Analysis of occupational accidents in the construction industry with regards to selected time parameters

Abstract The construction industry is characterized by a large variety and variability of implementation conditions. Construction works are carried out throughout the whole calendar year under varying atmospheric conditions, both during the day and also during the evening and night. Construction workers also often work overtime that exceeds the eight-hour working day. Work in such conditions creates a particularly high threat to the health and life of working people. In the issues of work safety the most important is worker, who in the accident process has a triple role: the decision maker, the perpetrator of an accident and the victim. The purpose of this article is to examine whether there is a dependence between the number of occupational accidents in the construction industry and time parameters such as: the time of an accident, the day of the week and the month in which an accident occurred. The article also attempts to determine the dependence between the age of the injured person, work experience and the number of hours worked by the injured person from the time of starting work to the moment of the accident.


Introduction
In recent years, a large number of scienti c works devoted to the problems of engineering construction processes has involved studies and scienti c research that deal with the subject of occupational work safety in the construction industry. This topic covers several research areas, including: • analysis of the mechanisms of the occurrence of occupational accidents [1,2], • analysis of the causes of accidents at work [3][4][5], *Corresponding Author: Mariusz Szóstak: Wroclaw University of Science and Technology, Faculty of Civil Engineering, Department of Construction Technology and Management, Poland, E-mail: mariusz.szostak@pwr.edu.pl • identi cation of direct and indirect factors which affect the accident rate [6][7][8] • analysis of the accident rate with regards to factors that generate costs [9][10][11].
In the issues of work safety the most important is worker, who in the accident process has a triple role: the decision maker, the perpetrator of an accident and the injured person, in the worst case, as the victim. Worker as a participant in the investment process is exposed to many dangerous factors that cause accidents. Sometimes worker is the source of these factors [12]. Human causes account for over 65% of all causes accident at work. Failure to use personal protective equipment by an employee, disregard of a danger and inappropriate psychophysical state of an employee due to the consumption of alcohol, drugs or psychotropic substances are the main causes of accidents [13]. The construction industry is characterized by a large variety and variability of working conditions and dangerous employee behaviour. Construction works are carried out throughout the whole calendar year under varying atmospheric conditions, both during the day and also during the evening and night. Construction workers also often work overtime that exceeds the eight-hour working day. Work in such conditions creates a particularly high threat to the health and life of working people.
The purpose of this article is to examine whether there is a dependence between the number of fatal and serious occupational accidents in the construction industry and time parameters such as: the time of an accident, the day of the week and the month in which an accident occurred. The article also attempts to determine the dependence between the age of the injured person, work experience and the number of hours worked by the injured person from the time of starting work to the moment of the accident.
The article was organized in the following way: Section 2 presents a review of the literature related to the topic of the article. Section 3 discusses the sources of information about the injured person in an accident at work, description of the analyzed parameters and the used re-search method. Section 4 contains the results of statistical analysis 630 people injured in occupational accidents in the construction industry and the dependencies between individual parameters using the Spearman rank correlation coe cient. Finally, section 5 is a summary that identi es the most important conclusions.

Literature review
Studies carried out to date have shown that there are dependencies between changing working conditions and the factors that characterize victims. Investigations of fatal accidents that took place in the years 2001-2005 at construction sites in Taiwan showed a relationship between the size of an enterprise in which the accident took place, work experience and the age of the injured person [14]. Similar studies were carried out in Iran. On the basis of 21 864 people injured in accidents in the construction industry, the existence of a dependence between the examined features was con rmed. These features included, among others, the age of the injured person; the date, time and place of the accident; as well as the causes and consequences of the accident [15].
The analyses carried out on more than 1.1 million occupational accidents in the Spanish construction industry, which took place during 2003-2008, showed an occurrence of dependencies between the e ects of occupational accidents in the construction industry and the following parameters: the age of the injured person, the type of conducted work, the size of an enterprise, work experience, the place of the accident, the day of the week, the number of days of absence from work, injuries and the climatic zone [16].
The analysis of 485 people injured in occupational accidents, which took place during 2008-2014 in the Polish construction industry, allowed the impact of the following characteristics on an occupational accident to be examined: the employment status of an injured person, occupation, age, work experience, preparation of an employee to perform duties at a workplace. As a result of the research the authors determined the pro le of a person who is the most often a ected by an accident in the construction industry [12].
Data analysis and literature studies also showed the existence of a dependence between physiological parameters of the human body, which change in subsequent periods of life, and accidents [17].

Research methodology
The main body of the supervision and control the principles of occupational health and safety of compliance in Poland is the State Labour Inspectorate (SLI). The State Labour Inspectorate are obliged to investigate the circumstances and causes of fatal and severe accidents. In addition, the District Labour Inspectorate (DLI) keep a record of the circumstances and causes of accidents at work causing minor personal injury and potentially accidental events, although this is not mandatory by law. After the postaccident investigation, a labour inspector draws up the post-accident protocol, in which he includes the conclusions from the investigation. The post-accident protocol contains a description of the actions carried out in connection with the claimed accident and presents the circumstances and the course of an accident, information on the time and place of the event. Moreover post-accident protocols contain data that characterizes an injured person, namely: sex, year of birth, citizenship, employment status, performed occupation, work experience at an occupied workplace in the plant in which the accident took place, and also preparation for work [18]. In this article the post-accident protocols were the source of information regarding injured people.
Statistical data on the accidents at work which have taken place in the construction industry are collected also by the Central Statistical O ce of Poland (CSO). However, the Central Statistical O ce of Poland mentioned above has a di erent set of data on accidents at work than the State Labour Inspectorate (SLI). Each year CSO publishes a report entitled: "Accidents at Work. Statistical Data and Compilations". Data on accidents at work are collected from employers are compiled there in tables. From this source one can learn, e.g., how many persons were injured in accidents, how many persons sustained injuries on building sites and in other places. The statistics contained in the sets were compiled in accordance with the Statistical Accident Card requirements.
On the basis of collected documents regarding occupational accidents in the construction industry that occurred in the years 2008-2017 in selected 5 Polish voivodeships, which were provided by District Labour Inspectorates in Poland, a database of occupational accidents was created. The article investigates 630 occupational accidents that took place in the Dolnoslaskie, Kujawsko-Pomorskie, Lubelskie, Lubuskie and Slaskie voivodships. In this paper only fatal and serious accidents at work have been studied. In the conducted research, the relationships between the following 6 parameters were de ned: 1. day of the week, 2. month 3. time of the accident, 4. age of the injured person, 5. work experience of the injured person, 6. number of hours worked by the injured person from the time of starting work to the moment of the accident.
The correlation analysis was used to determine the correlation between the studied parameters, and the Spearman's rank correlation coe cient was used to describe the relationship between the assessed parameters. To interpret the correlation and to determine the degree of dependence of two features (rxy), a 7-degree scale was assumed, namely: when rxy = , the variables are not correlated; when < rxy < . , there is a very weak correlation; when . ≤ rxy < . , there is a weak correlation; when . ≤ rxy < . , there is an average correlation; when . ≤ rxy < . , there is a high correlation; when . ≤ rxy < . , there is a very high correlation; and when . ≤ rxy < . , there is almost full correlation [19]. In the conducted research, when interpreting the correlation coe cient, the most important thing was to determine the signi cance of the correlation, and then to determine its degree. The calculations were performed using the STA-TISTICA 13 integrated statistical and analytical software package.

Analysis of statistical data
The analysis covered 630 people injured in occupational accidents in the construction industry, including 222 fatal accidents (which accounted for 35.24% of all accidents) and 408 severe accidents (which accounted for 64.76% of all accidents). Table 1 presents the results of the conducted calculations.

. Analysis of the day of the week
The structure of people a ected in an accident with regards to the day of the week in which the accident occurred is shown in Figure 1 and is as follows: • fatal accidents most frequently occurred on Wednesdays and Thursdays. The number of injured people was equal to 45 and 44, respectively, and their share was respectively 20.27% and 19.82% of all fatal accidents, • as many as 85 people su ered from a serious injury on Friday, which amounted to 20.83% of all serious accidents, • when analyzing all the victims, it should be noted that accidents occurred most frequently on Wednesdays. The number of injured people was equal to 123, and their share in the total number of victims amounted to 20.27%. On the basis of the analyzed data, the common opinion that occupational accidents most often occur at the beginning and end of the average standard 5-day working week cannot be con rmed. As with the analyzes carried out by other researchers [15], [20], the conducted research did not show any relation between the day of the week and the possibility of an occupational accident occurring. Analysis of accidents on particular days of the week showed that their number changes in a small range and oscillates between similar values. The analysis of the collected data does not indicate that the day of the week has an impact on the occurrence of an accident, although there is a slight upward trend in the number of accidents in the second half of the week.
. Analysis of the month in which an accident occurred Figure 2 shows the number of occupational accidents with regards to the month in which the accident occurred. The conducted research con rmed that the number of occupational accidents depends on the so-called construction season, the peak of which is in the third quarter of the year, i.e. from July to October [17], [21]. The analyzed data indicates a greater susceptibility of occupational accidents by employees during summer (33.02%, i.e. 165 occupational

. Analysis of the age of the injured person
The age structure of the victims is shown in Figure 4 and is as follows: • the largest group of people injured in occupational accidents consisted of people aged 20-29. The number of people injured in this age group amounted to 192 people, and their share in the total number of victims amounted to 30.48%. The obtained statistical data showed that in every third accident, a person who is under 30 years old is harmed, • in second place is the group of people aged 50-59. The share of this group in the total number of people affected during accidents amounted to 24.63%, with the number of victims amounting to 155 persons, • the most frequent value in the set of analyzed data (modal) was 29 years, and the average age of the injured person was equal to 39 ± 13 years, • the number of accidents among young employees (aged 20-29) is much higher than in other age groups, • the severity of the accident increases with the age of the employee. In the group of older workers aged over 50, fatal accidents are more frequent than severe accidents, whereas younger people are most often a ected by severe accidents. . Analysis of work experience Figure 5 shows the number of people injured in occupational accidents with regards to their work experience in an enterprise. The analysis showed that: • 57.46% of all injured people (i.e. 362 people) were involved in occupational accidents in the construction industry, which occurred in the rst year of their work at the occupied workplace, • up to 49 employees were injured on the rst day of their work, which accounted for 7.78% of all injured people. Among these employees, 14 people were killed and 35 people were seriously injured. • 152 people, i.e. 24.13% of all injured people were injured during the rst week of their work. From this number, 46 people were injured with a fatal outcome, and 106 people were severely injured, • 225 people were injured during the rst month of their work, which accounted for 35.71% of the total number of victims, of which 75 people died and 150 people were severely injured, • the longer the work experience at a workplace, the lower the number of accidents, • the most frequent value in the set of analyzed data (modal) was 1 day, and the average work experience of the injured person was equal to 182 days, i.e. half a year.
The analysis of statistical data showed that people with fewer years of work experience at a workplace -a few days or weeks -most often fall victim to accidents. This is caused by, among others: • a lack of or incorrect training in the area of occupational health and safety, • allowing an employee to work with medical contraindications or without medical examinations, • insu cient professional preparation of the employee, which is related to the performance of the activity without the required permission, • ignorance of the provisions and principles of occupational health and safety, • insu cient focus on the performed activities, • a lack of experience of the victim. Figure 6 shows the number of occupational accidents with regards to the number of hours worked since starting work to the moment of the accident. The analysis of the time of the accident allowed the following conclusions to be formulated: • the accident occurred most often in the rst hour of work. The number of injured people amounted to 86, and their share in the total number of victims amounted to 13.65%, • the same number of injured people were noticed in the group of people who worked overtime that exceeded eight hours of work. As many as 34 people had a fatal accident and 52 people had a severe accident, • the subsequent places regarding the number of injured people applied to time intervals between the second and third working hour (79 injured people, 12.54%), the fth and sixth working hour (76 injured people, 12.06%), and also between the seventh and eighth working hour (73 injured people, 11.59%).

. Analysis of the number of hours worked by an injured person
. Correlation analysis Table 2 shows the correlation matrix between the analyzed parameters and the calculated Spearman rank correlation coe cients.
In the carried out correlation analysis, no correlation was found for 13 pairs (relations) out of 15 possible, including (1) and (2), (1) and (3), etc. For 13 relations, the Spearman's rank correlation coe cient obtained values within the range < rxy < . , which means a faint or no correlation. Positive correlations were only obtained between the parameters for two relations, i.e. between (3) and (6) and between (4) and (5). The obtained positive correlation means that both parameters increase simultaneously, i.e. the high values of one parameter refer to the high values of the other.
The Spearman's rank correlation coe cient amounted to 0.9660 for parameters (3): time of an accident, and (6): the number of hours worked by the injured person from the time of starting the job until the accident. The obtained correlation coe cient indicates an almost full positive correlation between the analyzed parameters ( . < rxy < . ), which means that there is a strong relationship between the time of the accident and the number of worked hours. The analysis of dependencies showed that accidents most frequently happen in the following hours:

Conclusions
This article presents the results of statistical analysis carried out on 630 occupational accidents that took place in Poland in selected voivodeships in the years 2008-2017. In the conducted research, the following dependencies were determined between 6 time parameters: day of the week, month, time of the accident, age of the injured person, work experience of the injured person and the number of hours worked by the injured person from the time of starting work to the moment of the accident. The dependencies between individual parameters were determined using the Spearman rank correlation coe cient. The study showed a strong correlation between the time of the accident and the number of hours worked, as well as the average correlation between the age of the injured person and work experience. With an increase of peoples' age and their work experience, the number of occupational accidents decreases.
The analysis of people injured in occupational accidents in the Polish construction industry allowed the following conclusions to be formulated: • the tested parameters in uence the number of accidental events to a di erent degree, • the day of the week does not a ect the possibility of an occupational accident occurring, although there is a slight upward trend in the number of accidents in the second half of the week, • the analyzed data indicates a greater susceptibility to occupational accidents by employees during summer when compared to the winter period, • occupational accidents most often occur in three time intervals: between 07:00-07:59, 12.00-12.59 and 14.00-14.59, • analysis of the age structure of injured people showed that the most common age group of injured employees is those aged 20-29, • the probability of an accident occurring is the greatest in the rst year of work at an occupied workplace when compared to subsequent years of work, • an accident usually occurs in the rst hour of work, as well as during extra hours that exceed the eight-hour working shift.
The determination of the main time parameters and their impact on the occurrence of accidents in the construction industry is the rst step to minimize and reduce the number of occupational accidents. According to the author, the conclusions from this study can be used by construction companies in their work safety strategies, as well as in safety training programs. Sta training should be conducted while taking into account the speci c needs of each group of employees. Particular attention should be paid to a group of employees with zero or little work experience (up to 1 year). According to the author, some parameters should be further investigated in future research in order to determine other existing signi cant relationships between the analyzed parameters.