Assessing Factors for Occurrence of Road Accidents in Tanzania Using Panel Data Analysis: Road Safety Perspective

This study assessed critical factors for road traffic accidents and associated mitigation to reduce the accidents by the year 2030. The study was guided by research questions, what are the major causatives of road accidents and how to mitigate the problem. The study used secondary data collected from the repository database of traffic police at the division of Tanzania Road Safety Squad. Data were collected at the events of accident occurrences and reported annually by regions. Panel data analysis was used to allow for controlling variables which cannot be observed over time and across areas such as regions. Pooled Poisson model, fixed effect and random effect Poisson model was applied to assess factors for road traffic accidents. Fixed effect model was the best model with a reasonably good fit. Results indicated that all predictors are significant under fixed effect Poisson model with a p-value less than 0.05 but Passengers and Railway crossing road was found insignificant and dropped in the final model. Laws and regulatory frameworks should be formulated and enforced promptly for Tanzania may reach the target of 2 decade of action for roads safety 2021-2030.


Introduction
The 2 nd decade of action for roads safety 2021-2030 calls for all countries to achieve at least a 50 percent reduction of road deaths and injuries by 2030. This is a global performance targeted agreed by the member of state to achieve. The World Health ranking report published in 2018 indicated that about 50 percent of deaths globally are caused by fatal road accidents. The Report highlights that road traffic injuries are now the leading killer of people aged between 5 -29 years. Every day, almost 3700 people are killed globally in road traffic crashes involving cars, buses, motorcycles, bicycles, trucks, or pedestrians [1]. High proportional burden is borne by pedestrians, cyclists and motorcyclists, in particular those living in developing countries [1].
Road traffic accidents are the major cause of deaths and injuries from road vehicle crashes and the growing public health epidemic [2]. Road accident does not only kill people but it is a disaster that adversely affects people, property and environment. Economically, families are hardest hit by both direct medical and indirect costs such as lost wages that result from these injuries. At the national level, road traffic injuries lead to considerable financial costs, particularly in developing countries [3]. Based on these effects, achieving a reduction of road deaths and injuries by 2030, required more control measures on causes of road accidents to be effective [4].
Globally, several studies have been done on causes of the traffic road accidents. For instance, a study on the causes of road accidents was conducted in India and developed a dynamics model for reducing the severity of road accidents by considering human, road, vehicle and environmental factors [5]. It found that human factors were significant and accounted for 95% of all road traffic accidents. Another study formulated a multiple regression model that analyzed the factors that lead to motorcycle accidents in Kilimanjaro and Arusha regions [6].
The results showed that the number of motorcycle accidents in both Arusha and Kilimanjaro regions had a strong relationship with the experience of drivers, tarmac road, and personal status. Likewise, the study on modelling road traffic accidents was conducted using a logarithmic-linear regression to determine the main factors contributing to road traffic accidents in Turkey [7]. They found that the main factors contributing to road traffic accidents in Turkey were road defects, vehicle defects, passenger errors, driver defects and pedestrian errors, among general factors, driver defects had highest effects, while road defects had the lowest effects on road traffic accidents.
Expansion of road infrastructure and economic growth has been a significant increase in transportation equipment [8]. This increase has not been in line with road safety literacy and training among road users. There have been noted noncompliance practices especially with traffic rules and regulations among drivers and other road users [6]. This situation has resulted in an increase in traffic offenses and road accidents.
In Tanzania, road traffic accidents have become one of the concerns to most Tanzanians in recent times, about 19,058 road traffic accidents deaths were recorded in 2018 which are 5.24 percent of total deaths [9]. The Government of Tanzania has been taking several measures to reduce the road accidents such as  [11]. As part of the national initiative, Tanzania also implemented program campaign aimed at improving road safety in both urban and school children [2]. However, despite these initiatives, the challenge remains a big threat and burden to Tanzania's economic capital and human resources.
Road accidents increased the burden of expenses to the nation and to the individual particular money used in medical, emergence transport, cost of police and funeral. Therefore the increase in the road traffic accident will have direct or indirect effects on the economy of the country. Understanding the trends and patterns of road traffic accidents can help to predict the need for additional services and resources and guide policy-makers on implementing appropriate and sustainable prevention strategies on a long-term basis. This study used a panel data analysis using Poisson models to assess critical factors affecting road accidents for Tanzania Mainland from 2013 to 2020. The Poisson models are used as standard models for the count data [12]. Understanding critical factors associated with traffic road accidents is important for appropriate planning and control measures against road accidents. Therefore, this study aimed at assessing factors for occurrence of road accidents in Tanzania using panel data analysis.

Methodology
The study used secondary data collected by traffic police in the division of Tanzania Road Safety Squad from 2013 to 2020 in 30 regions in Tanzania Mailand.
These data are panel data collected occasionally as the event accident occur in region and reported annually. Data are collected at district level and compiled up to the regional level for each year. Panel data (also known as longitudinal or crosssectional time-series data) is a dataset in which the behavior of entities is observed across time [13].
The panel data analysis was used to allow for controlling variables which cannot be observed such as variables that change over time but not across areas such as regions in these cases.
However, the property for time independence permits a possible weakness of the serial correlation of residuals in the model specification that needs further investigation. Nevertheless, this model is used as a benchmark model: where, λ it is the expected number of accidents to occur during year t in region i, DangerD is the number of accidents due to dangerous driving, DriverNeg is the number of accident due to driver negligence, Cardef is for car defect, and other variables fire, cyclist, motorcycle, over speed, livestock collection, rumination, road blocks, bad roads, railway crossing, overtake, pedestrian, passenger, pulling devices on road and alcohol while TIME is the time trend to account for the differences in accident occurrence across year and Fixed effects Poisson model: For the panel count data for i region, each region observed a total of t times [15] and. Let η it be the count variable for individual i at time t. Then the expected value of η it is linked to regressors by: x it is a vector of regressors. According to [13] the conditional maximum likelihood method is used to estimate the model parameters As y it follows the Poisson distribution, the sum of accidents: The parameters were estimated by obtaining the joint distribution of (y i1 , …, y iT ) conditional on their sum [13]. The random effect Poisson model is that unlike the fixed effects Poisson model, the variation across entities is assumed to be random and uncorrelated with predictor or independent variables included in the model. The Poisson fixed effect model was considered the best model in this study.

Trend Analysis of the Factors for Road Accidents
There are 15 factors that contribute highly to the road accidents. These factors range from Road design engineering to social-economical activities on the roads.
The trends analysis of the roads accidents in Figure 1 shows The critical factors that contributed to the road accidents are shown in Figure   2. These factors can be discussed into three major groups. The group was categorized according to the percentage causal of road traffic accidents. The group was highly contributors with higher than 10 percent cause of road accidents, the medium with less than 10 to one percent and the low group with less than one percent. The highly contributing factors include negligence of the driver (23.74%), negligence of motorcyclist (21.1%) and dangerous driving (12.83%).
The second group of causal for road accidents includes driving with high speed

Number of accidents
Year Journal of Transportation Technologies accident cases reported were caused by Motorcycle and private cars [4].
In order to achieve at least 50% reduction of road accidents as per 2 nd Decade of Action for Road Safety 2021-2030 each factor listed as causative cannot be neglected in a road safety plan.
The distribution of road accidents by regions is presented in Figure 3. The results indicated that Kinondoni has higher cumulated road accidents than any    It should be noted that Dar es Salam includes Iliala Temeke and Kinondoni followed by Pwani, then Morogoro and Arusha. The road accident for school bus that happens in Arusha 2017 creates an argument that beyond human suffering, road traffic deaths and injuries impose significant economic and financial losses to individuals and to societies. [16] claim that an increase in economic performance is associated with improvement of road transportation infrastructure resulting in the increase motorisation. Therefore, more motorization resulted in more road accidents.   (Table 3). Livestock collection, Rumination, Road blocks, pulling over devices roads and time trend with exponential estimate of 0.9686463, 0.9890562, 0.9961342, 0.9985291 and 0.9865246 respectively will decrease the number of road accidents by 0.0318558%, 0.0110041%, 0.0038733%, 0.001472% and 0.2136283% respectively. Generally, the results show that fire, alcoholism and overtake contribute most to the occurrence of road accidents.

Discussion
The aims of this study were to assess factors for occurrence of road accidents in Tanzania using panel data analysis in order to inform decision-making and policy-makers for appropriate planning and control measures against road accidents. The number of road traffic accidents drastically reduced from 23,842 in 2013 to 1634 in 2020 probably due to the increased enforcement of traffic regulations, speed limit on road, the installation of roads signs as well as innovation of traffic light for pedestrian crossing and the presence of traffic police officers on the roads as well as intensified road safety campaigns by the national road safety council [19]. Similar study [17]  The traditional approach to reducing accidents at crossing intersections has been handled by road sign or by placing zebra crossing or by speed hampers.
Even though these signs are commonly used in enhancement of the pedestrian crossing required the use of engineering technology to enhance effectiveness of roads safety.
Despite installing road safety in some black sport areas Motorcycle known as Bodaboda still remain careless and caused more accidents on crossing section.
Since 2013 there is a high increase in Motorcyclist accidents in the country.
Bad roads are all about road geometry. This factor is mostly caused by poor design, absence of road sign, plating big trees nearby the roads, absence of sideways. Bad roads are recorded on road surface or pavement distress which include cracking, patching and potholes, surface deformation, and surface deflect [24]. The author argued that it is not intelligent to roads designer engineers to design good paved road but which kills. Road should save the purpose of transportation communication but also should save people's life. Road width that is too narrow or doesn't comply with standards, sharp curves, steep downhill and uphill, pavement surface damage, and non-illuminated roads are a source of road accidents [24].
Alcoholism is the most dangerous drink that contributes highly to road accidents. Many drivers do use some energy busters which are associated with alcohol. Either some drive long days with no rest. Several studies conclude that not necessary alcohol that may course road death but also dizziness which may be due to lack of good sleeping [25]. The current innovation of alocohl detector explored at Dar es Salaam Institute of Technology can be among the counter measure of alcoholic check before buses take off at any bus stand.

Conclusions and Recommendations
Based on the findings it is concluded that there is a promising reduction of road accidents between 2015 and 2020. With this reduction, Tanzania can reach the target of reducing road accidents by 50% by the year 2030 when control measures against factors which contribute more to road traffic accidents are effectively implemented. The study points out the factors leading to road traffic accidents in Tanzania. In general, prevention strategies in Tanzania have been mainly implemented for vehicle occupants, whereas little attention has been paid Journal of Transportation Technologies to other road users. However, road accidents caused by Dangerous driving, Driver negligence, Car defect, Fire, Cyclist, Motorcycle, over speed, bad roads, Overtake and Alcoholism remain major public health concerns. Laws and regulatory frameworks should be formulated and enforced promptly to avoid losses caused by the occurrence of the road accidents. Road traffic accidents require a collaborative approach from different sectors so as to address this public health concern in a holistic manner. Pedestrian walkways and pedal-cyclist lanes should be factored in road design. Road safety professionals should be trained, to monitor the magnitude, severity and burden resulting from road traffic accidents in Tan-zania.
Based on the conclusion, the study recommends a need to strengthen the technical pillars of the safe system (infrastructure, vehicles, road users, postcrash care and modal shift) in a holistic way. The study also emphasizes the role of road safety governance (priority setting, coordination, financing, human resources as well as knowledge management) in implementation process of road safety. The road safety program led by IRAP should bring and hold everything together through a cycle of continuous improvement. Finally, more research on prediction model for ascertaining the reduction of road accidents to half from 2021-2030 needs to be carried out.