See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/303104387

Spatial characteristics and public health consequences of road traffic injuries in

Romania

Article  in  Environmental engineering and management journal · November 2015

DOI: 10.30638/eemj.2015.286

CITATIONS READS
4 205

2 authors:

Silviu Marian Ciobanu József Benedek

Babeş-Bolyai University 1. Babeş-Bolyai University; 2. Miskolc University
5 PUBLICATIONS   24 CITATIONS    53 PUBLICATIONS   122 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Socio-economic and Political Responses to Regional Polarisation in Central und Eastern Europe – RegPol² View project

The Safety of Transnational Imported Second-Hand Cars: A Case Study on Vehicle-to-Vehicle Crashes in Romania View project

All content following this page was uploaded by József Benedek on 14 May 2016.

The user has requested enhancement of the downloaded file.

Environmental Engineering and Management Journal November 2015, Vol.14, No. 11, 2689-2702
http://omicron.ch.tuiasi.ro/EEMJ/

“Gheorghe Asachi” Technical University of Iasi, Romania

SPATIAL CHARACTERISTICS AND PUBLIC HEALTH

CONSEQUENCES OF ROAD TRAFFIC INJURIES IN ROMANIA

Silviu Marian Ciobanu, Jozsef Benedek

Babes-Bolyai University, Faculty of Geography, 5-7 Clinicilor, 400 006 Cluj-Napoca, Romania

Abstract

Romania has recorded a dramatic number of road traffic accidents in the last five years. The latest European documents and
statistics indicate that the country has the highest number of traffic fatalities across the Member States of the European Union.
The aim of the paper is twofold. First, the authors aim to analyze the spatial characteristics of traffic accidents for the period
2010-2014: spatial distribution of accidents at county level, number of vehicles involved, socio-spatial distribution of accidents,
and correlation between the urbanization level and accidents. Second, the intent is to calculate various indicators expressing the
severity of traffic accidents. The paper concludes with some remarks related to the consequences of road accident injuries on
public health.
GIS (Geographic Information System) techniques are employed in the paper, using official data sets provided by the General
Inspectorate of Romanian Police and the National Institute of Statistics.
The main results of the paper are: a) a set of maps and charts indicating the distribution characteristics and the crash
concentration areas, located along the major road connections and in the main urban areas; b) a chart of the social vulnerability,
indicating that the most frequently involved social groups are males aged 18-26, while the other age groups and females are less
involved in accidents; c) a set of maps and charts regarding the impact of accidents on health, indicating that about half of the
accidents have minor impact on health, the other half being represented by severe injuries and fatalities.
One of the main conclusions of the study is that during the last five years the traffic accident statistics in Romania has not
significantly improved. Therefore, traffic accidents represent the main risk and their reduction is one of the targets for the
transport and road safety policies.

Key words: GIS, public health, road safety, traffic accidents

Received: January, 2015; Revised final: October, 2015; Accepted: November, 2015

1. Introduction or serious crash on our roads is a tragedy. It is our

Road traffic accidents had, over time, a
significant negative impact on public health, and the
term “road safety” is directly associated to the
mitigation of this impact. A comprehensive long-
term strategy (2010-2020) on accident prevention
and improvement of road safety is implemented in
the European Union (EU). Approximately 25,700
fatalities were recorded last year (2014) in the EU as
a consequence of traffic accidents and more than
200,000 persons were seriously injured, and most
often their lives have radically changed. “Every fatal
moral obligation – our shared responsibility – to take
road safety seriously.” (European Commission,
2015).
The 2010 policies on road safety proved to be
very effective in improving road safety in EU
member states, as very good results were achieved in
reducing fatal accidents. The number of reported
fatalities per million inhabitants is the main
measurement unit for road safety. According to the
latest documents published by the European
Commission (EC), the fatal accidents rate in the EU
has constantly decreased over the last 5 years. As


Author to whom all correspondence should be addressed: e-mail: jozsef@geografie.ubbcluj.ro; Phone: +40264.592.214
 Ciobanu and Benedek/Environmental Engineering and Management Journal 14 (2015), 11, 2689-2702
shown in Fig. 1, the number of fatalities decreased by
5,700 in 2014 compared to 2010 (-18%), while
compared to 2013 the decline was very slow (1%),
far from the predicted evolution.
The EU average value of road fatality rate in
2014 was 51 dead per million inhabitants; the lowest
values were recorded in Sweden, Netherlands, UK
(less than 30 dead per million inhabitants). By
contrast, countries like Romania, Bulgaria, Latvia,
Lithuania and Poland have a road fatality rate of 80
dead per million inhabitants (European Commission,
2015).
In Romania, the European descending trend of
road fatality rate continued over the last 5 years
(2010-2014). In 2010, approximately 115 fatalities
per million were recorded, the highest road fatality
rate recorded in an EU Member State. This decreased
in 2014 to 91 fatalities per million inhabitants, still
placing Romania in a lagging position, outrun only
by Latvia with a fatality rate of 105 dead per million
inhabitants. According to the General Inspectorate of
Romanian Police, Road Department (GIRPRD),
approximately 130,000 accidents were recorded over
the last 5 years (2010-2014) resulting in 174,028
casualties of which 121,496 slightly injured, 42,416
seriously injured and 10,116 dead.
Table 1 presents the significant impact of
traffic accidents on public health in Romania,
causing tens of thousands of casualties every year.
Therefore, the need to improve road safety is of
utmost importance. In the last 5 years, a significant
decrease in the number of road traffic fatalities was
recorded (559 fewer dead in 2014 compared to
Fig. 1. Real vs. predicted evolution of fatal accidents (diagram adapted after European Commission, 2015)
Table 1. Traffic accidents statistics in Romania (2010-2014)
Year Accidents Cars Participants Slightly injured
2010 25,996 38,553 59,426
2011 26,648 39,674 59,992
2012 26,929 40,083 60,877
2013 24,828 37,121 56,701
2014 25,356 38,058 58,085
2690
2010), but the situation in Romania remains dramatic
compared to the average value in EU.
Road safety depends on several factors,
represented by a series of both quantitative and
qualitative data such as the number of traffic
accidents, traffic volume, road density, number of
lanes, lane width or curves density (Francello et al.,
2015). The authors intend to show how the traffic
accidents influence road safety in Romania and
which are their consequences and impact on public
health and health safety system (Marhavilas et al.,
2014). For this purpose, in the first part of the paper
the authors intend to identify the spatial
characteristics of traffic accidents occurred between
2010-2014, their socio-spatial distribution, the spatial
distribution of casualties and the number of vehicles
involved, as well as to outline a correlation at
national level between the number of accidents and
the urbanization level (Benedek and Kurkó, 2012).
The second part of the paper focuses on calculating
the gravity of several indicators of traffic accidents
such as injury rate, fatality rate or hazard rate, to
highlight their significant impact on public health in
each Romanian county. The authors also develop
here the vulnerability pyramid of social groups to the
occurrence of traffic accidents, but also the black
pyramids to highlight the social groups of the
victims. To achieve the proposed objectives the
authors used official data bases provided by GIRPRD
and by the National Institute of Statistics (NIS), as
well as GIS (Geographic Information System)
techniques available by ArcGis 10.1, and statistical
techniques implemented by means of Microsoft
Office 2010.
Severely injured Dead Total Casualties
23,905 8,509 2,377 34,791
24,723 8,768 2,018 35,509
25,349 8,860 2,042 36,251
23,306 8,158 1,861 33,325
24,213 8,121 1,818 34,152
 Spatial characteristics and public health consequences of road traffic injuries in Romania
2. Materials and methods
2.1. Data
Statistics experts state that such a study cannot
have credible results unless it uses records of
accidents over a period of at least 3 years
(Steenberghen et al., 2010). In this paper, the official
records of traffic accidents in Romania for a period
of 5 years (2010-2014) were used. This database was
provided by GIRPRD (the only institution in
Romania officially dealing with monitoring and
recording of traffic accidents) and contains a wide
range of information for each particular accident:
date and time, location (GPS coordinates), road
category, composition of the road, slope of the road,
number of lanes, lane width, weather conditions,
lighting conditions, number of vehicles involved,
number of minor injuries, number of seriously
injured, number of fatalities, details on the persons
involved (age, gender, nationality, alcohol,
profession) and details of vehicles involved (type,
year of manufacture, trade mark, technical condition,
and mileage). A total of 129,757 records were used,
but it should be noted that GIRPRD decided to store
the GPS coordinates for each accident starting with
2011, and the records which lacked the coordinates
were excluded from this point-type analysis.
Moreover, due to technical faults occurred in the
logistics available for GIRPRD, there were some
incorrect coordinates that also led to the removal of
those records from the analysis. A special case is that
of accidents recorded in Bacău and Buzău, where, for
unknown reasons, the GPS coordinates are missing in
99% of the cases. Another thing worth mentioning is
that not all traffic accidents are recorded by
GIRPRD, but only those of undeniable gravity,
where police intervention is needed.
Before starting this analysis, the data was
tested. More precisely, the authors wanted to see if
they can use only the accident records with the
correct coordinates, without influencing the results.
The database was divided into: accidents with
incorrect coordinates, accidents with missing
coordinates and accidents with correct coordinates.
Each of the 3 data sets was correlated with each other
according to 3 key parameters: number of slightly
injured, number of seriously injured, death toll, main
cause of accident, how the accident occurred,
participant category and participant guilt.
The linear regression model and Pearson
correlation coefficient were used for this, and the
results were positive, the correlations being high.
Therefore, it was possible to use only the data set
with the correct coordinates. It is also worth
mentioning that the correct coordinates data set was
used only in point analysis (calculation of hot spots
or general view of accident distribution). The rest of
analysis (accident rate, injury rate, fatality rate,
hazard rate and urbanization level) used all 129,757
records, because geographic coordinates were not
necessary in these cases.
Another data source was NIS (2015), which
provided statistical data about the Romanian
population (2014), namely, the number of inhabitants
for each county, as well as the rural urban population
separately.
2.2. Methodology and software
GIS has greatly developed in the last years in
the research field of traffic accidents. Some authors
use GIS for spatio-temporal analyses of road traffic
accidents (Ivan and Haidu, 2012, Plug et al., 2011;
Prasannakumar et al., 2011), others use it for simple
linear analyses and generation of maps by combining
various parameters such as number of accidents,
death toll and number of injured people or property
damage (Gundogdu, 2010; Pulugurtha et al., 2007;
Qin et al., 2013; Truong and Somenahalli, 2011),
others for spatial analyses and determination of risk
areas (Aguero-Valverde and Jovanis, 2006; Erdogan,
2009; Ivan et al., 2015; Lee et al., 2015, Yao et al.,
2015), while others use GIS to develop emergency
systems (Jones and Bentham, 1995; Rui et al., 2015).
This study uses GIS tools for spatial analyses
and representation provided by the ArcGIS 10.1
platform. The first tool used is Kernel Density
Estimation. The Kernel density calculates the
characteristics density of a value in a particular
location, as well as its surroundings. This function
can only be applied to primitive diagrams such as
points and lines whose attribute is numerical data. In
this case the authors used points represented by road
traffic accidents. The Kernel density is often used in
determining traffic accidents hotspots (Anderson,
2009; Bil et al., 2013; Xie and Yan, 2008, 2013).
This method is used in this paper to highlight the
spatial distribution of road traffic accident fatalities
and the number of vehicles involved.
Another GIS tool used in the paper is the
Geographically Weighted Regression (GWR), which
creates correlations (linear regressions) between
different parameters and represents their cartographic
outcome (Fotheringham et al., 1998, 2002). In this
case, this tool facilitated the display of spatial
distribution in the Romanian counties of the
correlation between the Urbanization Level (UL-
percentage of urban population in the total
population) and the Accident Rate (AR). The
intensity of correlation between the two variables
(UL and AR) was found by calculating the Pearson
correlation coefficient using the formula (Lee and
Wong, 2001) (Eq. 1):

n
xiyi
i 1
XY
r n
SxSy (1)
2691
 Ciobanu and Benedek/Environmental Engineering and Management Journal 14 (2015), 11, 2689-2702

where: r is the Pearson correlation coefficient (Wan the persons guilty for the occurrence of the accident,
et al., 2014), X and Y represent the average of x and taking into account their age, gender, category
(driver, pedestrian) and blood alcohol level. The
y S x while S y represents the standard deviation of x authors calculated the percentage for each and
divided them into 10 groups by gender and age
and y, calculated by the formulas (Eqs. 2-3) (Lee and (every 9 years), starting and ending with the 0-9 age
Wong, 2001): group and ending with the 80+ age group. By the
same token, the black pyramids were developed,
 while this time only the casualties of accidents were
n
x2
Sx  n 1
X2 taken into account, with the following parameters:
n (2)
number of slightly injured, number of seriously
injured and death toll.


n
y2 3. Results and discussion
Sy  n 1
Y 2
n (3) 3.1. Spatial distribution of accidents

To highlight the gravity of road accidents and Over the analysis period (2010-2014)
their impact on public health in Romania, the Romania recorded a slight decrease (2.5%) in the
calculation of 4 statistical indicators was used: number of traffic accidents (Table 1). It is worth
accident rate (AR), injury rate (IR), fatality rate (FR) mentioning that in this article both accidents between
and hazard rate (HR) by Eqs. (4-7) (Pirotti et al., vehicles and the vehicle – pedestrian accidents were
2011): taken into account.
By far, Bucharest has the highest accident rate
Ac (14.04%), with no less than 18,691 accidents, well
AR  * 100 (4) above the county average, i.e. 3,089 accidents.
At Considering an overview of accidents per county
(Fig. 2) one can say that the highest AR belongs to
where Ac is the number of accidents occurred in a Constanta County (4.9%), followed at a distance by
given area (in this case at county level) and At is the the percentage of Prahova County and Cluj County
total number of accidents (in this case, the total with a rate of 3.6%. AR values above average (2.3%)
number of accidents in the country); are also found in Iași, Suceava, Argeș, Ilfov, Brașov,
Dolj, Bacău and Timiș Counties. The opposite is
I found in Brăila, Ialomița, and Satu Mare at a rate of
IR  * 100 (5) about one percent of all accidents; the most favorable
At situation is in Covasna County, with AR = 0.7%.

where I is the number of injured persons in a given 3.1.1. Vehicles

area.
D vehicles/accident, but in some cases there were pile-
FR  *100 (6)
At
where D is the number of dead in a given area.
D bicycle etc.). Regarding their evolution over the last
HR  *100 (7)
DI
The result of these indicators was also mapped
using the ArcGIS 10.1 platform.
Another part of the methodology worth
mentioning is the outline of the pyramids of social
groups vulnerable to the production of traffic
accidents and the development of the black pyramids.
The notion of “pyramid” was taken up from
demography (Dannefer and Bahata, 2015; Hobbs,
2004; Horiuchi, 2015), more exactly from the
stratification of population on age and gender. To
build the pyramids of social groups vulnerable to the
production of traffic accidents, the authors used only
The analyzed accidents involved a number of
193,589 vehicles resulting in an average of 1.5
ups involving 3-7 vehicles and, in some extreme
situations, even 8-11 vehicles (Fig. 3). It should be
mentioned that this study took into account all types
of vehicles, be they motorized (car, truck, dumper,
tractor, bus, etc.) or using other type of traction (cart,
five years the authors conclude that the maximum
number of vehicles involved in traffic accidents was
40,083 vehicles in 2012, but the trend is decreasing,
as in 2014 the number reached 38,058 vehicles
(Table 1).
Similar to the situation concerning the number
of accidents, Bucharest ranks first in the number of
vehicles involved (28,357 vehicles), with an average
of 1.55 vehicles/accident. Applying the Kernel
density (with a 100 m radius) to the accidents that
occurred in the capital, considering the parameter
“number of cars involved”, the authors have noticed
(Fig. 3) the formation of numerous groups (hotspots)
on the main thoroughfares, especially in major
crossroads.

2692
 Spatial characteristics and public health consequences of road traffic injuries in Romania

The same trend as for AR can be noticed at Cluj counties (7,753 and 7,529 vehicles involved,
county level (Fig. 3). Constanța county has the respectively). The lowest values are found in Tulcea,
largest number of vehicles involved in traffic Brăila, Ialomița, Sălaj, Covasna and Satu Mare
accidents (9,910 vehicles), followed by Prahova and counties, with less than 2,250 vehicles.

Fig. 2. The general spatial distribution of traffic accidents (2010-2014) in Romanian counties

Fig. 3. The spatial distribution of the number of vehicles involved in traffic accidents in Romania

2693
 Ciobanu and Benedek/Environmental Engineering and Management Journal 14 (2015), 11, 2689-2702
3.1.2. Participants – nationality and profession
The approximately 130,000 traffic accidents
involved 295,081 persons - drivers, passengers,
passengers, pedestrians or cyclists (Table 2). While
the number of drivers, passengers and pedestrians is
slightly decreasing in the last years, the opposite can
be said about cyclists, as these were increasingly
involved in traffic incidents (dramatic increase by
66% in 2014 compared to 2010). Although the
number of pedestrians involved in traffic accidents
decreased by 6% over the last 5 years, Romania
ranks first in pedestrian fatalities among the EU
member state.
Regarding the nationality of the participants,
besides Romanians, other 103 nationalities are found,
totaling 5,715 foreign nationals. Their main countries
of origin can be easily noticed in the chart below
(Fig. 4).
The most numerous are the Bulgarians (675),
followed by Moldovans (579), Turkish nationals
(501), Hungarians (479) Germans (463), Italians
(438), Ukrainians (235) and Polish (229). Since these
are the 8 nationalities most frequently involved in the
analyzed accidents, they were the only one used in
this spatial distribution analysis. The locations of
accidents involving foreign nationals follow mostly
special patterns for each nationality (Fig. 5).
Table 2. The category of participants involved in traffic accidents occurring in Romania (2010-2014)
Year Drivers Passengers
2010 35,498 11,279
2011 35,860 11,234
2012 36,026 11,512
2013 33,186 10,970
2014 33,788 11,068
Fig. 4. Number of foreign nationals involved in traffic accidents in Romania (2010-2014)
2694
They generally occurred along the main
transit axes (European roads). Such accidents
involving Bulgarians occurred mainly on two transit
axes connecting EU to Bulgaria. As can be seen in
Fig. 5, the two axes are: Cenad Customs - DN6 –
Timișoara – E70 – Orșova – E70 – Craiova- E70 –
Alexandria – E70- Zimnicea (border crossing point
harbor) and Nădlac Customs – E68 – Arad – E68 -
Deva – A1 – Sibiu – E81 – Râmnicu Vâlcea – E81-
Pitești – A1 – Bucharest – E85 – Giurgiu (Ruse). The
first route will be further referred to as AXIS 1, and
the second as AXIS 2. Most of the accidents
involving Turkish nationals occurred in Bucharest,
where the largest Turkish community in the country
is found. The careful analysis of Fig. 5 revealed that
many accidents involving Turkish citizens occurred
also on AXIS 2, but also on the Nădlac Customs –
E68 – Arad – E68 - Deva – A1 – Sibiu- E68- Brașov-
E60 – Ploiești- A3 – Bucharest – E85 – Giurgiu
section (AXIS 3). The distribution of accidents
involving persons of Hungarian nationality follows a
well-defined pattern, i.e. on the route from Borș
Customs to Szeklerland: Borș Customs – E60 –
Oradea – E60 – Cluj Napoca – E60 – Târgu Mureș -
E60 - DN13A - Praid- DN13A - Odorheiul Secuiesc
– DN13A – Miercurea Ciuc – E578 – Sfântu
Gheorghe (AXIS 4).
Pedestrians Cyclists Others
10,179 2,080 390
9,943 2,525 430
9,777 2,886 676
9,272 2,824 449
9,594 3,141 494
 Spatial characteristics and public health consequences of road traffic injuries in Romania

Fig. 5. The spatial distribution of accidents in Romania (2010-2014) involving the main foreign nationalities

The distribution of accidents involving

Italians is revealed on the roads connecting the
border crossing points with the main cities
(Bucharest, Cluj-Napoca, Timișoara, Craiova, and
Sibiu).
Accidents involving Moldovan citizens are a
special case, as these are very frequent on the
Hungary - Republic of Moldova transit axis (Borș
Customs– E60 – Oradea – E60 – Cluj Napoca – E576
– Dej – E58 – Bistrița –E58 – Vatra Dornei – E58 –
Suceava – E85 – E583 – Iași – E583 – Sculeni
Customs) and on the various national and county
roads, especially in the counties near the border with
the Republic of Moldova, respectively Galați, Vaslui,
Iaşi, Neamţ, Botoşani and Suceava.
The analysis of occupations of persons
involved in road accidents revealed that 18.6% are
unemployed, and 10.2% are pensioners (Fig. 6). The
alarming fact is that a number of about 30,000 (10%)
of the participants to traffic incidents (most often Fig. 6. Occupations of persons involved in traffic accidents
accidents involving pedestrians) are students. This is in Romania (2010-2014)
all the more serious knowing that their number is
double compared to the persons whose occupation 3.2. Correlation between UL and AR
implies being in traffic all the time, namely the
professional drivers. The number of these persons is As mentioned above, another objective of this
about 14,000. paper is to discover the statistical link between UL
5% of the persons involved in traffic accidents and AR in the Romanian counties. In other words,
are farmers, engineers, economists, teachers and the authors wanted to see to what extent UL
employees of the Ministry of Internal Affairs (MAI), influences the number of traffic accidents. Simple
while 3,700 persons are official employees of the linear regression and calculation of the Pearson
Ministry of National Defense (MAN), employees of correlation coefficient were used for this purpose.
the Romanian Intelligence Service (SRI), diplomatic But in order to apply these two methods, the first step
personnel, penitentiary personnel, church employees was to calculate the UL values for each county, as the
and unemployed. AR variable was already calculated in section 4.

2695
 Ciobanu and Benedek/Environmental Engineering and Management Journal 14 (2015), 11, 2689-2702
Thus, the highest UL values were recorded in
Hunedoara, Braşov, Constanţa, Sibiu, Cluj, Timiș,
Alba, Arad, Galați, Caraș-Severin and Dolj counties,
where the urban population accounts for over 50% of
the total county population. The final result shows a
direct linear relationship between these two variables,
so that AR increases with the increase of UL (Fig. 7).
Fig. 7. The regression line between UL and AR
Since the coefficient of determination (R2) is
0.42, the authors conclude that 42% of accidents may
be explained by the UL.
The Pearson correlation coefficient was used
to see the intensity of this linear dependency between
the UL and AR variables. As shown in Fig. 7, the
coefficient has the value of 0.65 (positive value in the
range of 0.5 - 0.75), which indicates that AR is in
direct proportion to UL and the intensity of linear
dependency between them is strong.
To view this spatial correlation in Romania,
the authors used the GWR tool implemented by the
ArcMap 10.1 software. The result is a map (Fig. 8)
which indicates the correlation degree for each
county, divided into 5 distinct classes: very low, low,
medium, high and very high.
The extremes of these classes are represented
by Brăila county (very low, UL does not influence
AR) and Bucharest (very high, UL influences AR to
a great extent). The explanation is simple, Bucharest
has UL of 100% and has the highest AR in Romania
(14.04%), while Brăila county, although it has very
good UL (62.16%, 6th in the country) the number of
accidents is very low, with an AR reaching 1.09%.
The “low” category includes 15 counties, among
which only the following are worth mentioning:
Sibiu, Alba, Galați and Caraș – Severin, which have
a high UL (over 53%) and low AR (less than 2.2%
for Galaţi or below 1.4% for Caraș-Severin). There
are 9 counties that have high correlation between UL
and AR, following 3 models: either a high UL and
AR (Cluj and Prahova Counties) or an average UL
and average AR (Iași, Argeș, Bacău and Suceava), or
a low UL and low AR (Buzău, Giurgiu and
Dâmbovița).
The difference in the correlation coefficient of
counties is mainly explained by the urbanization
2696
level of the counties. Higher urbanization implies
higher economic development and higher spatial
mobility, which, combined with the low level of
transport infrastructure development, determines
higher accident rates. The authors are aware that
there are other causal factors involved as well, such
as the regional differences in driver behavior, but
they are out of the scope of this paper.
3.3 Social groups vulnerable to the occurrence of
traffic accidents
Another important objective of this paper is to
establish the traffic accident-prone social groups
divided by gender and age. To achieve this goal the
authors have analyzed only those persons guilty for
the occurrence of accidents and have applied the
social groups pyramid method. In this case, in
addition to accidents in the 2010-2014 period,
accidents occurring in January and February 2015
were also included. From the nearly 300,000 persons
involved in traffic incidents, 140,022 were found
guilty of their occurrence, 72% drivers, 20.5%
pedestrians, 7.5% other categories (passengers,
driving instructors or travelers). It is worth noting
that 11% of guilty persons were under the influence
of alcohol when the accident occurred.
The results include 4 pyramids of social
groups prone to the occurrence of traffic accidents
(Fig. 9): one for all guilty persons, one for drivers,
one for pedestrians and one for the guilty persons
under the influence of alcohol. A general look on the
total number of convicted persons pyramid (Fig. 9A)
reveals that the men/women ratio is about 80:20. It
can also be noticed that the “affected” group is the
male 18-26 age group (18.1%), followed by the male
27-35 age group (16.31%) and by the male 36-44 age
group. For women, the same 3 age groups are most
prone to accidents, but the percentages are much
lower, 3.5%, 3.4% and 2.8%, respectively.
In case of the alcohol pyramid (that of the
persons who have consumed alcohol and caused
traffic accidents), the men/women ratio is
overwhelmingly in favor of men, respectively 97:3
(Fig. 9B). Once again, the same 3 age groups (18-26,
27-35 and 36-44) are dominated by men, but the 45-
53 and 54-62 groups are not negligible either, with
more than 10 percents, while for women, only the 36-
44 group is significant (about half of the women who
caused traffic accidents under the influence of
alcohol fell in this group).
Regarding the drivers pyramid, the
men/women ratio is 87:13, and the highest value is
found in the age group of men aged 18-26,
cumulating over 22% of the guilty drivers, followed
by the 27-35 group, with a percentage of more than
20 % (Fig. 9D).
The situation of pedestrians pyramid (Fig. 9C)
is completely different, as the men/women ratio is
more balanced, i.e. 58.2:41.8. The most vulnerable
groups, both for men and for women are minors in
the 0-8 and 9-17 groups.
 Spatial characteristics and public health consequences of road traffic injuries in Romania

Fig. 8. The county distribution of the correlation between UL and AR

Fig. 9. The age and gender pyramid for A) total of guilty persons B) guilty persons under the influence of alcohol C) guilty
pedestrians D) guilty drivers

2697
 Ciobanu and Benedek/Environmental Engineering and Management Journal 14 (2015), 11, 2689-2702
For the male gender, the highest value is for
the 0-8 group (9.5%), while for the female gender it
is the 9-17 group (7.3%). It should also be mentioned
that the differences between the age groups in both
genders are not so large, except for the minor groups.
3.4. Impact of road accidents on public health
Each traffic accident resulting in casualties is
a stain on public health. In the analyzed period
(2010-2014) there were no less than 174,028
casualties, of which 70% slightly injured, 24%
seriously injured and 6% dead. To represent the
impact of road traffic accidents on public health in
Romania spatially the authors have calculated these
hotspots in terms of number of casualties. This was
achieved by applying the Kernel density with a 2.5
km search radius, implemented by means of the
ArcMap GIS 10.1 software. The results reveal
several hotspots located mainly in urban areas and
along the main transport routes (Fig. 10).
The most visible hotspots of the casualties are
located in the area of Bucharest, Constanța and Cluj-
Napoca cities, but also in Târgu Mureș, Oradea,
Timișoara, Deva, Hunedoara, Râmnicu-Vâlcea,
Craiova, Brașov, Sibiu, Focșani, Galați, Piatra
Neamț, Roman, and Iași. When examining the
hotspots along the main transport axes, it can be
noticed that they are located mainly on DN1
(Bucharest – Ploiești – Brașov section), E85
(Bucharest - Urziceni section), E85 (Adjud – Focșani
– Râmnicu Sărat section), E81 (Sibiu - Râmnicu
Vâlcea section), E81 (Turda – Alba Iulia), E60 (Cluj-
Napoca – Târgu Mureș), E576 (Cluj-Napoca – Dej),
E79 (Deva – Hațeg – Petroșani).
By far, the most delicate situation occurs in
Bucharest, where there have been more than 21,346
casualties (16,860 slightly injured, 4,079 seriously
injured and 407 dead) over the last 5 years. A
dramatic situation is that of Constanța municipality,
where road accidents have a significant impact on
public health because this is the city where the largest
number of fatalities was recorded (except Bucharest),
totaling a number of 127. Another city worth noting
in terms of number of casualties is Brașov, which,
during the analysis period has cumulated a total of
2,401 casualties, of which 1,706 slightly injured, 617
seriously injured and 78 dead. Cluj-Napoca is ranked
4th among the cities with the highest number of
casualties in Romania.
3.4.1. Injury rate, fatality rate and hazard ratio
To highlight the severity of traffic accidents
and to emphasize their impact on public health in
Romania, the authors calculated three statistical
indicators: IR, FR and HR. The findings were
mapped at county level using the ArcMap 10.1
software (Fig. 11).
IR represents the percentage of injured
persons in a certain area (in this case, the counties) in
relation to the total number of accidents occurred at
2698
national level. The careful examination of Fig. 11
revealed that the most favorable situation was in Satu
Mare, Sălaj, Covasna, Brăila, Tulcea, Ialomița and
Giurgiu counties, where IR ranged between 0.94 and
1.54. The opposite was found in Prahova and
Constanța (4.9 and 6.2, respectively), but the highest
value of IR was recorded in Bucharest (16.1).
FR represents the percentage of persons killed
in road accidents per total number of accidents in a
particular area. In Romania, the counties with the
highest FR are Constanța, Ilfov, Prahova and
Suceava (with FR values between 0.28 and 0.32).
The opposite is found in Botoșani, Tulcea, Covasna
and Sălaj, where FR has minimum values, due
primarily to the low death toll (127, 103, 90, and 115,
respectively).
HR is calculated relating the death toll to the
sum of the death toll and the number of injured
persons in a given area (all in percentages). In other
words, it calculates the percentage of dead from the
total number of casualties. Fig. 11 illustrates low
values of this indicator in Bucharest and in Tulcea,
Constanța, Argeș, Iași and Maramureș counties,
where, of all injured persons, dead are less than 5%.
The highest HR values are in Satu Mare and Ialomița
counties, with a percentage exceeding the 8%
threshold of all casualties.
3.4.2. The black pyramids
The black pyramids are used to emphasize the
social groups divided by gender and age, whose
health has been affected by accidents. The authors
believe that the black pyramid is an excellent tool to
divide the impact of traffic accidents on public health
in terms of various social groups.
To achieve this last objective the authors have
considered all types of casualties caused by accidents
(slightly injured, seriously injured and dead) and
have applied the same method as in the case of social
groups vulnerable to the occurrence of road
accidents, with the difference that each group is no
longer calculated in percentages, but the absolute
value (total number of persons for each group) is
used instead. It should be mentioned that data from
January and February 2015 were used this time. Four
pyramids were generated, one for the total number of
casualties, one for the slightly injured, one for the
seriously injured and one for the dead (Fig. 12).
The pyramid of the total number of casualties
(Fig. 12A) is a uniform, synchronized one for most
age groups, with a men/women ratio of 62:38 and the
age group most affected by the occurrence of
accidents is the 18-26 group, which includes no less
than 38,233 persons (24,611 men and 13,622
women). As people get older, the number of
casualties decreases gradually, reaching the lowest
value (3,117) in the 80+ group.
The shape of the slightly injured pyramid
(Fig. 12B) is very similar to that of the total
casualties, as the injured persons represent 70% of all
casualties. The men/women ratio is lower this time,
 Spatial characteristics and public health consequences of road traffic injuries in Romania

i.e. 60:40 for men, while the 18-26 age group is still
the most affected by traffic accidents (17,821 men
and 11,009 women). Regarding the seriously injured
pyramid (Fig. 12C), the authors consider that it does
not have a regular shape (the number of casualties
does not necessarily decrease with age every time)
and the female part is not synchronized with the male
part. The most affected male group is the 18-26 age
group (a total of more than 5,000 casualties), but the
27-35 and 36-44 age groups are not to be neglected
either, as they gather on average over 4,300
casualties each. For the female part, except the
extreme groups (0-8 and 80+), there are slight
differences between them, each group totaling an
average of about 1,700 casualties. Small differences
are in the 18-26 and 54-62 age groups, which record
a little more than 2,000 casualties. It should be noted
that the men:women ratio is 65:35.

Fig. 10. The hotspots distribution of road accidents in terms of number of casualties

Fig. 11. Accident rate, injury rate, fatality rate and hazard rate in Romanian counties

2699
 Ciobanu and Benedek/Environmental Engineering and Management Journal 14 (2015), 11, 2689-2702
Fig. 12. The black pyramids of traffic accidents casualties: A) total number of casualties B) slightly injured
C) seriously injured D) dead
A very interesting situation is that of the dead
pyramid (Fig. 12D) which was built based on the
10,329 fatalities caused by accidents, of which 7,793
men compared to 2,536 women (75:25 ratio). It can
be noticed that for men, the dominant age groups are
the 54-62 and 18-26 groups, weighing on average a
total of 1,200 persons. However, the 27-35, 36-44
and 45-53 age groups are equally important, totaling
on average approximately 1,800 casualties. As for
women, the situation is quite different because, with
few exceptions (the 18-26 group and the 80+ group),
the death toll increases with age. Thus, the most
affected group is that of 72-80 years (more than 380
casualties), and the least affected is the 0-8 age group
(105 fatalities).
4. Conclusions
An impressive number of traffic accidents
occurred in Romania over the last 5 years (2010-
2014), which places our country, alongside Latvia, in
the forefront of the EU member states with the lowest
road safety. This is revealed primarily by the large
number of reported fatalities per million inhabitants,
respectively 91 fatalities/1 million inhabitants. In the
given period, both the number of accidents and the
other additional parameters (number of persons
involved, number of casualties, number of vehicles,
etc.) follow a downward trend, but a very slow one,
which should compel us to invest more time, human
resources and money in the road safety concept.
In the first part of this paper the authors have
conducted a statistical analysis and spatial
distribution of road accidents in Romania according
2700
to their characteristics (location and number, number
of vehicles involved or profession and nationality of
the persons involved). Then, the analysis of the
relationship between UL and AR was used to see to
what extent UL influences the occurrence of
accidents, and the result was mapped for the entire
country. It was shown that the two variables are
linearly dependent, and the degree of dependence is
high, as AR is in direct proportion to UL, as 43% of
road accidents are explained by UL. All these are of
particular importance as they are the starting point
for setting up the road safety policies.
Another objective achieved in this paper was
to establish the traffic accidents-prone social groups.
Therefore, it was shown that women are much less
involved in traffic accidents than men, with a ratio of
80:20, while the most vulnerable social group is the
group of male drivers aged 18-26. On the other hand,
women are more vulnerable in the pedestrian group
aged 9 to 17. To highlight the impact of road
accidents on public health in Romania the paper
employed the calculation and spatial representation
of statistical indicators AR, IR, FR, HR, the
calculation and spatial representation of accident
hotspots (depending on the number of casualties) and
introduced the term black pyramids. Results show
that most casualties were recorded in Bucharest,
Constanța, Cluj-Napoca, Brașov and on the main
transport axes of Romania, especially DN1
(Bucharest – Ploiești – Brașov section), E85
(Bucharest – Urziceni section), E85 (Adjud – Focșani
– Râmnicu Sărat section), E81 (Sibiu - Râmnicu
Vâlcea section). High values of the above-mentioned
indicators were recorded in Constanța, Prahova, Cluj,
 Spatial characteristics and public health consequences of road traffic injuries in Romania
Iași, and Suceava, while low values were found in
Satu Mare, Sălaj, Brăila, Tulcea or Giurgiu. The
black pyramids proved to be an excellent tool to
divide the impact of road accidents on public health
into different social groups. This example proved that
most of the fatalities from road accidents are in the
54-61 and 18-26 age groups for men and in the 72-80
age group for women. This study has indicated with a
high degree of precision those places (hotspots) of
vulnerability to traffic accidents where interventions
are needed to improve traffic safety. This research
provides a tool for decision makers in traffic safety to
make better informed decisions and therefore to
better tailor their specific improvement objectives.
In conclusion, such a study is an essential tool
in institutions dealing with road accident analysis and
prevention, as it is the starting point for decision-
making, to improve road safety and reduce the
impact of traffic accidents on public health.
Acknowledgements
This paper is a result of a doctoral research made possible
by the financial support of the Sectoral Operational
Programme for Human Resources Development 2007-
2013, co-financed by the European Social Fund, under the
project POSDRU/159/1.5/S/132400 - “Young successful
researchers – professional development in an international
and interdisciplinary environment”.
References
Aguero-Valverde J., Jovanis P., (2006), Spatial analysis of
fatal and injury crashes in Pennsylvania, Accident
Analysis and Prevention, 38, 618–625.
Anderson T., (2009), Kernel density estimation and K-
means clustering to profile road accident hotspots,
Accident Analysis and Prevention, 41, 359-364.
Benedek J., Kurkó I., (2012), Convergence or divergence?
The Position of Romania in the Spatial Structure of the
European Union, Transylvanian Review, 21, 116-125.
Bil M., Andrasik R., Janosca Z., (2013), Identification of
hazardous road locations of traffic accidents by means
of kernel density estimation and cluster significance
evaluation, Accident Analysis and Prevention, 55, 265-
273.
Dannefer D., Bhatta T., (2015), Age stratification,
International Encyclopedia of the Social & Behavioral
Science, 2, 333-337.
Erdogan S., (2009), Explorative spatial analysis of traffic
accident statistics and road mortality among the
provinces of Turkey, Journal of Safety Research, 40,
341–351.
European Commission, (2015), Road safety in European
Union, Trends, statistics and main Challenges,
Mobility and Transport DG, BE-1049 Brussels.
Fancello G., Carta M., Fadda P., (2015), A decision
support system for road safety analysis,
Transportation Research Procedia, 5, 201 – 210.
Fotheringham A.S., Charlton M.E., Brunsdon C., (1998),
Geographically weighted regression: a natural
evolution of the expansion method for spatial data
analysis, Environment and Planning A, 30, 1905-1927.
Fotheringham A.S., Brunsdon C., Charlton M.E., (2002),
Geographically Weighted Regression: The Analysis of
Spatially Varying Relationship, Wiley, New York.
Gundogdu I.B., (2010), Applying linear analysis methods
to GIS-supported procedures for preventing traffic
accidents: Case study of Konya, Safety Science, 48,
763–769.
Hobbs F., (2004), Age and Sex Composition, In: The
Methods and Materials of Demography, 2nd Edition,
Siegel J.S., Swanson D.A. (Eds.), Elsevier Academic
Press, San Diego and London, 125–173.
Horiuchi S., (2015), Age structure, International
Encyclopedia of the Social & Behavioral Science, 2,
338-343.
Ivan K., Haidu I., (2012), The spatio-temporal distribution
of road accidents in Cluj-Napoca, Geographia
Technica, 2, 32-38.
Ivan K., Haidu I., Benedek J., Ciobanu S.M., (2015),
Identification of road traffic accidents risk-prone under
low lighting condition, Natural Hazards and Earth
System Science, 15, 2059-2068.
Jones A.P., Bentham G., (1995), Emergency medical
service accessibility and outcome from road traffic
accidents, Public Health, 109, 169-177.
Lee J., Abdel-Aty M., Choi K., Huang H., (2015), Multi-
level hot-zone identification for pedestrian safety,
Accident Analysis and Prevention, 76, 64-73.
Lee J., Wong D.W.S., (2001), Statistical Analysis with
ArcView GIS, John Wiley, New York.
Marhavilas P., Koulouriotis D., Mitrakas C., (2014), Fault
and event-tree techniques in occupational health-safety
systems - part ii: statistical analysis, Environmental
Engineering and Management Journal, 13, 2371-2386.
National Institute of Statistic, (2015), Resident population
on 1 January 2015 separated by age and sex area of
residence, On line at: http://www.insse.ro.
Pirotti F., Guarnieri A., Vettori A., (2011), Collaborative
Web-GIS design: A case study for road risk analysis
and monitoring, Transactions in GIS, 15, 213-226.
Plug C., Xia J.(C.), Caulfield C., (2011), Spatial and
temporal visualisation techniques for crash analysis,
Accident Analysis and Prevention, 43, 1937– 1946.
Prasannakumar V., Vijith H., Charutha R., Geetha N.,
(2011), Spatio-temporal clustering of road accidents:
GIS based analysis and assessment, Procedia Social
and Behavioral Sciences, 21, 317–325.
Pulugurtha S.S., Krishnakumar V.K., Nambisan S.S.,
(2007), New methods to identify and rank high
pedestrian crash zones: An illustration, Accident
Analysis and Prevention, 39, 800-811.
Qin X., Parker S. Liu Y., Graettinger A.J., Forde S.,
(2013), Intelligent geocoding system to locate traffic
crashes, Accident Analysis and Prevention, 50, 1034-
1041.
Rui Y., Shen D., Khalid S., Yang Z., Wang J., (2015), GIS-
based emergency response system for sudden water
pollution accidents, Physics and Chemistry of the
Earth, Parts A/B/Cy.
Steenberghen T., Aerts K., Thomas I., (2010), Spatial
clustering of events on a network, Journal of
Transport Geography, 18, 411–418.
Truong L.T., Somenahalli S.V.C., (2011), Using GIS to
identify pedestrian-vehicle crash hotspots and unsafe
bus stops, Journal of Public Transportation, 14, 99-
114.
Wan T., Zhang G.M., Gao F., Lu H.F., (2014),
Solubilization of heavy metals in sludge during
sonication: impact of sonication time and power
density, Environmental Engineering and Management
Journal, 13, 2625-2632.
2701
 Ciobanu and Benedek/Environmental Engineering and Management Journal 14 (2015), 11, 2689-2702

Xie Z., Yan J., (2008), Kernel Density Estimation of traffic
accidents in a network space, Computers, Environment
and Urban Systems, 32, 396–406.
Xie Z., Yan J., (2013), Detecting traffic accident clusters
with network kernel density estimation and local
spatial statistics: an integrated approach, Journal of
Transport Geography, 31, 64–71.
Yao S., Loo P.Y., Lam W.Y., (2015), Measures of activity-
based pedestrian exposure to the risk of vehicle-
pedestrian collisions: Space-time path vs. potential
path tree methods, Accident Analysis and Prevention,
75, 320-332.

2702

View publication stats