Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Factors affecting bus accident severity in Thailand: A multinomial logit model

Abstract

Bus accidents are a serious issue, with high rates of injury and fatality in Thailand. However, no studies have been conducted on the factors affecting bus accident severity in Thailand. A cross-sectional study was conducted by the Department of Highways, Thailand over the 2010–2019 period. A multinomial logit model was used to evaluate the factors associated with bus accident severity. This model divided accidents into three categories: non-injury, injury, and fatality. The risk factors consisted of three major categories: the bus driver, characteristics of the crash, and environmental characteristics. The results showed that characteristics of the bus driver, the crash, and the environment where the crash occurred all increased the probability of bus accidents causing injury. These three main factors included driving on sloped roads (relative risk ratio [RRR] 3.03, 95% confidence level [CI] 1.73 to 5.30), drowsy driving (RRR 2.60, 95% CI 1.71 to 3.96), and driving in the wrong direction (RRR 2.37, 95% CI 1.77 to 3.19). Moreover, the factors that increased the probability of the accidents causing fatality were drowsy driving (RRR 3.40, 95% CI 2.07 to 5.57) and drivers not obeying or following traffic rules (RRR 3.02, 95% CI 1.95 to 4.67), especially in the northern part of Thailand (RRR 3.01, 95% CI 1.98 to 4.62). The results can provide a valuable resource to help road authorities in development targeting road safety programs at sloped roads in the northern part of Thailand. Stakeholders should increase road safety efforts and implement campaigns, such as raising public awareness of the risks of not obeying or following traffic rules and drowsy driving which could possibly reduce the risk of both injury and fatality.

Citation: Mahikul W, Aiyasuwan O, Thanartthanaboon P, Chancharoen W, Achararit P, Sirisombat T, et al. (2022) Factors affecting bus accident severity in Thailand: A multinomial logit model. PLoS ONE 17(11): e0277318. https://doi.org/10.1371/journal.pone.0277318

Editor: Yanyong Guo, Southeast University, CHINA

Received: August 8, 2022; Accepted: October 24, 2022; Published: November 9, 2022

Copyright: © 2022 Mahikul et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the paper and its Supporting Information files.

Funding: This work was partly funded by Department of Disease Control, Ministry of Public Health, Thailand (FFB640004). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

In developed countries such as the United States and Canada, buses are considered to be safe [1, 2]. However, the situation in developing countries may be different [3]; in Thailand, approximately 300 bus accidents are reported by the Department of Land Transport each year, with about 2,000 injuries and 200 fatalities per year [4]. According to World Health Organization (WHO) estimates, the road traffic mortality rate including motorcyclists, car drivers, and passengers in Thailand remains at 32.2 per 100,000 population [5], which is ranked 1st in Asia in terms of crash fatality rates. While the rate of death among the drivers and passengers of buses is low (2%), bus accidents are a serious public health problem in terms of the property loss and personal damage caused by the accidents [6].

Studies on bus accidents are limited, and the factors associated with the severity of bus accident need to be addressed [7]. Several studies have been conducted on the safety of school buses [8, 9], and a few studies have analyzed the main risk factors associated with the occurrence of bus accident severity [1, 6, 10]. Risk factors comprise bus driver factors (such as not obeying or following traffic rules, abrupt driving, drunk driving, and falling asleep while driving), vehicle condition (such as the age of the bus and engine defects), the nature of the crash (such as the time of day and regionality), and the road factors (environmental factors such as road curves, median openings, road junctions, and highways without frontage roads). In terms of bus driver factors, several studies have examined their association with the severity of the outcome. For instance, in Iran, a study showed that bus accident rates were associated with drivers’ sleep disorders [11], while in the US, the socio-economic of bus driver and speeding were found to be associated with bus fatalities [12]. In Sweden, several studies investigated the correlation between bus accident rates and driver acceleration behavior [13–15]; the studies showed that the correlations were strong enough to confirm that acceleration behavior can be a predictive variable for bus accidents. In less economically developed countries such as Ghana, a study showed that the factors associated with fatalities on the highway included speeding, wrongful overtaking, careless driving, and a lack of experience by drivers [16]. In Asia, several studies have been conducted. In Sri Lanka, the drivers’ personal characteristics, such as high working hours and low salaries, were associated with private bus accidents [17], while in Thailand, the factors associated with the sleep quality among Thai intercity bus drivers were evaluated [18], but the study did not investigate its association with accident severity. In terms of crash characteristics, the time of day [19, 20], month [21], and region [22] where the crash occurred were associated with the severity of the outcome. These results also indicated that accidents occurring during peak evening hours were associated with a higher rate of fatality [23]. However, some studies have shown that crashes occurring during the morning rush hour were also associated with serious injury [23]. In the southern region of the U.S., the likelihood of light injuries and fatalities from bus accidents was lower than all other regions [1]. Additionally, in Iran, long holidays were significantly associated with the rate of accidents [24], which is similar to Thailand; there are usually higher incidences of road traffic injury over Thai long holidays such as over the New Year, in January, and during Songkran, in April [21].

As for road environmental factors, different kinds of road are known to affect the rate of bus accidents. For instance, in Canada, a study found that the factors associated with bus accidents were average daily traffic, public transportation and pedestrian traffic volumes, and transit characters [25]. In Denmark, the severity of bus accidents was associated with the weight of the vehicle, crossing intersections during a yellow or red light, open areas, and slippery road surfaces [19]. In Vietnam, a study revealed that the severity of the outcome increases due to rain, sparse traffic, the accident occurring in the evening or at night, or the accident occurring in a rural area, on roads with at least three lanes, or on curved roads or two-way roads without a physical barrier [26]. In term of vehicle features, a study in the US showed that accidents involving older buses caused more injuries and fatalities than those involving newer buses [6], and that road traffic accidents were associated with engine defects [27], while bus size was associated with the rate of bus accidents [10, 26]. The injury and fatality rate among school buses were lower than for other vehicle accidents [28].

Most studies divided the severity of accident into three categories: property damage only, injury, and fatality [23, 26, 29, 30]. However, there is little specialized research into injury and fatal accident severity, especially for bus accidents in Thailand. Therefore, this study aims to investigate the risk factors associated with bus accident severity in Thailand.

Methodology

Study design and data source

This study is cross-sectional study. Bus accident data were retrieved from the Department of Highways (DOH), Thailand over the 2010–2019 period [31]. Under the DOH composed of 18 offices of highways across the country, followed by “highway districts” composed of 106 districts [32]. Highways are main roads allocated for the transportation of people or goods that operating on urban and rural roads. It connects different regions, provinces, and districts [32]. The data were cleaned, and the missing information was removed from records; 2,911 events of bus accidents were used in this analysis. Some data were received immediately at scene; however, other data were gained afterward. Therefore, the data might not be representative of the real situation, with some of the accident causes missing. Based on a previous study [1], bus driver factors and environmental characteristics were classified as risk factors. A bus driver’s gender and age were not analyzed in this study because of the lack of availability of this information in the dataset.

Ethics approval

This study does not involve human participants. The study was approved by the Ethics Review Committee for Research in Human Subjects (Project Number Approval 64010), Thailand Ministry of Public Health, Department of Disease Control.

Bus driver factors

Drunk driving, drowsy driving, sleep, driving at excessive speeds, abrupt driving, and not obeying traffic rules by the bus driver were considered as potential risk factors affecting bus accident severity. These certain variables were classified into two groups as ‘‘Yes”, for cases where the bus driver displayed proper driving behavior, or ‘‘No”, for cases where the bus driver lacked sleep or disobeyed traffic rules through actions such as driving drunk or driving at excessive speeds. Based on the previous research literature, a bus driver’s behavior was divided two categories [1, 19, 20]. In the U.S., wrong-way accidents were associated with the severity of outcomes [20]; as a result, this study also classified the driving direction into three categories: unknown, right-way, and wrong-way.

Crash characteristics

Previous studies in Thailand over long holidays such as the New Year’s (in January) and Songkran holidays (in April) showed that there are usually higher road traffic accident and injury [21]. Therefore, our study divides the months into twelve categories related with Thai holidays. Time of day was also considered in this study. In previous study, the time of the bus accident was classified into four categories depending on traffic volume [19, 20]: early morning (6.00–9.00 a.m.), morning and afternoon (9.00 a.m.-3.00 p.m.), evening (3.00–6.00 p.m.), and night (6.00 p.m.-6.00 a.m.). However, our study grouped the time of the bus accident into three categories: morning (6.00–11.59 a.m.), afternoon and evening (12.00 p.m.-5.59 p.m.), and night (6.00 p.m.-5.59 a.m.). In the U.S., the probability of injury was reduced in the South region compared to other regions [1]. In Thailand, the estimated transport accident fatality rates were highest in the central region and lower northern region [22]. Therefore, the accidents in this study were divided into six regions: Central, Northern, Northeast, Eastern, Western, and Southern as shown in Fig 1.

thumbnail
Download:
Fig 1. Map of Thailand illustrating 6 regions.

https://doi.org/10.1371/journal.pone.0277318.g001

Environmental characteristics

Based on previous studies, accidents on highways were associated with greater hospitalization [33]. A study in Malaysia showed that increased accident severity was associated with poor horizontal alignment [34]. The road’s horizontal alignment in this study was categorized into three groups: straight, curved, and sharply curved. A previous study in China showed that vertical alignment was also associated with single-vehicle accidents [35]; therefore, in this study, road vertical alignment was divided into four categories: flat, upwardly curved, downwardly curved, and sloped. The type of intersection was associated with crash severity [23]. In this study, the type of intersection was divided into five groups, including no intersection, four-way, T-shaped, Y-shaped, and others. Previous studies showed that access points or median openings are critical locations that influence safety performance [36]; for example, in Croatia, more people were injured than killed at urban junctions [37].

Statistical approach

The factors associated with the severity of bus accident in Thailand were evaluated by a multinomial logit model (MNL). Bus accident severity was divided into three categories: non-injury (property damage only), injury (hospitalization), and fatality. A ‘non-injury’ means an accident which leads to property damage only without any injury cases. An ‘injury’ means an accident which leads to the injury and needs hospitalization of a victim on the bus at least one case. A ’fatal accident’ means an accident which leads to the death of a victim at least one death. Previous studies [1, 38] used the ordered logit or ordinal probit model to analyze the severity outcome, which is characterized by natural ordering (from a low level to a high level of severity). However, some limitations were found from these models [39]. The increased probability of the severest class was related to a decreased probability of the least severe class in an ordered model [40]. Therefore, it is not always suitable to analyze an ordered model for an ordered outcome [41]. Hence, the multinomial logit model has been recommended to be used for evaluation of the severity outcomes [16, 42]. MNL extends the binary logit model for situations where the independent variable has more than two categories [43]. The nominal category of the response variable was transformed into a numerical scale [44]. Let Y = (Y2,…., YI) be a vector of dummy variables given that category 1 is the reference value. Y is a random variable with a multinomial distribution, where Y ∼ MN(p, 1), and probability mass function in (1): (1) Where p = (p1,…, pI) is a vector of probabilities of bus driver experiencing the accident.

For a set of n explanatory variables, denoted by x =(x1,…, xn) in (2): (2) which is a multinomial probability such that .

The multinomial logit model that is used in the study is presented below [23] in (3): (3) where pni is the probability of the bus driver n experiencing the severe injury of i, βi is a coefficient of the accident severity i, and Xni is an explanatory variable.

The maximum likelihood approach was used to estimate βi. The non-injury from bus accident category, which had the highest reporting rate, was used as the reference category. According to the nonlinear assumptions of the multinomial logit model, the estimated coefficients is not suitable for indicating the direct effect on the severity outcomes; therefore, the relative risk ratio (RRR) was used to evaluate the risk factors which is computed relative to the base category. The relative probability of severity outcome (i = 2) to the non-injury category (i = 1) is (4)

The RRR represents the risk or protective factor for bus accident severity. When an RRR greater than 1, it shows an increase in risk, while an RRR lower than 1, it shows a decrease in the risk of a specific injury in terms of severity level compared with the base category [45]. In this study, the MNL and the associated RRR at the 95% confidence level were estimated using Stata (version 14.0). Marginal effects were computed and used to assess the effect of each variable on the bus accident severity outcome probabilities.

The independence of irrelevant alternatives (IIAs) for the response group is the assumption of the MNL [46]. The probability ratio was assumed to be the same for both levels of bus accident severity. The Hausman-McFadden test was used to evaluate the IIA [23, 47] as shown in Table 1. There is no reason to reject the null hypothesis or independence of irrelevant alternatives given the p-value is greater than 0.05. Therefore, the MNL model can be used in this study based on the certainty that the groups are independent of each other.

thumbnail
Download:
Table 1. Independence of Irrelevant Alternatives (IIAs) property test.

https://doi.org/10.1371/journal.pone.0277318.t001

Results

The count and percentage of the factors which might associated with the level of bus severity accident are represented in Table 2. More than half of bus driver (68.5%) was reported having excessive speed. The percentage of bus driver who not obeying traffic rules was similar for each category approximately 30.0%. Only 5–10% of bus driver was reported having abrupt driving and fallen asleep. However, drunk driving was reported with the high proportion of accident causing fatality (60.0%). In addition, the proportions of wrong way driving direction were 27.0%. The largest proportion of accident was reported in April (13.8%) and morning rush hour (50.4%). The central of Thailand was reported the high proportion of bus accident (44.5%). However, in the norther part of Thailand was reported the large percentage of bus accidents causing injury (30.6%). Two-third of accident was found at highways without a frontage road (74.6%). While one-third of accident causing fatality was found at curve road and Y-shaped road approximately 30.0%. About half of the bus accident causing injury was reported at median opening (43.5%) and road junctions (59.1%).

thumbnail
Download:
Table 2. Descriptive statistics (%) of risk factors from 2010 to 2019 in Thailand.

https://doi.org/10.1371/journal.pone.0277318.t002

The results of the MNL model’s estimation are shown in Table 3. In this table, the only statistically significant variable is presented. The estimated coefficient using maximum likelihood approach explains the differences compared to the non-injury outcome. The results show that the factors increasing the probability of injury were the driver falling asleep; driving in the right- or wrong-way direction; driving in the northern, northeast, western, or southern regions; driving on highways without frontage roads; or driving on sloped roads or at road junctions. In addition, the factors increasing the chance of fatality were similar to those causing injury, except for the driver not obeying traffic rules, abrupt driving, drunk driving, driving in the eastern region, or driving at median openings. Also, the results showed that only driving during the afternoon hours can reduce the probability of fatal injuries.

thumbnail
Download:
Table 3. The results of the MNL model for bus accidents in Thailand.

https://doi.org/10.1371/journal.pone.0277318.t003

Bus driver factors

The results show that bus drivers who do not obey traffic rules are 3.02 times (95% CI: 1.95–4.67) more likely to cause a fatality than a non-injury during a bus accident. The results revealed that bus drivers who are drunk while driving are 4.79 times (95% CI: 1.05–21.94) more likely to cause a fatality than a non-injury in bus accidents. There was no association between driving at excessive speeds and the severity of outcome in this study. The results of the model also showed that bus drivers who attempted to overtake another vehicle or displayed abrupt maneuvers were 2.48 times (95% CI: 1.79–3.43) more likely to cause a fatality compared to a non-injury. Accidents occurring due to the driver falling asleep or experiencing fatigue were 2.60 times (95% CI: 1.71–3.96) and 3.40 times (95% CI: 2.07–5.57) more likely to result in injury or fatality compared to non-injury, respectively. Accidents occurring from the driver driving in the wrong direction were 2.37 times (95% CI: 1.77–3.19) and 1.49 times (95% CI: 1.03–2.16) more likely to result in injury or fatality compared to non-injury, respectively.

Crash characteristics

Crashes occurring during peak morning traffic are 1.05 times (RRR = 1/0.95) and 1.45 times (RRR = 1/0.69) more likely to result in injury or fatality compared to non-injury, respectively. Crashes occurring in Northern Thailand are 2.19 times (95% CI: 1.51–3.16) and 3.01 times (95% CI: 1.98–4.62) more likely to cause in injury or fatality compared to non-injury, respectively.

Environmental characteristics

Bus accidents that occur in highways without frontage roads are 1.72 times (95% CI: 1.38–2.15) and 2.45 times (95% CI: 1.77–3.38) more likely to result in injury or fatality compared to non-injury, respectively. The results show that bus accidents occurring on curved roads are 1.81 times (95% CI: 1.28–2.56) more likely to cause a fatality than non-injury. Bus accidents that occur on sloped roads are 3.03 times (95% CI: 1.73–5.30) and 2.81 times (95% CI: 1.49–5.27) more likely to result in injury or fatality compared to non-injury, respectively. The results show that bus accidents that occur at median openings are 1.51 times (95% CI: 1.01–2.26) more likely to cause a fatality than to a non-injury. Bus accidents that occur at road junctions are 2.18 times (95% CI: 1.07–4.44) more likely to result in a fatality than non-injury.

The results of the marginal effects for bus accidents in Thailand are given in Table 4. Drunk driving greatly increases the probability of fatal injury by 42.4% in bus accidents compared to accidents occurring in non-drunk driving. While, wrong-way driving direction increases the probability of injury by 18.2%. In terms of crash characteristics, the crashes occurring in Northern Thailand increases the probability of fatal injury by 11.1% in bus accidents compared to accidents occurring in Central region. Other crash characteristic variables show that the crashes occurring in Northeast Thailand increases the probability of injury by 10.5% compared to accidents occurring in Central region. Environmental characteristics were also found to be significant. Curve roads was found to significantly increase the probability of accident fatality by 8.7% compared to straight roads. While, sloped roads were found to significantly increase the probability of injury by 17.2% compared to flat roads.

thumbnail
Download:
Table 4. The results of the marginal effects for bus accidents in Thailand (%).

https://doi.org/10.1371/journal.pone.0277318.t004

Discussion

Bus driver factors

The study showed that bus drivers who do not obey traffic rules are more likely to cause a fatality than a non-injury during a bus accident. This result is consistent with previous studies [48], which demonstrated that bus drivers not obeying traffic rules tend to experience more severe injuries. The results revealed that bus drivers who are drunk while driving are more likely to cause a fatality than a non-injury in bus accidents. This result is consistent with a previous study [49], which reported that drunk driving among bus drivers is associated with more severe outcomes. In Thailand, the Ministry of Transport has implemented strict pre-departure screening to ban drunk bus drivers since 2011; however, accidents due to the bus drivers being intoxicated still occur [50]. There was no association between driving at excessive speeds and the severity of outcome in this study. In contrast, Prato CG et al [19] reported that high speed limits may increase bus accident severity. A recent study on road traffic injuries showed that drivers driving at excessive speeds were two to three times more likely to cause severe injuries [48, 51]. The results of the model also showed that bus drivers who attempted to overtake another vehicle or displayed abrupt maneuvers were more likely to cause a fatality compared to a non-injury, which is consistent with a previous study [48], which demonstrated that those who were abrupt driving were more than two times likely to be more severe injury. Accidents occurring due to the driver falling asleep or experiencing fatigue were more likely to result in injury or fatality compared to non-injury, respectively. Another study [52] showed that the rate of accident was associated with asleep and tired drivers [53, 54]; this indicates that driver sleepiness or tiredness is a problem for city bus drivers. Accidents occurring from the driver driving in the wrong direction were more likely to result in injury or fatality compared to non-injury, respectively. This result is consistent with a previous study [55], which reported that wrong-way driving crashes are associated with more severe outcomes. To prevent the bus driver risk factors, previous studies showed that awareness campaigns in reducing the accidents on highways should be implemented [56] and the driver should be encouraged to register the driver licensing [57].

Crash characteristics

Crashes occurring during peak morning traffic are more likely to result in injury or fatality compared to non-injury, respectively. This is due to the fact that crashes occurring during the morning hours tend to be at higher speeds or are affected by the conditions of rush hour driving. This result is consistent with previous studies [19, 23, 58]. Crashes occurring in Northern Thailand are more likely to cause in injury or fatality compared to non-injury, respectively. This might be due to the fact that landslides occur in the upper northern region of Thailand, where the roads are mostly curved and the terrain is mountainous [59]. Moreover, Phayao was found to have the highest fatality rate (60.0%) of all provinces. In the U.S., regionality was associated with severity of outcome [1]. In Thailand, the estimated transport accident fatality rates were highest in the central region or lower northern region [22]. Previous studies showed that rescheduling of rush hour driving could reduce the incidence of crashes [60] and avoiding unfamiliar place could be reduced the risk of accident [61].

Environmental characteristics

Bus accidents that occur in highways without frontage roads are more likely to result in injury or fatality compared to non-injury, respectively. This result is consistent with previous studies [33, 62], which reported that accidents on highways were associated with more severe outcomes. The results show that bus accidents occurring on curved roads are more likely to cause a fatality than non-injury. These results are supported by previous studies [63], which reported that the most vulnerable sites for crash occurrence are on horizontally curved roads. Bus accidents that occur on sloped roads are more likely to result in injury or fatality compared to non-injury, respectively. Another study [1, 64] demonstrated that roads with curved and sloped alignments were associated with a greater probability of fatality. The results show that bus accidents that occur at median openings are more likely to cause a fatality than to a non-injury. Previous research [36] has also shown that median openings are associated with a higher risk of severe injury. Bus accidents that occur at road junctions are more likely to result in a fatality than non-injury. This result is consistent with previous studies [37, 65], which revealed that more people were injured than killed at urban junctions. To prevent bus accident severity, previous studies suggested that policy recommendations for improving road conditions should be proposed [26] and road authorities should concern the safety on the high risk roads [66–68].

Limitations

This study only analyzed secondary data at the scene, which were limited to factors available from the Department of Highways database and may not fully explain the causal relationship. Additional factors such demographics, attitudes, and perceptions of bus drivers may help to create a more accurate and inclusive study. Previous studies showed that affecting parameters such as the red-light running behaviors [69], individual characteristics [70], and speed limit standards [71] on the injury severity outcomes may have heterogeneity and temporal instability across different time periods that could exist in the models. As such, random parameters model such as random parameters logit model with heterogeneity [72] and Bayesian approach [69, 70] could be used in further study. Moreover, there was no association between excessive speed and severity of outcome in this study; however, this might be caused by a limitation of the data on speed at the scene of the accident. Therefore, more studies utilizing cohort design, GPS, and detective devices to not only collect data from accidents but also measure the behavior of bus drivers and weather conditions could help evaluate the real risk factors associated with bus accidents.

Conclusions

This study aimed to investigate the risk factors associated with bus accident severity in Thailand. To achieve this purpose, 2,911 bus accidents recorded by the Department of Highways in Thailand over the 2010–2019 period were analyzed using a multinomial logit model. The severity of bus accidents was divided into three categories: non-injury, injury, and fatality. Risk factors included bus driver factors (such as not obeying or following traffic rules, abrupt driving, driving in the wrong direction, drunk driving, and falling asleep), vehicle condition (such as age of the bus and engine defects), crash conditions (such as the time of day and regionality), and the road condition (environmental factors such as road curves, median openings, road junctions, and highways without frontage roads). The relative risk ratio (RRR) was used in this study to evaluate these risk factors; the results showed that the risk factors increasing the probability of injury included driving in the morning, driving in the wrong direction, driving on highways without a frontage road, driving on sloped roads, driving at road junctions, the driver falling asleep, or driving in the northern part of Thailand. Moreover, the factors that increased the likelihood of fatality were similar to those increasing the probability of injury, except for the curve of the road, driving at median openings or road junctions, pulling abrupt maneuvers while driving, drunk driving, and not obeying traffic rules by performing actions such as turning off lights in the nighttime, illegally overtaking other vehicles, or driving through a red light.

The results from the study can provide a valuable resource to help road authorities in development targeting road safety improvements at road curves, median openings, road junctions and highways without frontage roads, as they may need to establish more traffic signs to warn bus drivers at these places. Roads and junctions should be wide and well-lit for the safety of the bus driver. Rapid notification, rescue accessibility, and availability of care should be prepared at highways without frontage roads, especially in the northern part of Thailand and during the morning rush hours. Stakeholders should increase road safety efforts and enact campaigns to promote road safety, such as raising public awareness of the risks of not obeying or following traffic rules, drowsy driving, driving in the wrong direction, drunk driving, and abrupt driving in order to reduce the possible risk of injury or fatality. Driving licenses should be strictly checked and continually renewed to ensure current traffic rules are understood by bus drivers. The results from this analysis obtained in different regions reduced the geographical bias in relation to bus accidents. To narrow the research, this study has exclusively focused on bus accident severity in a developing country, Thailand, which is important to inform the public awareness on bus accident.

Acknowledgments

We thank the Department of Highways (DOH) and Ministry of Public Health, Thailand, for the data inputs. We gratefully acknowledge all experts especially Dr. Chamaiparn Santikarn.

References

  1. 1. Kaplan S, Prato CG. Risk factors associated with bus accident severity in the United States: a generalized ordered logit model. J Safety Res. 2012;43(3):171–80. Epub 2012/09/15. pmid:22974682.
  2. 2. Morency P, Strauss J, Pépin F, Tessier F, Grondines J. Traveling by Bus Instead of Car on Urban Major Roads: Safety Benefits for Vehicle Occupants, Pedestrians, and Cyclists. J Urban Health. 2018;95(2):196–207. pmid:29500736.
  3. 3. Pearce T, Maunder DAC, Mbara TC, Babu DM, Rwebangira T. Bus Accidents in India, Nepal, Tanzania, and Zimbabwe. Transportation Research Record. 2000;1726(1):16–23.
  4. 4. Department of Land Transport. Bus accident data https://web.dlt.go.th/statistics/index.php 2021 [cited October 14, 2021]. Available from: https://web.dlt.go.th/statistics/index.php.
    • 5. WHO. Global Status Report on Road Safety 2018 https://www.who.int/thailand/news/detail/19-12-2018-launch-of-the-global-status-report-on-road-safety-2018-in-thailand2018 [cited October 14, 2021]. Available from: https://www.who.int/thailand/news/detail/19-12-2018-launch-of-the-global-status-report-on-road-safety-2018-in-thailand.
      • 6. Feng S, Li Z, Ci Y, Zhang G. Risk factors affecting fatal bus accident severity: Their impact on different types of bus drivers. Accid Anal Prev. 2016;86:29–39. Epub 2015/10/30. pmid:26513334.
      • 7. Barua U, Tay R. Severity of urban transit bus crashes in Bangladesh. Journal of Advanced Transportation. 2010;44(1):34–41.
      • 8. Anund A, Dukic T, Thornthwaite S, Falkmer T. Is European school transport safe?—The need for a “door-to-door” perspective. European Transport Research Review. 2011;3(2):75–83.
      • 9. Tetali S, Edwards P, Murthy GVS, Roberts I. Road traffic injuries to children during the school commute in Hyderabad, India: cross-sectional survey. Injury Prevention. 2016;22(3):171. pmid:26701985
      • 10. Chimba D, Sando T, Kwigizile V. Effect of bus size and operation to crash occurrences. Accident Analysis & Prevention. 2010;42(6):2063–7. pmid:20728663
      • 11. Razmpa E, Sadegh Niat K, Saedi B. Urban bus drivers’ sleep problems and crash accidents. Indian J Otolaryngol Head Neck Surg. 2011;63(3):269–73. Epub 2012/07/04. pmid:22754808; PubMed Central PMCID: PMC3138945.
      • 12. Blower D, Green P. Type of Motor Carrier and Driver History in Fatal Bus Crashes. Transportation Research Record: Journal of the Transportation Research Board. 2010;2194.
      • 13. af Wåhlberg AE. Characteristics of low speed accidents with buses in public transport: part II. Accid Anal Prev. 2004;36(1):63–71. Epub 2003/10/24. pmid:14572828.
      • 14. af Wåhlberg AE. Speed choice versus celeration behavior as traffic accident predictor. J Safety Res. 2006;37(1):43–51. Epub 2006/02/28. pmid:16499929.
      • 15. Af Wåhlberg AE. The relation of non-culpable traffic incidents to bus drivers’ celeration behavior. J Safety Res. 2008;39(1):41–6. Epub 2008/03/08. pmid:18325415.
      • 16. Damsere-Derry J, Adanu EK, Ojo TK, Sam EF. Injury-severity analysis of intercity bus crashes in Ghana: A random parameters multinomial logit with heterogeneity in means and variances approach. Accid Anal Prev. 2021;160:106323. Epub 2021/08/12. pmid:34380083.
      • 17. Jayatilleke AU, Nakahara S, Dharmaratne SD, Jayatilleke AC, Poudel KC, Jimba M. Working conditions of bus drivers in the private sector and bus crashes in Kandy district, Sri Lanka: a case-control study. Inj Prev. 2009;15(2):80–6. Epub 2009/04/07. pmid:19346419.
      • 18. Chaiard J, Deeluea J, Suksatit B, Songkham W. Factors associated with sleep quality of Thai intercity bus drivers. Industrial Health. 2019;57(5):596–603. pmid:30686814
      • 19. Prato CG, Kaplan S. Bus accident severity and passenger injury: evidence from Denmark. European Transport Research Review. 2014;6(1):17–30.
      • 20. Ponnaluri RV. The odds of wrong-way crashes and resulting fatalities: A comprehensive analysis. Accident Analysis & Prevention. 2016;88:105–16. pmid:26745273
      • 21. Riyapan S, Thitichai P, Chaisirin W, Nakornchai T, Chakorn T. Outcomes of Emergency Medical Service Usage in Severe Road Traffic Injury during Thai Holidays. West J Emerg Med. 2018;19(2):266–75. Epub 2018/02/20. pmid:29560053.
      • 22. Klinjun N, Lim A, Bundhamcharoen K. Epidemiologic patterns of transport accident mortality in Thailand. Southeast Asian J Trop Med Public Health. 2016;47(2):318–27. Epub 2016/06/02. pmid:27244970.
      • 23. Abrari Vajari M, Aghabayk K, Sadeghian M, Shiwakoti N. A multinomial logit model of motorcycle crash severity at Australian intersections. Journal of Safety Research. 2020;73:17–24. pmid:32563389
      • 24. Khorshidi A, Ainy E, Hashemi Nazari SS, Soori H. Temporal Patterns of Road Traffic Injuries in Iran. Arch Trauma Res. 2016;5(2):e27894. Epub 2016/10/06. pmid:27703958; PubMed Central PMCID: PMC5037289.
      • 25. Shahla F, Shalaby AS, Persaud BN, Hadayeghi A. Analysis of Transit Safety at Signalized Intersections in Toronto, Ontario, Canada. Transportation Research Record. 2009;2102(1):108–14.
      • 26. Nguyen TC, Nguyen MH, Armoogum J, Ha TT. Bus Crash Severity in Hanoi, Vietnam. Safety. 2021;7(3).
      • 27. Ahmed M, Khanom K, Shampa RM, Bari MH. Road traffic accident among motor vehicle drivers in selected high ways. Mymensingh Med J. 2004;13(2):165–8. Epub 2004/07/31. pmid:15284694.
      • 28. Yang J, Peek-Asa C, Cheng G, Heiden E, Falb S, Ramirez M. Incidence and characteristics of school bus crashes and injuries. Accident Analysis & Prevention. 2009;41(2):336–41. pmid:19245893
      • 29. Abellán J, López G, de Oña J. Analysis of traffic accident severity using Decision Rules via Decision Trees. Expert Systems with Applications. 2013;40(15):6047–54.
      • 30. de Lapparent M. Empirical Bayesian analysis of accident severity for motorcyclists in large French urban areas. Accident Analysis & Prevention. 2006;38(2):260–8. pmid:16280119
      • 31. Department of Highways. Bus accident data https://www.doh.go.th/visual/accident 2021 [cited October 14, 2021]. Available from: https://www.doh.go.th/visual/accident.
      • 32. Champahom T, Jomnonkwao S, Banyong C, Nambulee W, Karoonsoontawong A, Ratanavaraha V. Analysis of Crash Frequency and Crash Severity in Thailand: Hierarchical Structure Models Approach. Sustainability. 2021;13(18).
      • 33. Zimmerman K, Mzige AA, Kibatala PL, Museru LM, Guerrero A. Road traffic injury incidence and crash characteristics in Dar es Salaam: a population based study. Accid Anal Prev. 2012;45:204–10. Epub 2012/01/25. pmid:22269502.
      • 34. Musa MF, Hassan SA, Mashros N. The impact of roadway conditions towards accident severity on federal roads in Malaysia. PLoS One. 2020;15(7):e0235564–e. pmid:32628689.
      • 35. Chen Y, Wang K, King M, He J, Ding J, Shi Q, et al. Differences in Factors Affecting Various Crash Types with High Numbers of Fatalities and Injuries in China. PLoS One. 2016;11(7):e0158559. pmid:27439113
      • 36. Wang Z, Huang B, Lu J, Zhao J, Zhang Y. Exploring the Impact of Access Designs on Crash Injury Severity on Multilane Highways. Traffic Injury Prevention. 2014;15(1):102–9. pmid:24279973
      • 37. Vorko-Jović A, Kern J, Biloglav Z. Risk factors in urban road traffic accidents. J Safety Res. 2006;37(1):93–8. Epub 2006/03/07. pmid:16516927.
      • 38. Islam S, Mannering F. Driver aging and its effect on male and female single-vehicle accident injuries: Some additional evidence. Journal of Safety Research. 2006;37(3):267–76. pmid:16820170
      • 39. Savolainen P, Mannering F. Probabilistic models of motorcyclists’ injury severities in single- and multi-vehicle crashes. Accident Analysis & Prevention. 2007;39(5):955–63. pmid:17324370
      • 40. Geedipally SR, Turner PA, Patil S. Analysis of Motorcycle Crashes in Texas with Multinomial Logit Model. Transportation Research Record. 2011;2265(1):62–9.
      • 41. Çelik AK, Oktay E. A multinomial logit analysis of risk factors influencing road traffic injury severities in the Erzurum and Kars Provinces of Turkey. Accident Analysis & Prevention. 2014;72:66–77. pmid:25016457
      • 42. Salum JH, Kitali AE, Bwire H, Sando T, Alluri P. Severity of motorcycle crashes in Dar es Salaam, Tanzania. Traffic Injury Prevention. 2019;20(2):189–95. pmid:30888877
      • 43. Withers SD. Data Analysis, Categorical. In: Kobayashi A, editor. International Encyclopedia of Human Geography (Second Edition). Oxford: Elsevier; 2009. p. 159–65.
        • 44. Hashimoto EM, Ortega EMM, Cordeiro GM, Suzuki AK, Kattan MW. The multinomial logistic regression model for predicting the discharge status after liver transplantation: estimation and diagnostics analysis. J Appl Stat. 2020;47(12):2159–77. Epub 2019/12/24. pmid:35706842; PubMed Central PMCID: PMC9041638.
        • 45. Rifaat SM, Tay R, de Barros A. Effect of street pattern on the severity of crashes involving vulnerable road users. Accident Analysis & Prevention. 2011;43(1):276–83. pmid:21094325
        • 46. Savolainen PT, Mannering FL, Lord D, Quddus MA. The statistical analysis of highway crash-injury severities: A review and assessment of methodological alternatives. Accident Analysis & Prevention. 2011;43(5):1666–76. https://doi.org/10.1016/j.aap.2011.03.025.
        • 47. Hausman J, McFadden D. Specification Tests for the Multinomial Logit Model. Econometrica. 1984;52(5):1219–40.
        • 48. Yagoub U, Saiyed NS, Rahim B-EEA, Musawa N, Al Zahrani AM. Road Traffic Injuries and Related Safety Measures: A Multicentre Analysis at Military Hospitals in Tabuk, Saudi Arabia. Emerg Med Int. 2021;2021:6617381-. pmid:33708446.
        • 49. Sam EF, Daniels S, Brijs K, Brijs T, Wets G. Modelling public bus/minibus transport accident severity in Ghana. Accid Anal Prev. 2018;119:114–21. Epub 2018/07/18. pmid:30016751.
        • 50. CTN NEWS. Thailand’s Transport Ministry to Ban Bus Drivers from Alcohol Consumption 2017 [27 April 2022]. Available from: https://www.chiangraitimes.com/news-asia-thailand/thailands-transport-ministry-to-ban-bus-drivers-from-alcohol-consumption/.
        • 51. Hussain Q, Feng H, Grzebieta R, Brijs T, Olivier J. The relationship between impact speed and the probability of pedestrian fatality during a vehicle-pedestrian crash: A systematic review and meta-analysis. Accid Anal Prev. 2019;129:241–9. Epub 2019/06/09. pmid:31176144.
        • 52. Sabbagh-Ehrlich S, Friedman L, Richter ED. Working conditions and fatigue in professional truck drivers at Israeli ports. Inj Prev. 2005;11(2):110–4. Epub 2005/04/05. pmid:15805441; PubMed Central PMCID: PMC1730197.
        • 53. Anund A, Ihlström J, Fors C, Kecklund G, Filtness A. Factors associated with self-reported driver sleepiness and incidents in city bus drivers. Industrial health. 2016;54(4):337–46. Epub 2016/04/19. pmid:27098307.
        • 54. Rey de Castro J, Gallo J, Loureiro H. [Tiredness and sleepiness in bus drivers and road accidents in Peru: a quantitative study]. Rev Panam Salud Publica. 2004;16(1):11–8. Epub 2004/08/31. pmid:15333261.
        • 55. Kemel E. Wrong-way driving crashes on French divided roads. Accid Anal Prev. 2015;75:69–76. Epub 2014/12/03. pmid:25460093.
        • 56. Gopalakrishnan S. A public health perspective of road traffic accidents. J Family Med Prim Care. 2012;1(2):144–50. Epub 2012/07/01. pmid:24479025; PubMed Central PMCID: PMC3893966.
        • 57. Hartling L, Wiebe N, Russell K, Petruk J, Spinola C, Klassen TP. Graduated driver licensing for reducing motor vehicle crashes among young drivers. Cochrane Database Syst Rev. 2004;(2):Cd003300. Epub 2004/04/24. pmid:15106200.
        • 58. Abdul Manan MM, Várhelyi A, Çelik AK, Hashim HH. Road characteristics and environment factors associated with motorcycle fatal crashes in Malaysia. IATSS Research. 2018;42(4):207–20.
        • 59. Dechkamfoo C, Sitthikankun S, Kridakorn Na Ayutthaya T, Manokeaw S, Timprae W, Tepweerakun S, et al. Impact of Rainfall-Induced Landslide Susceptibility Risk on Mountain Roadside in Northern Thailand. Infrastructures. 2022;7(2).
        • 60. Wang SY, Wu KF. Reducing intercity bus crashes through driver rescheduling. Accid Anal Prev. 2019;122:25–35. Epub 2018/10/10. pmid:30300796.
        • 61. Stewart BT, Yankson IK, Afukaar F, Medina MC, Cuong PV, Mock C. Road Traffic and Other Unintentional Injuries Among Travelers to Developing Countries. Med Clin North Am. 2016;100(2):331–43. Epub 2016/02/24. pmid:26900117; PubMed Central PMCID: PMC4764791.
        • 62. Milton JC, Shankar VN, Mannering FL. Highway accident severities and the mixed logit model: an exploratory empirical analysis. Accid Anal Prev. 2008;40(1):260–6. Epub 2008/01/25. pmid:18215557.
        • 63. Buddhavarapu P, Banerjee A, Prozzi JA. Influence of pavement condition on horizontal curve safety. Accid Anal Prev. 2013;52:9–18. Epub 2013/01/10. pmid:23298704.
        • 64. Chang F, Li M, Xu P, Zhou H, Haque MM, Huang H. Injury Severity of Motorcycle Riders Involved in Traffic Crashes in Hunan, China: A Mixed Ordered Logit Approach. Int J Environ Res Public Health. 2016;13(7):714. pmid:27428987.
        • 65. Zhao XG, He XD, Wu JS, Zhao GF, Ma YF, Zhang M, et al. Risk factors for urban road traffic injuries in Hangzhou, China. Arch Orthop Trauma Surg. 2009;129(4):507–13. Epub 2009/02/18. pmid:19221774.
        • 66. Abdollahzadeh Nasiri AS, Rahmani O, Abdi Kordani A, Karballaeezadeh N, Mosavi A. Evaluation of Safety in Horizontal Curves of Roads Using a Multi-Body Dynamic Simulation Process. Int J Environ Res Public Health. 2020;17(16). Epub 2020/08/23. pmid:32824601; PubMed Central PMCID: PMC7459981.
        • 67. Jones AP, Haynes R, Harvey IM, Jewell T. Road traffic crashes and the protective effect of road curvature over small areas. Health & Place. 2012;18(2):315–20. pmid:22079862
        • 68. Polus A, Pollatschek MA, Farah H. Impact of infrastructure characteristics on road crashes on two-lane highways. Traffic Inj Prev. 2005;6(3):240–7. Epub 2005/08/10. pmid:16087465.
        • 69. Guo Y, Li Z, Wu Y, Xu C. Exploring unobserved heterogeneity in bicyclists’ red-light running behaviors at different crossing facilities. Accident Analysis & Prevention. 2018;115:118–27. pmid:29558688
        • 70. Guo Y, Li Z, Wu Y, Xu C. Evaluating factors affecting electric bike users’ registration of license plate in China using Bayesian approach. Transportation Research Part F: Traffic Psychology and Behaviour. 2018;59:212–21.
        • 71. Alnawmasi N, Mannering F. A statistical assessment of temporal instability in the factors determining motorcyclist injury severities. Analytic Methods in Accident Research. 2019;22:100090.
        • 72. Xie P, Qin D, Zhu T. Impact of rule-violating behaviors on the risk of bus drivers being at-fault in crashes. Traffic Inj Prev. 2022;23(6):364–8. Epub 2022/06/08. pmid:35670548.