Human Injury causing Road Tra � c Accident at Debre Markos Town


 Objective: A Road traffic accident (RTA) is when in a road Vehicle collides with another Vehicle, pedestrian, animal or geographical or architectural obstacle. The RTAs can result in human injury, property damage and death. RTA result in the deaths of 1.2 million people worldwide each year and injuries about 4 times this number. The objective of this study is to identify the main causing-factors that contribute to road traffic accidents involving human injuries. Literature suggested factors considered for analysis are: Driver's Age, Driver’s Education status, Driver's experience, Vehicle type, Driver Vehicle Ownership, Vehicle year of service, Road type, Road division, Road condition, Problem with car, Weather condition, and accident time (day or night). Results: Among the candidate variables, Pearson Chi-Square method identified weather condition, driver’s experience, Vehicle year of service; Road division, Driver Vehicle Ownership, and accident time (day or night ) as significantly associated variables. Furthermore, percentage is used to describe the magnitude of associated variable. The result from Poisson regression analyses revealed that low driver experience, high Vehicle year of service (old cars), and Road division (one way road) are the significant contributing factors for increment of human injuries accidents.


Introduction
A Road tra c accident (RTA) is a type of accident occurred at road when a Vehicle collides with another Vehicle, pedestrian, animal or geographical or architectural obstacle. The RTAs can result in injury, property damage and death. RTA is a cause of death for 1.2 million people worldwide each year, and injures about 4 times this number (WHO, 2004). In this study, a road tra c accident is de ned as accident which took place on the road between two or more objects, one of which must be any kind of a moving Vehicle [1]. Road Tra c Accidents (RTAs) are increasing with rapid pace and presently these are one of the leading causes of death in developing countries.
The morbidity and mortality burden in developing countries is rising due to a combination of factors, including rapid motorization, poor road and tra c infrastructure as well as the behavior of road users [2]. However, in technologically advanced countries injuries are reducing [3,4]. Lack of protection, drivers, passengers, and the involved pedestrians are more likely to sustain injury or fatality at the impact of tra c collision [5].
Human factors, including road user behavior and incapacitation, are the most common factors, accounting for more than 85% of all tra c crashes among them, the two key known contributing factors are speeding, and drinking and driving. Other contributing factors are unsafe vehicles, unsafe road design, and the related lack of effective law enforcement and safety regulations [6]. Shankar et al. (1995) explored the frequency of occurrence of highway accidents on the basis of a multivariate analysis (by negative binomial model) of road way geometries (e.g. horizontal and vertical alignments), weather and other seasonal effects. The results of the analysis uncover important determinants of accident frequency. Karlaftis and Golias (2002) studied the effects of road geometry and tra c volumes on rural roadway accident rates using Generalized linear modeling. The results showed that although the importance of isolated variables differs between two-lane and multilane roads, 'geometric design' variables and 'pavement condition' variables are the two most important factors affecting accident rates. Particularly all accident frequency increased with speed to the power of approximately 2.5.
The literature review shows that pedestrian crashes account for more than 40 percent of crashes in most of Africa countries. For example: pedestrians accounted for 55% of road tra c deaths in Mozambique between 1993 and 2000 [7]. Pedestrians account for 46% of road tra c deaths in Ghana between 1994 and 1998 [8]. Pedestrian and passenger crashes represented 80% of all road tra c deaths in Kenya in 1990 [9].
In a study of motorcycle collisions using hospital data from Nigeria, motor cyclist involved crashes are the second most common cause of road tra c injuries in Nigeria. From the victims more than half are passengers (39.5) and pedestrians (13.8) [5]. Most of car accidents on pedestrian happened on male, with a three to one ratio to female fatalities. The death of the most productive member exerts a devastating impact to the families, pushing many into poverty with long lasting effect to their children and their community at large [10]. This creates an overwhelming burden to the most vulnerable road users and their families, which tend to be poor and less educated.
RTA and injuries constitute major health, economic, and developmental challenges to developing countries, especially those in Africa. Africa has the highest fatality rate in relation to population (28.3 per 100,000 population after adjusting for underreporting), which is higher than motorized countries in the world, such as those in North America (12.1 to 16.2 per 100,000 population) [11].
Accidents and injuries due to RTAs is the second most common & accounted for 22.8% of all such incidents in Ethiopia. RTAs contributed to 43.8% of all fatalities next to other accidents and injuries. Among RTA causalities, 21.9% were drivers, 35.0% were passenger vehicle occupants and 36.0% were vulnerable road users including: motorcyclists (21.0%), pedestrians (12.1%) and cyclists (2.9%) [12].
Recent studies on RTA in Ethiopia have shown the escalation of the problem at the national level, at least 70 people die for every 10,000 vehicle accidents annually. Particularly Amhara region accounted for 27.3% of the total road tra c accident related deaths in Ethiopia during the year 2008/9, it takes the highest share among all regions. This indicates the need to examine the cause of accidents on the region [13].
Pedestrians and passengers of commercial vehicles are the most vulnerable in Ethiopia. Factors like poor road network, absence of knowledge on road tra c safety, mixed tra c ow system, poor legislation and failure of enforcement, poor conditions of vehicles, poor emergency medical services, and absence of tra c accident compulsory insurance law have been identi ed as key determinants of the problem [14].
Road tra c accident in Amara region indicates that freight vehicles (51%) and passenger vehicles (34.5) are the main causes of accidents. And interstate highways takes 54.8% of the accidents occurred. Mainly pedestrians passengers were accounted for the largest part of road tra c deaths victims in the urban areas. This accident is mainly due to drivers problem of failure to give priority to pedestrians, failure to stay on the right side of the road, speeding, failure to maintain distance between vehicles and failure to yield the right of way for other vehicles [13].
Recent report on Debre Markos town tra c accident indicated that there is an increase in accident and huge loss due to many factors. This enhances socio-economic instability in addition to physiological and physical damage. In response to this, road and transport authority o ce plan to work with other organizations, to have good behaved and well trained human resources, and based on information and studies developing awareness and work with the society, to achieve a gole of creating stable accident free transportation system. The approach is based on developing awareness of society on rules of road and transportation system, which helps the society to know and respect rules for the purpose of reducing car accident. In addition creating good behaved and well trained human resources is another input. It known that to reduce level and number of accident the in uence of diver and related factors are unrepeatable hence much concern should have to give. However, the quality of drivers, cars, and road standard does not considered. As a result, this study attempts to investigate factors those contribute to road tra c accident involving human injuries, by considering literature suggested factors which are recorded by tra c police o cers.

Data Source
The Cross-sectional study is applied for the number of human injuries per Vehicle accident at Debre Markos town in a year from 2014 to 2019(1615 consecutive days). The data is taken from the record and report book of Debre Markos town tra c police. Accidents were recorded by the tra c police on daily basis for the purpose of reports and public service.

Missing Value Treatment
The data set is extracted from recorded document, and some values were incomplete or missed. So in ordered to treat such problem I preferred to use median value for categorical variable and exception maximization method for continuous variable.

Variables of the study
The response variable is number of human injuries per accident over a day. This includes the number of people dead, lightly injured or heavily injured due to a tra c accident. The considered predictors are Driver's Age, Driver's Education status, Driver's experience, Vehicle type, Driver Vehicle Ownership, Vehicle year of service, Road type, Road division, Road condition, Problem with car, Weather condition, and accident time (day or night) [See Table S1].

Variable selection
The analyses were started by considering the whole explanatory variables, and a signi cantly associated factor with a dependent variable is taken to check the causal effect on a response. Pearson chi-square is used to select which variable should be entered into reduced model. Finally, the log-likelihood of the full model and reduced model were compare to select the variables which stayed in the model.

Poisson Regression Model
In statistical analysis of the count of rare event, it is often assumed that the dependent variable follows a Poisson distribution, with the assumption of the mean (expected) count equal to its variance. In practice, the variance is much larger than the mean. This is often referred to as "over dispersion" with respect to the Poisson distribution.
The number of accidents occurred in a day follow a Poisson distribution with parameter λ, and for a given independent variables (x 1 , x 2 , ..., x m ) the probability of an event occurred is modeled as: Where the log of the mean occurrence (λ) is assumed to be a linear function of the independent variables. Which implies that λ is the exponential function of independent variables:

Descriptive Statistics
Descriptive statistics is used to compare and describe the magnitude of signi cantly associated factors using percentile. The result in Table S3 indicates that, explanatory variables which have signi cant association with number of human injure in road tra c accident are: weather condition, driver's experience, Vehicle year of service; Road division, Driver Vehicle Ownership, and accident time (day or night).
The accident occurred by diver problem (56.2%) takes the larger proportion compared to technical problem (43.8%), this result is supported by previous studies [7 & 14]. Vehicle owner ship were also considered as a source of accident and it indicates that, employed diver takes the larger portion of accident (58.8%) than owners (41.2%).

Inferential Statistics
Poisson Regression is tted for the full model, and the reduced model by using only signi cantly associated explanatory variables (six variables). The result in Table S4 indicates that, even if the full model is signi cant, only two variables from 13 are signi cant. Where as the result in Table 1 indicates that, from six signi cantly associated variables 3 variables (with one additional variable) have statistically signi cant causal effect on Vehicle accident.
Hence, tting by using signi cantly associated 6 variables gave a better model comparing to the full model. As a result the Poisson model for Vehicle accident is as follow: Log(λ)= β 0 + β 1 X 1 + β 2 X 2 + β 3 X 3 = 0.818 + 0.091X 1 -0.022X 2 + 0.102X 3 Where X 1 is Vehicle year of service, X 2 is driver experience, X 3 is road division.
Based on the tted model above: The odds of number of vehicle accident for one year older vehicle is 1.095 times higher. So as a vehicle becoming older the accident will increase. Older vehicles are more likely to cause an accident comparing to a new one. The odds of number of vehicle accident for one more year experienced drivers is 0.022 times lower. Un-experienced drivers are more likely to cause an accident comparing to a experienced drivers. The odds of vehicle accident is higher for one directional road (1.107 times) as compared to odds of two directional road. The result from Poisson regression reveals that driver experience, road division, and Vehicle year of service are the signi cantly causing factors which affect the number of human injuries. In order to reduce the number injury in Debre Markos Town, for effective and safe tra c management, the concerned transportation authorities can consider the above mentioned predictors as potential causes of accidents in their order of importance in order to take preventive measures. Speci cally, controlling new driver, removing old Vehicles, and controlling one way road (or building two-way road) can help to reduce human injuries Vehicle accidents.

Limitation
This study is a cross-sectional study which does not consider the spatial and temporal analysis od the accident so further study on this aspect is advisable. The data is secondary data design for public use and report by Debre Markos town tra c police o ce, and any researcher can take and do a research. And formal ethical ethical approval is not required.

Consent for Publication
Author proves consent of publication for this research.

Availability of data and material
All important data and material are available. It is a free access data any researcher can take and make a study for the improvement of public service.

Competing interests
The author declare no competing interests regarding this paper.