 years also increase risk of road traffic injury. Furthermore, recent surveys have shown that only around one-third of car users regularly wear safety belts [6], and regular helmet use on motorcycles was only just over 40% [7].
National statistics such as those reported here however only give an overall view of the road traffic safety situation. Each road traffic incident has its own set of, generally preventable, drivers. Examining individual accident events in detail can help us understand where preventative measures could have been employed. The Haddon Matrix is a tool which can help to examine road traffic accidents in detail using a qualitative analysis method [8]. It allows the combined assessment of conditions leading to the accident, and the severity of injuries, in terms of human, vehicular, physical environment and socio-economic environment factors. This analytical method then adds a temporal dimension examining how these three factors operated pre-crash, during the crash, and post-crash.
The purpose of this paper is to use the Haddon Matrix approach to examine in detail a number of fatal and non-fatal injury incidents in Thailand and consider the appropriate injury prevention techniques. This information can be used to support the Thai government in implementing policies to reduce the number of road traffic injuries in Thailand. The objectives of this study were: 1. To understand drivers of road traffic injuries in Thailand by using indepth road traffic investigations; and 2. To determine the factors influencing injury and injury severity using the Haddon's matrix model.

Study design
This study comprises a content analysis of road traffic investigation reports for traffic incidents occurring between November 2006 and April 2019 in the lower southern region of Thailand. The criteria for including incidents in the investigation are as follows: 1) Road traffic injuries resulting in five or more deaths, or 15 or more injuries; 2) Road traffic incidents at sites which have had injuries and deaths occur previously, more than 2-5 times in the same month; 3) Particularly noteworthy incidents from a community perspective, such as injuries to students, or accidents involving public transportation or ambulances [9]  Thailand. This is the responsible government body for documenting road traffic injuries in this part of Southern Thailand. The content of these reports was analyzed using Haddon's matrix to understand the key characteristics of each incident.

Data extraction and Filtering Criteria
The first step of the study was to extract the full in-depth reports of selected road traffic injury incidents from the database of the reginal Office of Disease Prevention and Control 12, Songkhla Province, Thailand (ODPC12). The reports were prepared by multidisciplinary teams from the Ministry of Public Health including epidemiologists, nurses, public health officers, civil engineers, police and other related agencies. The reports included inspection outcomes from crash scenes, in-depth interviews with victims, eye-witnesses, and health staff, and information extracted from the medical records concerning ambulance and hospital treatment [10].
Recording forms were developed by the authors to extract the relevant variables from these reports for this study. Variables extracted comprised: 1. Event characteristics: vehicle-related crash and/or other parties involved in incident (such as pedestrians), involvement of roadside object hazards, collision types, and type of vehicle.
2. Road user information: sex, age-groups, road user type (driver, passenger, pedestrian), characteristics of those injured (driver, passenger, pedestrian), number of victims, number of injured, and number of deaths.
3. Injury details: region of body injured consisted of head and neck, face, chest, abdomen and pelvis, extremities and pelvic girdle, and external and skin injuries. [11] 4. Prehospital trauma care: type of assistance provided, and mode of delivery to hospitals as reported by eyewitnesses, police officers, or rescue terms (First response unit: FR., Basic life support unit: BLS, Advanced life support unit: ALS) 5. Factors influencing the injury: this followed Haddon's matrix system addressing human, vehicle, physical environment, and socio-economic environment factors which influenced likelihood of injury within three phases of influence (pre-event, event, post-event) [10,12,13] In total, 24 road traffic investigation reports were chosen with a total of 25 events because two events were included in the same report. These reports covered events in five provinces in the lower southern region of Thailand namely, Trang, Satun, Phatthalung, Songkhla, and Pattani.

Data Analysis
The data extraction was performed during April 2020. The first analysis step involved calculating descriptive statistics for the traffic incidents. Event characteristics, road user information, and injured body parts were described using frequencies and percentages.
Next the extracted content from the traffic accident reports was evaluated qualitatively using the Haddon's matrix model as an analytical framework. Factors influencing the likelihood and severity of injuries for each traffic accident were categorized in terms of the host, the vehicle, the physical environment and the socio-economic environment in three crucial temporal stages (the pre-event phase, the event phase and the post-event phase) [10,14]. This analysis aims to identify determinants of road accidents themselves, and also the determinants of injury, injury severity and death, at multiple levels, and before, during and after the accident. The advantages of this approach are the ability to identify structural and behavioral risk factors. Table 1 shows the details of the 25 road traffic investigation reports analyzed, involving 407 victims, 47 vehicles, and 27 roadside object hazards. Of the victims, 268 (65.8%) were injured, including 59 deaths (14.5%). Roadside objects were a factor in 18 events. The incidents with the highest number of injuries were a multiple-vehicle crash in front of a school with 34 injured, followed by an overturned bus crashing into a tree with 24 injured. The incidents with the highest number of deaths were a school pickup crashing into a tree with 10 casualties, followed by a worker pickup running off the road and crashing into trees with 7 persons deceased. Events in which all victims died consisted of a car burning after hitting a tree, a prime mover truck crashing into a pedestrian bridge and a pickup crashing into a motorcycle.   Table 3 shows approximately 52.0 % (n=13) of events were single-vehicle accidents.

Overview of event characteristics
The most common types of single-vehicle accidents were crashes with fixed objects such as a tree, or electric pole, 7 events (53.8%). 72% (n=18) of all events involved fixed object crashes.

Type of vehicles
Of the 47 vehicles in total, there was 17 (36.2%) vehicles whose occupants were injured. The maximum vehicle number involved in a single event was eight vehicles. Figure   1a shows approximately 42.6% (n=20) of vehicles involved in accidents were pick-ups, and 17.0% (n=8) were cars. Figure 1b shows the distribution of vehicle types was similar when we looked only at vehicles whose occupants were injured, approximately 42.4% (n=14) of vehicle type involving injury were pick-ups, and 21.2% (n=7) cars.

Body region of injuries
One person could be injured in several body parts.   Table 7 shows that the first call in the scene in the case of a serious injury accident was call to Medical Emergency Call from 1669 (44.0%).

Risk factors
More than half of the accidents overall involved four risk factors, 14 events (56.0%), followed by three risk factors, 10 events (40.0%). All accidents involved a combination of human error in conjunction with physical environment risk factors. Vehicle defects were a risk factor in 84.0% of accidents. All road traffic accidents which caused injury involved a combination of human errors in conjunction with vehicle defects. Physical environment defects were a factor in 96.0% of injury causing accidents. Table 9 shows a Haddon's Matrix analysis of risk factors in three crucial stages of the road traffic injuries. In the pre-event phase, speeding was the most common human error, accounting for 17 events (68.0%). Fixed objects in safety zone was a physical environment defect for 15 events (60.0%), and accident black spots were a social-economic environment defect involved in 12 events (48.0%). In the event phase, the human error of not using the safety belt and being ejected from the seats or the vehicles, was most common in causing road traffic injuries. Vehicle hitting fixed objects in safety zones was found among most road traffic injuries. In the post-event phase, delayed call to EMS and no persons trained in first aid on scene were found to be the largest contributor to injury risk. 'An additional

Discussion
The reviews of 25 in-depth road injury investigations presented here reveal some of the common patterns and risk factors connected to the major problem of road injuries in Thailand.
The Haddon Matrix approach provided a useful tool to examine the multiple factors connected to injury and severity of injury in physical and temporal dimensions. This adds to much previous literature focusing on mainly driver behavior or other singular risks. This study showed that multiple injuries and more serious injuries most often involved vehicles leaving the road and collisions with roadside hazards. This pattern has also been found in other indepth road traffic analyses [15,16,17]. Moreover, the pickup was the most common vehicle type in road traffic accidents and was associated with the most injuries and deaths. Severity of injuries and fatalities were also linked to the number of occupants in the vehicles, as found elsewhere [18], and carrying passengers in the cab and cargo area [19,20]. In fatal accidents, most victims died at the scene. Head or neck trauma was the main cause of deaths [21,22].
The results of the study confirmed other research showing that road traffic injuries are caused by the main risk factors comprising human errors, vehicle defects and environment defects [13,23,24,25]. The work of multidisciplinary road traffic investigation teams were important keys to recognizing human errors increasing risk pre-crash including behaviors of driver (speeding, drowsiness from night or long-distance driving, insufficient driving skills), and risk behaviors of others (insufficient understanding of traffic rules, risk behaviors of pedestrians or passengers). This study confirmed that exceeding the speed limit was the most common violation involved in road accidents [24] and lead to more severe injuries [26,27].
Moreover, the drowsy driver appears widespread and has been found in several countries [28,29].
The vehicle defects causing the pre-crash consisted mainly of overloading and modification. The environment defects pre-event were documented as many fixed objects in safety zone, no traffic lights, no guard rails, no traffic signs. It was also worth noting that there have been frequent crashes at these accident blackspot areas. This study is similar to observations in other studies [10,16,30]. However, this study showed that multiple risk factors at the pre-crash phase lead to the severity of traffic accidents at that time more than any single risk factor such as "high speed in the presence of overloading and fixed objects in safety zone", "high speed while distracted in spite of overloading and fixed objects in safety zone", "high speed in spite of drowsy from night driving and long-distance driving together with overloading and fixed objects in safety zone", " high speed while bent down to pick up things on floor", "Distracted driving from impaired shoes which go against the accelerator while U-turn" Based on the in-depth road traffic injury investigation in crash risk factors, the human errors involved non use of seat belt, sitting in a place without safety device such as the cargo area and the cab of the pickups, and being ejected from seats or vehicles. Some other event factors were documented such as no occupant's safety belt in vehicle, vehicle structural damage, collision with object on and off road, the body impacted ground or fixed objects on side road. This study is similar to publications, and observations in other studies [10,16,17,31,32,33]. Then, multiple inter-related risk factors at the crash phase lead to the severity of traffic injuries such as combinations of: "non use of seat belt, the body collision with object on vehicle or ejected from vehicle and hit into the tree", or "the pickups rolled over and hit the tree, passenger were sitting in the cargo area of the pickups, no self-defense, ejected from the cargo", "the pickups ran off the road hitting the tree making the passenger were sitting in the cab of the pickups that no seat belt clashing with object on vehicle" Post-crash a further set of risks affected severity of injury. The important post-crash risk factors were reported as not having any persons trained in first aid at the scene, delayed call to EMS, or non-availability of EMS. Fatality and disability resulting from accidents was also associated with other post-event factors including being stuck in vehicle, flames in the car, vehicle over the body, difficult to access and evacuate, and lack of slitting equipment. This phase is important to reduce fatalities and disabilities [10,25].
Searching for in-depth information on all elements of the road injury is important for finding the root cause of the problem, where the immediate injury may be just the tip of the iceberg. [34,35,36]. The analyses presented in this paper provide multiple avenues to improve road safety in Thailand. These include improvements in human risk behaviors, the improvement of physical road infrastructure, safety standards for vehicles, higher rates of first aid training among the general population, and better coordination of EMS services. A holistic approach to road traffic accident prevention in Thailand can both reduce the injury rate and the severity of injuries incurred.

Availability of data and materials
The datasets used and analyzed during this study are available from the corresponding author follow by reasonable request.