Risk factors of fatal occupational accidents in Iran

Design and setting of the study

This analytical cross-sectional study utilized data from the 2013 registry system of the Ministry of Health and Medical Education of Iran. This patented system of worker’s accident information was launched in 2013 for the first time in Iran with the help of medical universities and related health care units. The target population was industrialized and non-industrialized adults, aged 18 years or older. It worth mentioning that as their main outcome of interest was death but there was not a nationwide study before it. So, in order to increase the precision of estimates the proportion of death was preliminary assumed as 50% (p = 0.50) and with confidence level of 95% (α =0.05) for an error rate of 1.5% (d = 0.015) using [TeX:] \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ n=\frac{Z_{1-\alpha /2}^2\times p\times \left(1-p\right)}{d^2} $$\end{document}n=Z1−α/22×p×1−pd2 the minimum sample size for this study was computed at least as 4269. Fortunately, they could gather data on 6052 individual accidents all over the country.

Measures

The dependent variable in this study was the consequence of the accidents in injured people that were categorized into two groups including the survived and dead. The independent variables were including sex, age (a quantitative variable), education, tenure in current job (a quantitative variable), occupation, usual work schedule, injury cause, damaged part of the body, date of the accident and referred location of the injured people. While sex has two groups, male and female, for analyses, we grouped age into two groups: < 45 and 45 or older years and education into “Not graduate from high school” and “Graduated from high school”. Job tenure was assessed in terms of the number of years that injured people had held their current job. To investigation the relationship of this measure with depend variable it categories into two groups: 1 year or less and more than 1 year. About injured workers occupations (130 different occupations) variable nine different categories based on the 2010 U.S. Census Occupation Code were made [13]. Categories including day shift, evening shift and night shift were used for shift work variable. Injuries causes measure in the registry system of accidents data categorized as: falling, collision & throw & hit, trapping, excessive force, high or low temperature, electrical shock, acute cause and other causes but for analyzing goals this measure recoded into three groups including: falling & collision & throw & hit, electrical shock and trapping & excessive force & other reasons finally.

For damaged parts of the body variable seven groups including: upper extremity, lower extremity, trunk & wall, head & neck, abdomen & pelvis, thorax and other parts were used in the statistical analysis. Season of accident has four groups including spring, summer, fall and winter while for accident day variable three groups including 1–10, 11–20 and 21–31 were made. For referred location of the injured people in the recording system of accidents data six category including: work place healthcare center, urban or rural healthcare center, hospital, clinic, forensics and others were existed but for statistical analyses this measure categorized into the healthcare center, hospital and other groups.

Data analysis

Data were analyzed by the Stata 14.0 (Stata Corp. USA) and R (version 3.3.1) software. Two sets of analyses were performed. First, univariate analyses were performed to assess the relationship of each potential risk factors with the consequence of occupational injury. For this purpose, Chi-Squared test, and Fisher exact test were used. Subsequently, an exact multiple logistic regression analysis was performed using the variables that were significant at p < 0.10 in the univariate analyses in order not to lose important variables due to confounding effect. As in this analysis, cells with small count was under investigation, exact procedure with backward elimination method was employed. In exact multiple logistic regression analysis first group of each variable was chosen as a reference group.

To calculate the YLL for each dead work force, age of death was subtracted from his/her life expectancy that it was derived from the life table of Iranian population [14]. One-way ANOVA test was used to compare mean of YLL related to each COIs (Causes of Injuries). Statistically significant differences were accepted at the p < 0.05.

Univariate analyses

The effect of different independent factors on death of subjects understudy are also shown in Table 1. As Table 1 shows, the annual incidence of death was higher in injured subjects whose job tenure were higher than 1 year (0.9% vs 0.3%; p = 0.01). The annual incidence of death due to electrical shock was the highest (7.40%) in comparison with falling, collision, throw and hit (0.5%), Others (Trapping & Excessive force & Other reasons) (0.6%) (p < 0.001). Also, incidence of death in subjects with other reasons of trauma (0.87%) was higher in compare to the other parts of body trauma (p < 0.001).

Table 1

Frequency, percentage and univariate analyzing results for all independent variables by consequence, 2013

Exact multiple logistic regression analysis

It was observed that amongst injury causes, electric shock enforced the highest chance of death (OR = 7.04; 95% CI: 1.01–43.74; p = 0.02). However, chance of death for injured subjects with job tenure more than one year was nearly 80% lower (OR = 0.21; 95% CI: 0.05–0.70; p = 0.005) (Table 2).

Table 2

Exact multiple logistic regression analysis result

Years of life lost

In 2013 the life expectancy at birth for Iranian males was about 73 years, according to that cumulative YLL for 33 dead people were calculated 1289.12 years. Such that cumulative YLL in effect of falling was 389.11 years (n: 10; 95% CI: 33.24–44.57), collision & throw & hit, 279.60 years (n: 7; 95% CI: 27.07–52.79), trapping, 90.26 years (n: 2: 95% CI: 27.36–62.89), high or low temperature, 60.48 years (n: 2; 95% CI: -115.71, 176.20), electrical shock, 199.27 years (n: 5; 95% CI: 35.75–43.95), acute cause, 58.60 years (n: 2; 95% CI: -116.27, 174.83), others, 211.80 years (n: 5; 95% CI: 37.05–47.61).

The mean (±SD) of YLL in effect of different COIs are shown in fig. 1. The difference between the means was not statistically significant (df: 2; F = 0.06; p = 0.08).

Fig. 1

The mean (±SD) of YLL in effect of different causes of injuries

Discussion

This study was the first study that conducted with using of a national wide data related to the occupational accidents in Iran. The main goal of this study was examination of relationships between the fatal occupational accidents and various socio-demographic factors, factors related to the work characteristics, and other available factors.

The multivariate results (Exact multiple logistic regression analysis) showed that current job tenure with fatal occupational accidents is associated. As expected, workers with less or equal to 1-year job tenure had more death (N = 15) in comparison to the workers with more than 1-year job tenure. It seems to work experience has protective effect on the happening of fatal accidents. Related to this, a study conducted by Chung J. and et al., positive correlation between age and work experience was observed (p < 0.01, correlation coefficient: 0.31) [15]. Also the study of J-M. Cellier and et al. indicated that low experience in beginning workers at comparison with intermediate workers, cause more frequent and seriousness of accidents [16]. The studies that were carried out in Korean, Chinese, Japanese, western cuisine restaurants showed that in all countries having longer period of work experience cause persons less to be injured by accidents [17]. Also, several studies have shown due to the effects of subcontracted work, labor turnover and short-term employment, new and inexperienced workers have the greatest risk of fatal occupational injury [18–20]. The question is, how work experience decreases the happening of accidents or seriousness of accidents, it may be related with the worker’s knowledge about job risks. For people who work in a company the knowledge about the risks which are specific to that place, increase with the passage of time so maybe this knowledge plays a protective role. On the other hand, Salthouse’s study revealed that the job experience can be effected basic cognitive processes or job performance positively [21].

As can see in Table 1 “falling & collision & throw & hit” is responsible for 16 deaths in total of 33 recorded deaths. But Exact multiple logistic regression analysis showed that just electrical shock as a cause of injury has statistical significant relationship with response variable. As shown in Table 1, electrical shock has more percentage of deaths (7.4%) in comparison to falling & collision & throw & tit group (0.5%) so this finding was not far from the mind. According to the Y.-H. Lin and et al. study, in dead male workers, falls (38.2%), electric shock (14.5%), collapse (11.3%) were the most leading causes of death [22]. In our study and in the dead workers like Y.-H. Lin and et al. study, falling & collision & throw & hit group was the first reasons of death so that it was responsible for 48.48% of deaths. Also our results were generally consistent with findings of Im and et al. study results. They found that in the construction industry in Korea, falling is the most frequent (52.7%) cause of fatal injuries. Also they found that deaths due to structural collapse and electric shock in construction industry in compare to the other industries is significantly higher [23]. In line with the findings of our study Cheng et al. reported that at small construction enterprises, the reason for 58 % of occupational accidents is falls and tumbles while electrical shock was responsible for 8% of occupational accidents [24]. It seems that lack of appropriate safeguards in result of inadequate safety management or not using of personal protection equipments are the probable reasons for the increase in the rates of falling, electrical shocks and so on.

In the calculation of the YLL because of any data was not registered for the dead women so we just did this calculation for men. Although because of cultural status and exist of some cultural limitations for woman in Iran this phenomenon was expectable. As we expected women in comparison with men employ mostly in less hazardous occupations in Iran so fatal occupational accidents less happen to them. Cohen et al. reported that average years of potential life lost (YPLL) per occupational fatality for workers employed in the remediation sites is 38.09 years [25]. In this study we found that the average YLL for each dead worker is about 39.06. A study conducted by mohammadfam et al. in Tehran (capital of Iran) showed that cumulative years of life lost for all dead insured workers because of occupational fatalities is 7552 years such that the average YLL for each dead worker was about 32.69 years [26]. It seems that difference in years studied and number of years studied are two main reasons to make differences in findings.

The primary interest of researchers was to investigate the effects of work organization factors like: work-family interference, management employee relations, organizational effectiveness, safety climate and job content on the fatal consequence of occupational accidents but unfortunately related factors were not recorded in the registry system of the Ministry of Health and Medical Education of Iran. Also, access to the one year of data related to the occupational accidents was other limitation of our study. So we suggest that in addition to the factors that investigated in this study other mentioned factors to be studied in the future research. Also investigation of fatal occupational accidents and related factors in the passing of time for studying of trends will be interested.