- Email: [email protected]
A statistical study on temporary work and occupational accidents: Specific risk factors and risk management strategies
Available online at www.sciencedirect.com
Safety Science 46 (2008) 535–544 www.elsevier.com/locate/ssci
A statistical study on temporary work and occupational accidents: Specific risk factors and risk management strategies Bruno Fabiano *, Fabio Curro`, Andrea P. Reverberi, Renato Pastorino Chemical and Process Engineering Department ‘‘G.B. Bonino”, University of Genoa, via Opera Pia 15, 16145 Genoa, Italy Received 23 March 2006; received in revised form 26 April 2007; accepted 4 May 2007
Abstract Temporary work, supplied by temporary-help agencies and sometimes referred to as ‘‘job in leasing”, was only recently introduced in Italy, and has since spread considerably thanks to its flexibility and cost effectiveness. In this study, trends in the rates of occupational injuries in different sectors of Italian industries in the period 2000–2004 are explored, contrasting direct employment and temporary work. Data on occupational injuries from the National Organization for Labour Injury Insurance as well as data directly obtained through a field survey in three large manufacturing firms were analysed to highlight the interaction between injury frequency index (FI) and the characteristics of the labour force. FI for temporary workers ranged between 89.23 and 94.10, i.e., between 136.4% and 175.2% more than the value found for direct employees in the most hazardous industrial sector. Also accident severity (assessed on the basis of time lost due to injuries) is twice the overall value of the severity index. The results from the field survey confirmed the trend: FI for direct employees ranges from 25.7 to 45.0 in respect of total hours of work in the range 1.99 106–2.40 106; whereas FI in temporary workers ranges from 86.2 to 163.5 in respect of total hours of work ranging from 1.25 105 to 1.28 105. Given this evidence, an empirical statistical analysis, based on responses from injured temporary workers was carried out. The questionnaire was intended to collect data on a variety of control variables relating to personal characteristics of respondents. The analysis of the questionnaire data using ANOVA and response surface methodology highlighted the interaction between short duration of work assignments, inadequate training period and FI increase. Reasons can be traced to lack of experience in the activity, insufficient specific knowledge (formal and informal knowledge) about a particular installation and to inadequate training period. Means of promoting safety include diffusion of information on regulatory provisions, management training in safety, enhancement of specific training and formation. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Frequency index; Injury; Questionnaire; Response surface methodology; Temporary work; Training
*
Corresponding author. Tel.: +39 0103532585. E-mail address: [email protected] (B. Fabiano).
0925-7535/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.ssci.2007.05.004
536
B. Fabiano et al. / Safety Science 46 (2008) 535–544
1. Introduction On a national level, Italy has enjoyed declining trends in the number and rates of occupational fatalities and injuries for more than three decades. While these figures are encouraging, far too many preventable injuries are still occurring. The factors influencing accident frequency can be divided into the following categories: technical factors: low automation, multi-product industries, discontinuous operating cycles, non standardized production affect safety negatively, since they require a greater interaction between man and devices. On the other hand, a reduction in individual exposure to severe hazards was reported where mechanized processes and equipment were introduced in the mining (Asogwa, 1988) and the logging industry (Laflamme and Cloutier, 1988); economical factors, e.g., general economic situation (Saari, 1982), unemployment rate, labour and socialinsurance legislation, (Blank et al., 1996); labour organization, e.g., management system and performance, work practice, oversight, communication structure, etc.; environmental conditions: about half the accidents in Italy are related to labour environment (Fabiano et al., 2001) and they could be prevented by rather simple lay-out and protection measures, which however prove extremely difficult or even unfeasible in small concerns, because of operating, economic and/or space constraints (Fabiano et al., 2004); human factors, both individual and inter-individual, e.g., workload, experience and training, competence, fatigue, etc. Current market conditions often make it necessary to resort to outsourcing to remain competitive, particularly by employing external and precarious human resources. In fact, in the last 20 years there has been a significant growth in labour governed by casual, part-time, subcontract or franchised arrangements, virtually in all OECD countries (less so in Italy, Canada and Luxenbourg) (Ferrie et al., 1999). Results of international research seem to point out that precarious labour is associated with increased fatalities, occupational injuries, illness, in various industry sectors across a number of countries (Kochan et al., 1994; Morris, 1999; Mayhew and Quinlan, 1999). Temporary work (i.e., a job arrangement whereby workers are hired out to firms by Temporary Work Agencies) is a sort of labour outsourcing and precarious work. This kind of work arrangement, sometimes referred to as ‘‘job in leasing”, has only recently become widespread in Italy. Due to the flexibility and cost effectiveness of this contract, in the last few years Italy has experienced constant growing rates of temporary work. The whole number of contracts provided by temporary work agencies soared from 194,835 in 1999 to 744,438 in 2003, corresponding to an average increase of about 56.4% per year. Virtually, every industrial sector shared this trend, with an average daily percentage of temporary workers corresponding to 1% of ‘‘in-house employees”, for all industrial sectors in the year 2003. Process, metal–mechanic and transport sectors displayed the highest average daily percentage of temporary workers, amounting to about 3% of direct employees. In the same year, the average EC daily incidence of temporary work was 1.5%, with higher values in the UK (3.2%) and in the Netherlands (4.5%), while the daily incidence in the United States was 2.5%. Despite the international diffusion of temporary work, in the open literature there is little information on possible safety implications and contributing factors of occupational injuries. Moreover, even if most safety-related regulations and safety management systems already incorporate provisions for the management of contractors, a striking number of injuries occurs among ‘‘non in-house employees”. In Spain, 60% of occupational injuries occurred to temporary workers, with a rate nearly three times that of those in permanent jobs (Artiles and Alos-Moner, 1999). A survey conducted among three manufacturing industries of the plastic sector in the USA, showed that the trend of accident rate over the years 1993–1998 does not appear to be a steady one, either for direct employees, or for temporary workers. However, the evidence obtained from such research indicates that the injury rate for temporary workers is constantly higher (from 2.6 to 3.8 times) than the one recorded for permanent employees (Morris, 1999). This paper offers a perspective on the trends in the rates of occupational injuries in the main sectors of Italian industries, during the period 2000–2004, contrasting direct employment and temporary work and
B. Fabiano et al. / Safety Science 46 (2008) 535–544
537
focusing attention on the human factor. Generally speaking, the different reasons for people failure can be grouped as follows: lack of training or instruction; lack of motivation; lack of physical or mental ability; slips and lapses of attention (Kletz, 1993). On this basis, an investigation into the relationship between occupational accidents and temporary work is performed, adopting a questionnaire survey, for the definition of peculiar risk factors and for setting priorities to improve safety standards in this context. 2. Materials and methods 2.1. Research design Raw data were obtained from the National Organization for Labour Injury Insurance (I.N.A.I.L. Istituto Nazionale per l’Assicurazione contro gli Infortuni sul Lavoro) (INAIL, 1970–2004). One of the main measures of performance is the lost time accident (LTA), which however has only limited value and should be supplemented by other measurements, such as the total accident rate, the cost of the damage caused by accidents and other dangerous incidents and, if possible, a numerical measure of the results of plant audits (Kletz, 1993). The frequency index (FI), allows to find out the incidence of accidents of workers exposed to risk and is defined as follows: FI ¼
Number of total accidents 6 10 : Number of worked hours
ð1Þ
As ‘‘worked hours” can be calculated regularly, this index may offer a proper instrument for evaluation of accidents. To estimate trends, we modelled the rate of injury frequency index as a function of the year, by adopting a multiple regression model with a log transformed rate (Bailer et al., 1997; Fabiano et al., 2001), as follows ln FI ¼ a þ bðyeari year0 Þ;
ð2Þ
where year0 is the first year of the study, to which corresponds the baseline of injury rate. Data on temporary workers were collected directly from the 16 leading temporary-job agencies in Italy, covering nearly 90% of the whole market. Data from INAIL and from direct field survey (temporary job agencies) were combined with data from the State Statistics Institute, in order to highlight the interaction between the trend of the injury frequency index and the characteristics of the labour force, throughout the years from the introduction of the new work arrangements. The critical issue is whether there is a statistical relation between occupational injuries and temporary work, and whether or not such relation differs for different occupation groups. In order to confirm the impact of new working regulations on the occurrence of occupational injuries, a detailed case study was made on three large firms of the manufacturing sector, characterized by high proportion of temporary work. Finally, an empirical statistical analysis was carried out, based on responses by injured temporary workers to an ‘‘ad-hoc” questionnaire specially designed for the purpose. 2.2. Questionnaire development The questionnaire was intended to collect data on a variety of control variables relating to personal characteristics of respondents. The main section of the questionnaire was designed to characterize technical, organizational and individual factors of suffered injuries, as well as to find out how firms try to reduce the occupational risk to which temporary employees may be exposed. The content of the questionnaire is described below. Demographic and work details: respondents were asked to give information on their sex, age, job title, employer, extent of their job experience and specific work task, and task rotation if any. Safety attitudes: respondents were asked to rate their job and safety training, extent of the temporary contract, whether or not they were accompanied by a job tutor, how much support they received from fellow workers on the job.
538
B. Fabiano et al. / Safety Science 46 (2008) 535–544
Accident and near misses: respondents were asked to describe the accident and its severity based on lost time accident (LTA), time elapsed from starting of the contract until the accident, perceived accident cause (distraction, work pressure, difficulty in job performing, lack of specific/safety training, insufficient preventive measures in plant/machine/equipment, inadequate personal protective equipment, fatalistic event). Data from the structured questionnaire were coded and entered into a database for subsequent univariate analysis of variance (ANOVA). The independent variables whose effects on number of injuries and severity were evaluated included: worker’s age, job position, training period, on-site experience, temporary contract duration, perceived accident cause. Significant results were analysed adopting response surface methodology (RSM), this latter being a collection of mathematical and statistical techniques widely adopted to optimise different processes (e.g., Dasu and Panda, 2000), used in accident analysis for the first time, at least to our knowledge. 2.3. Sample The study population was identified from 16 staffing companies in Milan, Italy, covering nearly 90% of the whole market of temporary work. The questionnaire was distributed to 700 injured workers resulting from the investigation. All ‘‘in itinere” accidents (accidents occurred on the way to or from the workplace) were excluded from the sample. Respondents were informed that the interview data were to be treated as anonymous and that only aggregated results would be produced. The response rate for this distribution was 43.3% (n = 303). The average age of all respondents was 29.2 years (SD = 8.23) in the range from 18 to 45 years. 3. Results and discussion 3.1. Injury trends As shown in Fig. 1, the overall value of FI on a national level decreased, in the last three decades, from 94.92 in 1970 to 22.54 in 2000, with a mean reduction per year corresponding to 5.48%, obtained from the fit of the regression model (Eq. (2)). This encouraging trend was common to every industrial sector, with some differences depending on sector. It must be pointed out that raw data from INAIL obviously do not include injuries in case of irregular or unregistered employment (e.g. moonlighting phenomena), as well as accidents
Fig. 1. Overall injury frequency index (FI) over the period 1970–2000 and regression model with prediction intervals (95%).
B. Fabiano et al. / Safety Science 46 (2008) 535–544
539
involving less than three days absence from work, since, in Italy, they need not be reported. However, these limitations concern not only the Italian reality, but generally the worldwide safety community; in fact, the problem of improving the organizational memory and the need of a new approach to the collection and analysis of accident data are becoming more and more important (Kletz, 1993). Table 1 shows the values of the national frequency index and mean values over five years, for different industrial sectors (in-house workers) and for temporary workers, over the period 2000–2004, i.e., since the introduction in Italy of temporary-work arrangements. Again, a general decline may be observed in the frequency of injuries, for direct employees, across all industries, from 2000 until 2004, at an average estimated rate of decline of 4.29% per year, from a baseline of 22.54 in 2000. Practically speaking, all sectors are reaching minimum values of the frequency index, which can be regarded as a statistically reinforced reference for the minimum risk levels per individual employed in the different industrial activities. Obviously, this is only a statistical consideration, not to be regarded as an attempt to renounce the effort of reducing even the lowest injury frequency recorded. The trend for temporary work is not decreasing, but shows some variations from year to year, mainly because of the different percentage distribution of temporary workers in the various industrial sectors. However, it should be pointed out that the FI value for temporary workers is 2.65 times the mean value for the highest risk sector at national level (Building sector). It is worth noting that the average FI value of temporary workers over the time span considered is comparable to the value of the national frequency index for direct employees recorded in 1971, i.e., 91.83 (see Fig. 1). In order to offer a better insight on this observation, Table 2 shows FI values calculated for the Lombardia region, where nearly 63% of temporary workers were employed in the years considered. FI values in Lombardia for direct employees are constantly lower than the corresponding values at national level, with mean differences ranging from 5.35% in the food sector to 20.68% in the process sector. Considering this difference in the reference region, the injury trend declines constantly by 3.76% per year and FI of temporary workers is by far higher than FI of the most hazardous sector, to a percentage of 410.9%. These findings are in good agreement with the few studies available in the international literature (Artiles and Alos-Moner, 1999). A case-study conducted over five years in a process industry in India concluded that accident incidence, accident frequency and severity rates were significantly higher for temporary workers than for permanent workers, although the two groups were similar in relation to age, level of education, habits and nature of work. Relative risk varied from 2.3 to 18.0 in case of time loss accidents and from 1.1 to 2.6 in case of no time-loss accidents (Saha et al., 2004). Higher accident risk of temporary workers might have been due to less effective experience and to lack of job security (Saha et al., 2005). The results presented here underline that temporary work may be considered a risk factor. Reasons are to be traced to its intrinsic characteristics: age of workers, lack of experience in the activity, insufficient specific knowledge, low level of experience in the same firm, lack of a sufficient training period and heavy workload. In order to confirm the impact of innovative employment arrangements on the occurrence of occupational injuries, a detailed case study was carried out on three large firms of the manufacturing sector, characterized by a high proportion of temporary work. The three firms belong to the metal and mechanic sector and their employees have similar characteristics: number of hours worked, percentage of temporary workers, position
Table 1 Injury frequency index (FI) calculated, on a national level, for direct workers in the different industrial sectors and for temporary workers Mean
Year 2004
2003
2002
2001
2000
Metal and mechanic Process Food Services Building All sectors
34.07 ± 3.07 19.39 ± 1.69 22.61 ± 1.49 20.18 ± 1.51 34.77 ± 2.78 20.76 ± 1.42
30.09 17.25 20.62 18.76 32.02 18.98
32.08 18.11 21.90 18.93 32.76 19.70
34.15 19.60 22.60 19.87 34.19 20.87
36.52 21.02 23.38 20.91 35.93 21.70
37.51 20.98 24.56 22.41 38.95 22.54
Temporary work
91.63 ± 2.22
89.42
89.23
94.10
93.30
92.10
540
B. Fabiano et al. / Safety Science 46 (2008) 535–544
Table 2 Injury frequency index (FI) calculated, in the reference region (Lombardia), for direct workers in the different industrial sectors Industrial sector
Metal and mechanic Process Food Services Building All sectors
Mean
Year
28.35 ± 2.70 15.16 ± 1.19 21.31 ± 0.97 18.08 ± 1.28 29.16 + 2.05 17.84 ± 1.27
2005
2004
2003
2002
2001
2000
25.39 14.03 20.88 17.59 26.95 16.60
25.88 13.89 20.67 17.12 28.63 16.79
27.27 14.53 20.39 16.88 27.10 16.99
28.62 15.49 21.26 17.91 29.76 18.07
30.88 16.40 21.56 18.64 30.12 18.84
32.08 16.61 23.10 20.36 32.39 19.76
and tasks within the firm. The only remarkable difference regards the training period and the tutoring applied to temporary workers: firms A and C developed and implemented an integrated management system encompassing quality (ISO 9000), environmental management (ISO 14000) and safety and health. This integrated management system includes a specific job-training methodology that applies to sub-contractors and temporary workers hired through staffing agencies. Firm B established separate management systems for environment, health and safety and quality, with standard training of personnel addressed to both ‘‘in house” workers and temporary workers. Table 3 shows the results of the statistical analysis by direct survey in these three firms: the injury frequency index for temporary workers in firms A and C appears to be 1.9 times FI of direct employees. This figure is consistent with results from Morris, 1999), which indicate that the injury rate for temporary workers is constantly higher (from 2.6 to 3.8 times) than the one recorded for permanent employees in three manufacturing industries. Firm B scored a strikingly high difference, with FI referred to temporary workers 6.35 times FI of direct employees. As shown in Table 3, in the three firms examined, also severity (LTA) resulted higher than the value calculated for direct employees by a percentage from 87% to 97%. Even if limited, the experience of three Italian firms seems to show how a robust health, safety and quality management system, including different training for ‘‘in-house” and temporary workers can affect positively safety performance. Furthermore, management system integration avoids duplication or conflicting of procedures established for quality management with similar procedures for health and safety (Holdsworth, 2003). Clearly, in the area of safety training and education, the effectiveness of intervention should be carefully checked at an in-depth level of analysis (Hale, 1984), especially when dealing with new and precarious job arrangements.
3.2. Questionnaire statistics In order to test the main effects of workers age, job position, employment sector, training period, on-site experience, duration of temporary assignment and perceived accident cause on injury probability and injury severity, a univariate analysis of variance (ANOVA) was carried out. Industrial sector and job position of the sample are shown, respectively, in Tables 4 and 5. We may notice that the sample of this research is composed for 89.21% by low-level line workers, mainly in the process and metal–mechanic sectors (83.2%). No significant differences emerged on the workers age, job position, assignment duration and perceived cause. Three
Table 3 Injury frequency index and severity for in-house workers and temporary workers in three metal–mechanics settings Reference firm A
Worked hours Number of injuries FI Severity (LTA)
Reference firm B
Reference firm C
In house workers
Temporary workers
In house workers
Temporary workers
In house workers
Temporary workers
2,401,210 106 44.14 60
125,898 11 87.4 112
2,135,544 55 25.75 70
128,424 21 163.52 138
1,998,272 90 45.04 76
127,590 11 86.21 142
B. Fabiano et al. / Safety Science 46 (2008) 535–544
541
Table 4 Employment industrial sector of respondent temporary workers Industrial sector
Number
Process Metal and mechanics Food Services Building
47 205 16 23 12
% 15.51 67.66 5.28 7.59 3.96
Total
303
100.00
Table 5 Job position of respondent temporary workers Number
%
Skilled worker Unskilled worker Storeman Other
45 203 37 18
14.85 67.00 12.21 5.94
Total
303
100.00
interactions were statistically significant (see Table 6), namely industrial sector (p < 0.01%), training period (p < 0.05%) and on site experience (p < 0.05%). Five categories of on-site experience were entered (less than 7 days; 7–30 days; 31–60 days; 61–90 days; more than 90 days) and four categories of training period (less then 3 h; 3–5 h; 6–7 h; more than 7 h). Industrial sector categories were as follows: process, metal and mechanics, food, building, services. Fig. 2 shows the three-dimensional graph representing the response surface for injury probability as a function of training and on-site experience. The significant interaction of the independent variables indicates that an increase of the training period (professional training and job tutoring) greatly reduces the probability of injury. It must be noted that safety programs include training as a part of the risk management process. However, the implementation of rules and training may often prove to be insufficient in reducing unsafe practices, since safety rules are often considered to apply only in certain situations while being impossible to follow in the many exceptional situations which are seen to be the reality of the shop floor situation (Hale, 1990). Equally, it seems that the worker’s permanency on the same job site entails an increase of experience and acquaintance with one’s duties, reducing the probability of an accident. In other words, even if employees are initially unaware of occupational risks, they can often acquire on-the-job experience. A key aspect is that, as reported by Asogwa (1988), an adaptation period is required for workers to perform adequately in new work assignments and in a changed environment, whereas in conditions of production intensification/pressure (as are those usually corresponding to the utilization of temporary workers) such training period can be curtailed or eliminated. Moreover, intensification of work and production in combination with an increase in overtime was referred to constitute a set of interacting factors increasing injury risk (Saari, 1982). Available evidence suggests that workers have some idea about risk inherent in their chosen occupation,
Table 6 Univariate tests of significance
Intercept Industrial sector Training On-site experience Error
Sum of squares
DF
Mean square
F value
P value
64.6354 59.0997 17.9502 27.7416 302.5752
1 4 3 4 137
64.63538 14.77492 5.98341 6.93540 2.20858
29.26560 6.68979 2.70917 3.14021
0.000000 0.000060 0.047606 0.016589
Sigma-restricted parameterization.
542
B. Fabiano et al. / Safety Science 46 (2008) 535–544
Fig. 2. Response surface for injury probability, as a function of training and on-site experience (legend: a = less than 7 days; b = 7–30 days; c = 31–60 days; d = 61–90 days; e = more than 90 days; r = less then 3 h; s = 3–5 h; t = 6–7 h; u = more than 7 h).
even if this information is rarely perfect or processed correctly (Rundmo, 1992). The ANOVA results for accident severity (LTA) show a statistical significance on training period (p < 0.05%) and on site experience (p < 0.05%). Fig. 3 shows the response surface plot of injury severity (LTA, days) as a function of training
Fig. 3. Response surface plot of injury severity (LTA, days), as a function of training and on-site experience (legend: a = less than 7 days; b = 7–30 days; c = 31–60 days; d = 61–90 days; e = more than 90 days; r = less then 3 h; s = 3–5 h; t = 6–7 h; u = more than 7 h).
B. Fabiano et al. / Safety Science 46 (2008) 535–544
543
Fig. 4. Perceived accident cause (legend: a = distraction; b = work pressure; c = difficulty in job performing; d = lack of specific job/safety training; e = insufficient preventive measures in plant/machine/equipment; f = non adequate personal protective equipment, g = fatalistic event).
and on-site experience. Of primary interest, in this study, was also to examine the accident cause perceived by injured workers. The distribution of perceived causes, grouped into seven main clusters, is shown in Fig. 4. Two main categories of causes can be identified: those referred to workplace conditions (c, e, f) and those referred to human active failures (a, b, d). In the 303 accidents studied, human failures, at different levels, cover the majority (55.12%). Even though the sample is limited, these findings seem consistent with those referred in the literature: in industrial accidents, human failures are estimated to be the majority, up to 70– 80% (Kosmowski and Kwiesielewicz, 2000). Among human failures, nearly 45% are perceived as connected to performance pressure. This factor has been found to increase the likelihood that employees will violate safety rules by taking short cuts (e.g., Hofman et al., 1995). The least frequent category (6.93%) was associated with personal protective equipment (PPE) unsuitable for work-place conditions. This value is consistent with results from Feeney (1986), who found that workers avoid wearing personal protective equipments where poor functional design of equipments leads to discomfort, or when it creates interference with work tasks. Whenever production is perceived to compete with safety, the temptation is to neglect safety procedures, so as to increase work output (Quinlan and Bohle, 1991). Considering that the sample of this research is mainly composed by low-level line workers, in the process and metal–mechanic sectors, their perceived work pressure, also due to temporary contract, would be likely to lead to increased psychological pressure which may, in turn, increase the probability of inducing unsafe behaviours or errors, as discussed for US grain industry workers by Dong-Chul (2005).
4. Conclusions In spite of the development of safety technique in both hardware and software and of the adoption of more stringent legislative constraints on health and safety issues, still higher consideration should be paid to the human factor. Temporary workers, engaged both in the process industries and in other productive sectors, suffer a higher frequency index than in-house workers employed in the same activities. Reasons are to be traced to lack of experience in the activity, insufficient specific knowledge (formal and informal knowledge) about a particular installation and to inadequate training period. This experimental work, with the many people involved in it, provides useful suggestions for promoting safety among non-direct employees. Some of
544
B. Fabiano et al. / Safety Science 46 (2008) 535–544
these measures may include diffusion of information on regulatory provisions, management training in safety; enhancement of specific training and formation targeted to personnel hired through staffing agencies, as well as implementation of health and safety regulations for temporary workers. Further research is required to document injury experience of temporary employees, identify training needs and devise interventions in different sectors, especially process and metal–mechanics. Moreover, it is important to conduct studies, which can embed in a model the management influences on safety in outdoor job leasing. References Artiles, A., Alos-Moner, R., 1999. Flexible employment policies and working conditions: flexibility strategies and working conditions in Spain. Report to the European Foundation, Universitad Autonoma de Barcelona ed., Spain. Asogwa, S.E., 1988. The health benefits of mechanization at the Nigerian coal corporation. Accid. Anal. Prev. 20 (1), 103–108. Bailer, A.J., Reed, L.D., Stayner, L.T., 1997. Modelling fatal injury rates using Poisson regression: a case study of workers in agriculture, forestry and fishing. J. Safety Res. 28, 177–186. Blank, V.L.G., Diderichsen, F., Andersson, R., 1996. Technological development and occupational accidents as a conditional relationship: a study over eighty years in the Swedish mining industry. J. Safety Res. 27, 137–146. Dasu, W., Panda, T., 2000. Optimization of microbiological parameters for enhanced griseofulvin production using response surface methodology. Bioprocess Eng. 22 (1), 45–49. Dong-Chul, S., 2005. An explicative model of unsafe work behavior. Safety Sci. 43 (3), 187–211. Fabiano, B., Curro`, F., Pastorino, R., 2001. Occupational injuries in Italy: risk factors and trend over the long period. Occup. Environ. Med. 58 (5), 330–338. Fabiano, B., Curro`, F., Pastorino, R., 2004. A study of the relationship between occupational injuries and firm size and type in the Italian industry. Safety Sci. 42 (7), 587–600. Feeney, R.J., 1986. Why is there resistance to wearing protective equipment at work? Possible strategies for overcoming this. J. Occup. Accid. 12, 3–20. Ferrie, J., Marmot, M., Griffiths, J., Ziglio, E., 1999. Labour Market Changes and Job Insecurity: A Challenge for Social Welfare and Health Promotion. Regional Office for Europe of the World Health organization, Denmark. Hale, A.R., 1984. Is safety training worthwhile? J. Occup. Accid. 6, 17–33. Hale, A.R., 1990. Safety rules O.K.? Possibilities and limitations in behavioural safety strategies. J. Occup. Accid. 12, 3–20. Hofman, D.A., Jacobs, R., Landy, F., 1995. High reliability process industries: individual, micro and macro organizational influences on safety performance. J. Safety Res. 26, 131–149. Holdsworth, R., 2003. Practical applications approach to design, development and implementation of an integrated management system. J. Hazard. Mater. 104 (1–3), 193–205. Istituto Nazionale per l’Assicurazione contro gli Infortuni sul Lavoro, 1970–2004. Notiziario Statistico, Roma. Kletz, T.A., 1993. Accident data – the need for a new look at the sort of data that are collected and analyzed. Safety Sci. 16 (3–4), 407–415. Kochan, T., Smith, M., Wells, J., Rebitzer, J., 1994. Human resource strategies and contingent workers: the case of safety and health in petrochemical industry. Hum. Resource Manage. 33 (1), 55–77. Kosmowski, K.T., Kwiesielewicz, M., 2000. A methodology for incorporating human factors in risk analysis of industrial systems. In: Proceedings ESREL 2000, SARS and SRA Europe Annual Conference, SRA ed., Edinburgh, pp. 351–358. Laflamme, L., Cloutier, E., 1988. Mechanization and risk of occupational accidents in the logging industry. J. Occup. Accid. 10, 191–198. Mayhew, C., Quinlan, M., 1999. The relationship between precarious employment and patterns of occupational violence: survey evidence from thirteen occupations. In: Isaksson, K., Hogstedt, C., Theorell, T. (Eds.), Health Effects of the New Labour Market. Kluwer Academic/Plenum Publishers, New York, pp. 183–205. Morris, J., 1999. Injury experience of temporary workers in a manufacturing setting. Factors that increase vulnerability. Am. Assoc. Occup. Health Nurses 47 (10), 470–478. Quinlan, M., Bohle, P., 1991. Managing Occupational Health and Safety in Australia: A Multidisciplinary Approach. Macmillan, Melbourne. Rundmo, T., 1992. Risk perceptions and safety on offshore petroleum platforms. Part II: perceived risk, job stress and accidents. Safety Sci. 15, 53–68. Saari, J., 1982. Accidents and progress of technology in Finnish industry. J. Occup. Accid. 4, 133–144. Saha, A., Ramnath, T., Chaudhuri, R.N., Saiyed, H.N., 2004. An accident risk assessment study of temporary piece rated workers. Indian J. Med. Sci. 42 (2), 240–245. Saha, A., Kulkarni, P.K., Chaudhuri, R., Saiyed, H., 2005. Occupational injuries: is job security a factor? Indian J. Med. Sci. 59 (9), 375– 381.
Safety Science 46 (2008) 535–544 www.elsevier.com/locate/ssci
A statistical study on temporary work and occupational accidents: Specific risk factors and risk management strategies Bruno Fabiano *, Fabio Curro`, Andrea P. Reverberi, Renato Pastorino Chemical and Process Engineering Department ‘‘G.B. Bonino”, University of Genoa, via Opera Pia 15, 16145 Genoa, Italy Received 23 March 2006; received in revised form 26 April 2007; accepted 4 May 2007
Abstract Temporary work, supplied by temporary-help agencies and sometimes referred to as ‘‘job in leasing”, was only recently introduced in Italy, and has since spread considerably thanks to its flexibility and cost effectiveness. In this study, trends in the rates of occupational injuries in different sectors of Italian industries in the period 2000–2004 are explored, contrasting direct employment and temporary work. Data on occupational injuries from the National Organization for Labour Injury Insurance as well as data directly obtained through a field survey in three large manufacturing firms were analysed to highlight the interaction between injury frequency index (FI) and the characteristics of the labour force. FI for temporary workers ranged between 89.23 and 94.10, i.e., between 136.4% and 175.2% more than the value found for direct employees in the most hazardous industrial sector. Also accident severity (assessed on the basis of time lost due to injuries) is twice the overall value of the severity index. The results from the field survey confirmed the trend: FI for direct employees ranges from 25.7 to 45.0 in respect of total hours of work in the range 1.99 106–2.40 106; whereas FI in temporary workers ranges from 86.2 to 163.5 in respect of total hours of work ranging from 1.25 105 to 1.28 105. Given this evidence, an empirical statistical analysis, based on responses from injured temporary workers was carried out. The questionnaire was intended to collect data on a variety of control variables relating to personal characteristics of respondents. The analysis of the questionnaire data using ANOVA and response surface methodology highlighted the interaction between short duration of work assignments, inadequate training period and FI increase. Reasons can be traced to lack of experience in the activity, insufficient specific knowledge (formal and informal knowledge) about a particular installation and to inadequate training period. Means of promoting safety include diffusion of information on regulatory provisions, management training in safety, enhancement of specific training and formation. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Frequency index; Injury; Questionnaire; Response surface methodology; Temporary work; Training
*
Corresponding author. Tel.: +39 0103532585. E-mail address: [email protected] (B. Fabiano).
0925-7535/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.ssci.2007.05.004
536
B. Fabiano et al. / Safety Science 46 (2008) 535–544
1. Introduction On a national level, Italy has enjoyed declining trends in the number and rates of occupational fatalities and injuries for more than three decades. While these figures are encouraging, far too many preventable injuries are still occurring. The factors influencing accident frequency can be divided into the following categories: technical factors: low automation, multi-product industries, discontinuous operating cycles, non standardized production affect safety negatively, since they require a greater interaction between man and devices. On the other hand, a reduction in individual exposure to severe hazards was reported where mechanized processes and equipment were introduced in the mining (Asogwa, 1988) and the logging industry (Laflamme and Cloutier, 1988); economical factors, e.g., general economic situation (Saari, 1982), unemployment rate, labour and socialinsurance legislation, (Blank et al., 1996); labour organization, e.g., management system and performance, work practice, oversight, communication structure, etc.; environmental conditions: about half the accidents in Italy are related to labour environment (Fabiano et al., 2001) and they could be prevented by rather simple lay-out and protection measures, which however prove extremely difficult or even unfeasible in small concerns, because of operating, economic and/or space constraints (Fabiano et al., 2004); human factors, both individual and inter-individual, e.g., workload, experience and training, competence, fatigue, etc. Current market conditions often make it necessary to resort to outsourcing to remain competitive, particularly by employing external and precarious human resources. In fact, in the last 20 years there has been a significant growth in labour governed by casual, part-time, subcontract or franchised arrangements, virtually in all OECD countries (less so in Italy, Canada and Luxenbourg) (Ferrie et al., 1999). Results of international research seem to point out that precarious labour is associated with increased fatalities, occupational injuries, illness, in various industry sectors across a number of countries (Kochan et al., 1994; Morris, 1999; Mayhew and Quinlan, 1999). Temporary work (i.e., a job arrangement whereby workers are hired out to firms by Temporary Work Agencies) is a sort of labour outsourcing and precarious work. This kind of work arrangement, sometimes referred to as ‘‘job in leasing”, has only recently become widespread in Italy. Due to the flexibility and cost effectiveness of this contract, in the last few years Italy has experienced constant growing rates of temporary work. The whole number of contracts provided by temporary work agencies soared from 194,835 in 1999 to 744,438 in 2003, corresponding to an average increase of about 56.4% per year. Virtually, every industrial sector shared this trend, with an average daily percentage of temporary workers corresponding to 1% of ‘‘in-house employees”, for all industrial sectors in the year 2003. Process, metal–mechanic and transport sectors displayed the highest average daily percentage of temporary workers, amounting to about 3% of direct employees. In the same year, the average EC daily incidence of temporary work was 1.5%, with higher values in the UK (3.2%) and in the Netherlands (4.5%), while the daily incidence in the United States was 2.5%. Despite the international diffusion of temporary work, in the open literature there is little information on possible safety implications and contributing factors of occupational injuries. Moreover, even if most safety-related regulations and safety management systems already incorporate provisions for the management of contractors, a striking number of injuries occurs among ‘‘non in-house employees”. In Spain, 60% of occupational injuries occurred to temporary workers, with a rate nearly three times that of those in permanent jobs (Artiles and Alos-Moner, 1999). A survey conducted among three manufacturing industries of the plastic sector in the USA, showed that the trend of accident rate over the years 1993–1998 does not appear to be a steady one, either for direct employees, or for temporary workers. However, the evidence obtained from such research indicates that the injury rate for temporary workers is constantly higher (from 2.6 to 3.8 times) than the one recorded for permanent employees (Morris, 1999). This paper offers a perspective on the trends in the rates of occupational injuries in the main sectors of Italian industries, during the period 2000–2004, contrasting direct employment and temporary work and
B. Fabiano et al. / Safety Science 46 (2008) 535–544
537
focusing attention on the human factor. Generally speaking, the different reasons for people failure can be grouped as follows: lack of training or instruction; lack of motivation; lack of physical or mental ability; slips and lapses of attention (Kletz, 1993). On this basis, an investigation into the relationship between occupational accidents and temporary work is performed, adopting a questionnaire survey, for the definition of peculiar risk factors and for setting priorities to improve safety standards in this context. 2. Materials and methods 2.1. Research design Raw data were obtained from the National Organization for Labour Injury Insurance (I.N.A.I.L. Istituto Nazionale per l’Assicurazione contro gli Infortuni sul Lavoro) (INAIL, 1970–2004). One of the main measures of performance is the lost time accident (LTA), which however has only limited value and should be supplemented by other measurements, such as the total accident rate, the cost of the damage caused by accidents and other dangerous incidents and, if possible, a numerical measure of the results of plant audits (Kletz, 1993). The frequency index (FI), allows to find out the incidence of accidents of workers exposed to risk and is defined as follows: FI ¼
Number of total accidents 6 10 : Number of worked hours
ð1Þ
As ‘‘worked hours” can be calculated regularly, this index may offer a proper instrument for evaluation of accidents. To estimate trends, we modelled the rate of injury frequency index as a function of the year, by adopting a multiple regression model with a log transformed rate (Bailer et al., 1997; Fabiano et al., 2001), as follows ln FI ¼ a þ bðyeari year0 Þ;
ð2Þ
where year0 is the first year of the study, to which corresponds the baseline of injury rate. Data on temporary workers were collected directly from the 16 leading temporary-job agencies in Italy, covering nearly 90% of the whole market. Data from INAIL and from direct field survey (temporary job agencies) were combined with data from the State Statistics Institute, in order to highlight the interaction between the trend of the injury frequency index and the characteristics of the labour force, throughout the years from the introduction of the new work arrangements. The critical issue is whether there is a statistical relation between occupational injuries and temporary work, and whether or not such relation differs for different occupation groups. In order to confirm the impact of new working regulations on the occurrence of occupational injuries, a detailed case study was made on three large firms of the manufacturing sector, characterized by high proportion of temporary work. Finally, an empirical statistical analysis was carried out, based on responses by injured temporary workers to an ‘‘ad-hoc” questionnaire specially designed for the purpose. 2.2. Questionnaire development The questionnaire was intended to collect data on a variety of control variables relating to personal characteristics of respondents. The main section of the questionnaire was designed to characterize technical, organizational and individual factors of suffered injuries, as well as to find out how firms try to reduce the occupational risk to which temporary employees may be exposed. The content of the questionnaire is described below. Demographic and work details: respondents were asked to give information on their sex, age, job title, employer, extent of their job experience and specific work task, and task rotation if any. Safety attitudes: respondents were asked to rate their job and safety training, extent of the temporary contract, whether or not they were accompanied by a job tutor, how much support they received from fellow workers on the job.
538
B. Fabiano et al. / Safety Science 46 (2008) 535–544
Accident and near misses: respondents were asked to describe the accident and its severity based on lost time accident (LTA), time elapsed from starting of the contract until the accident, perceived accident cause (distraction, work pressure, difficulty in job performing, lack of specific/safety training, insufficient preventive measures in plant/machine/equipment, inadequate personal protective equipment, fatalistic event). Data from the structured questionnaire were coded and entered into a database for subsequent univariate analysis of variance (ANOVA). The independent variables whose effects on number of injuries and severity were evaluated included: worker’s age, job position, training period, on-site experience, temporary contract duration, perceived accident cause. Significant results were analysed adopting response surface methodology (RSM), this latter being a collection of mathematical and statistical techniques widely adopted to optimise different processes (e.g., Dasu and Panda, 2000), used in accident analysis for the first time, at least to our knowledge. 2.3. Sample The study population was identified from 16 staffing companies in Milan, Italy, covering nearly 90% of the whole market of temporary work. The questionnaire was distributed to 700 injured workers resulting from the investigation. All ‘‘in itinere” accidents (accidents occurred on the way to or from the workplace) were excluded from the sample. Respondents were informed that the interview data were to be treated as anonymous and that only aggregated results would be produced. The response rate for this distribution was 43.3% (n = 303). The average age of all respondents was 29.2 years (SD = 8.23) in the range from 18 to 45 years. 3. Results and discussion 3.1. Injury trends As shown in Fig. 1, the overall value of FI on a national level decreased, in the last three decades, from 94.92 in 1970 to 22.54 in 2000, with a mean reduction per year corresponding to 5.48%, obtained from the fit of the regression model (Eq. (2)). This encouraging trend was common to every industrial sector, with some differences depending on sector. It must be pointed out that raw data from INAIL obviously do not include injuries in case of irregular or unregistered employment (e.g. moonlighting phenomena), as well as accidents
Fig. 1. Overall injury frequency index (FI) over the period 1970–2000 and regression model with prediction intervals (95%).
B. Fabiano et al. / Safety Science 46 (2008) 535–544
539
involving less than three days absence from work, since, in Italy, they need not be reported. However, these limitations concern not only the Italian reality, but generally the worldwide safety community; in fact, the problem of improving the organizational memory and the need of a new approach to the collection and analysis of accident data are becoming more and more important (Kletz, 1993). Table 1 shows the values of the national frequency index and mean values over five years, for different industrial sectors (in-house workers) and for temporary workers, over the period 2000–2004, i.e., since the introduction in Italy of temporary-work arrangements. Again, a general decline may be observed in the frequency of injuries, for direct employees, across all industries, from 2000 until 2004, at an average estimated rate of decline of 4.29% per year, from a baseline of 22.54 in 2000. Practically speaking, all sectors are reaching minimum values of the frequency index, which can be regarded as a statistically reinforced reference for the minimum risk levels per individual employed in the different industrial activities. Obviously, this is only a statistical consideration, not to be regarded as an attempt to renounce the effort of reducing even the lowest injury frequency recorded. The trend for temporary work is not decreasing, but shows some variations from year to year, mainly because of the different percentage distribution of temporary workers in the various industrial sectors. However, it should be pointed out that the FI value for temporary workers is 2.65 times the mean value for the highest risk sector at national level (Building sector). It is worth noting that the average FI value of temporary workers over the time span considered is comparable to the value of the national frequency index for direct employees recorded in 1971, i.e., 91.83 (see Fig. 1). In order to offer a better insight on this observation, Table 2 shows FI values calculated for the Lombardia region, where nearly 63% of temporary workers were employed in the years considered. FI values in Lombardia for direct employees are constantly lower than the corresponding values at national level, with mean differences ranging from 5.35% in the food sector to 20.68% in the process sector. Considering this difference in the reference region, the injury trend declines constantly by 3.76% per year and FI of temporary workers is by far higher than FI of the most hazardous sector, to a percentage of 410.9%. These findings are in good agreement with the few studies available in the international literature (Artiles and Alos-Moner, 1999). A case-study conducted over five years in a process industry in India concluded that accident incidence, accident frequency and severity rates were significantly higher for temporary workers than for permanent workers, although the two groups were similar in relation to age, level of education, habits and nature of work. Relative risk varied from 2.3 to 18.0 in case of time loss accidents and from 1.1 to 2.6 in case of no time-loss accidents (Saha et al., 2004). Higher accident risk of temporary workers might have been due to less effective experience and to lack of job security (Saha et al., 2005). The results presented here underline that temporary work may be considered a risk factor. Reasons are to be traced to its intrinsic characteristics: age of workers, lack of experience in the activity, insufficient specific knowledge, low level of experience in the same firm, lack of a sufficient training period and heavy workload. In order to confirm the impact of innovative employment arrangements on the occurrence of occupational injuries, a detailed case study was carried out on three large firms of the manufacturing sector, characterized by a high proportion of temporary work. The three firms belong to the metal and mechanic sector and their employees have similar characteristics: number of hours worked, percentage of temporary workers, position
Table 1 Injury frequency index (FI) calculated, on a national level, for direct workers in the different industrial sectors and for temporary workers Mean
Year 2004
2003
2002
2001
2000
Metal and mechanic Process Food Services Building All sectors
34.07 ± 3.07 19.39 ± 1.69 22.61 ± 1.49 20.18 ± 1.51 34.77 ± 2.78 20.76 ± 1.42
30.09 17.25 20.62 18.76 32.02 18.98
32.08 18.11 21.90 18.93 32.76 19.70
34.15 19.60 22.60 19.87 34.19 20.87
36.52 21.02 23.38 20.91 35.93 21.70
37.51 20.98 24.56 22.41 38.95 22.54
Temporary work
91.63 ± 2.22
89.42
89.23
94.10
93.30
92.10
540
B. Fabiano et al. / Safety Science 46 (2008) 535–544
Table 2 Injury frequency index (FI) calculated, in the reference region (Lombardia), for direct workers in the different industrial sectors Industrial sector
Metal and mechanic Process Food Services Building All sectors
Mean
Year
28.35 ± 2.70 15.16 ± 1.19 21.31 ± 0.97 18.08 ± 1.28 29.16 + 2.05 17.84 ± 1.27
2005
2004
2003
2002
2001
2000
25.39 14.03 20.88 17.59 26.95 16.60
25.88 13.89 20.67 17.12 28.63 16.79
27.27 14.53 20.39 16.88 27.10 16.99
28.62 15.49 21.26 17.91 29.76 18.07
30.88 16.40 21.56 18.64 30.12 18.84
32.08 16.61 23.10 20.36 32.39 19.76
and tasks within the firm. The only remarkable difference regards the training period and the tutoring applied to temporary workers: firms A and C developed and implemented an integrated management system encompassing quality (ISO 9000), environmental management (ISO 14000) and safety and health. This integrated management system includes a specific job-training methodology that applies to sub-contractors and temporary workers hired through staffing agencies. Firm B established separate management systems for environment, health and safety and quality, with standard training of personnel addressed to both ‘‘in house” workers and temporary workers. Table 3 shows the results of the statistical analysis by direct survey in these three firms: the injury frequency index for temporary workers in firms A and C appears to be 1.9 times FI of direct employees. This figure is consistent with results from Morris, 1999), which indicate that the injury rate for temporary workers is constantly higher (from 2.6 to 3.8 times) than the one recorded for permanent employees in three manufacturing industries. Firm B scored a strikingly high difference, with FI referred to temporary workers 6.35 times FI of direct employees. As shown in Table 3, in the three firms examined, also severity (LTA) resulted higher than the value calculated for direct employees by a percentage from 87% to 97%. Even if limited, the experience of three Italian firms seems to show how a robust health, safety and quality management system, including different training for ‘‘in-house” and temporary workers can affect positively safety performance. Furthermore, management system integration avoids duplication or conflicting of procedures established for quality management with similar procedures for health and safety (Holdsworth, 2003). Clearly, in the area of safety training and education, the effectiveness of intervention should be carefully checked at an in-depth level of analysis (Hale, 1984), especially when dealing with new and precarious job arrangements.
3.2. Questionnaire statistics In order to test the main effects of workers age, job position, employment sector, training period, on-site experience, duration of temporary assignment and perceived accident cause on injury probability and injury severity, a univariate analysis of variance (ANOVA) was carried out. Industrial sector and job position of the sample are shown, respectively, in Tables 4 and 5. We may notice that the sample of this research is composed for 89.21% by low-level line workers, mainly in the process and metal–mechanic sectors (83.2%). No significant differences emerged on the workers age, job position, assignment duration and perceived cause. Three
Table 3 Injury frequency index and severity for in-house workers and temporary workers in three metal–mechanics settings Reference firm A
Worked hours Number of injuries FI Severity (LTA)
Reference firm B
Reference firm C
In house workers
Temporary workers
In house workers
Temporary workers
In house workers
Temporary workers
2,401,210 106 44.14 60
125,898 11 87.4 112
2,135,544 55 25.75 70
128,424 21 163.52 138
1,998,272 90 45.04 76
127,590 11 86.21 142
B. Fabiano et al. / Safety Science 46 (2008) 535–544
541
Table 4 Employment industrial sector of respondent temporary workers Industrial sector
Number
Process Metal and mechanics Food Services Building
47 205 16 23 12
% 15.51 67.66 5.28 7.59 3.96
Total
303
100.00
Table 5 Job position of respondent temporary workers Number
%
Skilled worker Unskilled worker Storeman Other
45 203 37 18
14.85 67.00 12.21 5.94
Total
303
100.00
interactions were statistically significant (see Table 6), namely industrial sector (p < 0.01%), training period (p < 0.05%) and on site experience (p < 0.05%). Five categories of on-site experience were entered (less than 7 days; 7–30 days; 31–60 days; 61–90 days; more than 90 days) and four categories of training period (less then 3 h; 3–5 h; 6–7 h; more than 7 h). Industrial sector categories were as follows: process, metal and mechanics, food, building, services. Fig. 2 shows the three-dimensional graph representing the response surface for injury probability as a function of training and on-site experience. The significant interaction of the independent variables indicates that an increase of the training period (professional training and job tutoring) greatly reduces the probability of injury. It must be noted that safety programs include training as a part of the risk management process. However, the implementation of rules and training may often prove to be insufficient in reducing unsafe practices, since safety rules are often considered to apply only in certain situations while being impossible to follow in the many exceptional situations which are seen to be the reality of the shop floor situation (Hale, 1990). Equally, it seems that the worker’s permanency on the same job site entails an increase of experience and acquaintance with one’s duties, reducing the probability of an accident. In other words, even if employees are initially unaware of occupational risks, they can often acquire on-the-job experience. A key aspect is that, as reported by Asogwa (1988), an adaptation period is required for workers to perform adequately in new work assignments and in a changed environment, whereas in conditions of production intensification/pressure (as are those usually corresponding to the utilization of temporary workers) such training period can be curtailed or eliminated. Moreover, intensification of work and production in combination with an increase in overtime was referred to constitute a set of interacting factors increasing injury risk (Saari, 1982). Available evidence suggests that workers have some idea about risk inherent in their chosen occupation,
Table 6 Univariate tests of significance
Intercept Industrial sector Training On-site experience Error
Sum of squares
DF
Mean square
F value
P value
64.6354 59.0997 17.9502 27.7416 302.5752
1 4 3 4 137
64.63538 14.77492 5.98341 6.93540 2.20858
29.26560 6.68979 2.70917 3.14021
0.000000 0.000060 0.047606 0.016589
Sigma-restricted parameterization.
542
B. Fabiano et al. / Safety Science 46 (2008) 535–544
Fig. 2. Response surface for injury probability, as a function of training and on-site experience (legend: a = less than 7 days; b = 7–30 days; c = 31–60 days; d = 61–90 days; e = more than 90 days; r = less then 3 h; s = 3–5 h; t = 6–7 h; u = more than 7 h).
even if this information is rarely perfect or processed correctly (Rundmo, 1992). The ANOVA results for accident severity (LTA) show a statistical significance on training period (p < 0.05%) and on site experience (p < 0.05%). Fig. 3 shows the response surface plot of injury severity (LTA, days) as a function of training
Fig. 3. Response surface plot of injury severity (LTA, days), as a function of training and on-site experience (legend: a = less than 7 days; b = 7–30 days; c = 31–60 days; d = 61–90 days; e = more than 90 days; r = less then 3 h; s = 3–5 h; t = 6–7 h; u = more than 7 h).
B. Fabiano et al. / Safety Science 46 (2008) 535–544
543
Fig. 4. Perceived accident cause (legend: a = distraction; b = work pressure; c = difficulty in job performing; d = lack of specific job/safety training; e = insufficient preventive measures in plant/machine/equipment; f = non adequate personal protective equipment, g = fatalistic event).
and on-site experience. Of primary interest, in this study, was also to examine the accident cause perceived by injured workers. The distribution of perceived causes, grouped into seven main clusters, is shown in Fig. 4. Two main categories of causes can be identified: those referred to workplace conditions (c, e, f) and those referred to human active failures (a, b, d). In the 303 accidents studied, human failures, at different levels, cover the majority (55.12%). Even though the sample is limited, these findings seem consistent with those referred in the literature: in industrial accidents, human failures are estimated to be the majority, up to 70– 80% (Kosmowski and Kwiesielewicz, 2000). Among human failures, nearly 45% are perceived as connected to performance pressure. This factor has been found to increase the likelihood that employees will violate safety rules by taking short cuts (e.g., Hofman et al., 1995). The least frequent category (6.93%) was associated with personal protective equipment (PPE) unsuitable for work-place conditions. This value is consistent with results from Feeney (1986), who found that workers avoid wearing personal protective equipments where poor functional design of equipments leads to discomfort, or when it creates interference with work tasks. Whenever production is perceived to compete with safety, the temptation is to neglect safety procedures, so as to increase work output (Quinlan and Bohle, 1991). Considering that the sample of this research is mainly composed by low-level line workers, in the process and metal–mechanic sectors, their perceived work pressure, also due to temporary contract, would be likely to lead to increased psychological pressure which may, in turn, increase the probability of inducing unsafe behaviours or errors, as discussed for US grain industry workers by Dong-Chul (2005).
4. Conclusions In spite of the development of safety technique in both hardware and software and of the adoption of more stringent legislative constraints on health and safety issues, still higher consideration should be paid to the human factor. Temporary workers, engaged both in the process industries and in other productive sectors, suffer a higher frequency index than in-house workers employed in the same activities. Reasons are to be traced to lack of experience in the activity, insufficient specific knowledge (formal and informal knowledge) about a particular installation and to inadequate training period. This experimental work, with the many people involved in it, provides useful suggestions for promoting safety among non-direct employees. Some of
544
B. Fabiano et al. / Safety Science 46 (2008) 535–544
these measures may include diffusion of information on regulatory provisions, management training in safety; enhancement of specific training and formation targeted to personnel hired through staffing agencies, as well as implementation of health and safety regulations for temporary workers. Further research is required to document injury experience of temporary employees, identify training needs and devise interventions in different sectors, especially process and metal–mechanics. Moreover, it is important to conduct studies, which can embed in a model the management influences on safety in outdoor job leasing. References Artiles, A., Alos-Moner, R., 1999. Flexible employment policies and working conditions: flexibility strategies and working conditions in Spain. Report to the European Foundation, Universitad Autonoma de Barcelona ed., Spain. Asogwa, S.E., 1988. The health benefits of mechanization at the Nigerian coal corporation. Accid. Anal. Prev. 20 (1), 103–108. Bailer, A.J., Reed, L.D., Stayner, L.T., 1997. Modelling fatal injury rates using Poisson regression: a case study of workers in agriculture, forestry and fishing. J. Safety Res. 28, 177–186. Blank, V.L.G., Diderichsen, F., Andersson, R., 1996. Technological development and occupational accidents as a conditional relationship: a study over eighty years in the Swedish mining industry. J. Safety Res. 27, 137–146. Dasu, W., Panda, T., 2000. Optimization of microbiological parameters for enhanced griseofulvin production using response surface methodology. Bioprocess Eng. 22 (1), 45–49. Dong-Chul, S., 2005. An explicative model of unsafe work behavior. Safety Sci. 43 (3), 187–211. Fabiano, B., Curro`, F., Pastorino, R., 2001. Occupational injuries in Italy: risk factors and trend over the long period. Occup. Environ. Med. 58 (5), 330–338. Fabiano, B., Curro`, F., Pastorino, R., 2004. A study of the relationship between occupational injuries and firm size and type in the Italian industry. Safety Sci. 42 (7), 587–600. Feeney, R.J., 1986. Why is there resistance to wearing protective equipment at work? Possible strategies for overcoming this. J. Occup. Accid. 12, 3–20. Ferrie, J., Marmot, M., Griffiths, J., Ziglio, E., 1999. Labour Market Changes and Job Insecurity: A Challenge for Social Welfare and Health Promotion. Regional Office for Europe of the World Health organization, Denmark. Hale, A.R., 1984. Is safety training worthwhile? J. Occup. Accid. 6, 17–33. Hale, A.R., 1990. Safety rules O.K.? Possibilities and limitations in behavioural safety strategies. J. Occup. Accid. 12, 3–20. Hofman, D.A., Jacobs, R., Landy, F., 1995. High reliability process industries: individual, micro and macro organizational influences on safety performance. J. Safety Res. 26, 131–149. Holdsworth, R., 2003. Practical applications approach to design, development and implementation of an integrated management system. J. Hazard. Mater. 104 (1–3), 193–205. Istituto Nazionale per l’Assicurazione contro gli Infortuni sul Lavoro, 1970–2004. Notiziario Statistico, Roma. Kletz, T.A., 1993. Accident data – the need for a new look at the sort of data that are collected and analyzed. Safety Sci. 16 (3–4), 407–415. Kochan, T., Smith, M., Wells, J., Rebitzer, J., 1994. Human resource strategies and contingent workers: the case of safety and health in petrochemical industry. Hum. Resource Manage. 33 (1), 55–77. Kosmowski, K.T., Kwiesielewicz, M., 2000. A methodology for incorporating human factors in risk analysis of industrial systems. In: Proceedings ESREL 2000, SARS and SRA Europe Annual Conference, SRA ed., Edinburgh, pp. 351–358. Laflamme, L., Cloutier, E., 1988. Mechanization and risk of occupational accidents in the logging industry. J. Occup. Accid. 10, 191–198. Mayhew, C., Quinlan, M., 1999. The relationship between precarious employment and patterns of occupational violence: survey evidence from thirteen occupations. In: Isaksson, K., Hogstedt, C., Theorell, T. (Eds.), Health Effects of the New Labour Market. Kluwer Academic/Plenum Publishers, New York, pp. 183–205. Morris, J., 1999. Injury experience of temporary workers in a manufacturing setting. Factors that increase vulnerability. Am. Assoc. Occup. Health Nurses 47 (10), 470–478. Quinlan, M., Bohle, P., 1991. Managing Occupational Health and Safety in Australia: A Multidisciplinary Approach. Macmillan, Melbourne. Rundmo, T., 1992. Risk perceptions and safety on offshore petroleum platforms. Part II: perceived risk, job stress and accidents. Safety Sci. 15, 53–68. Saari, J., 1982. Accidents and progress of technology in Finnish industry. J. Occup. Accid. 4, 133–144. Saha, A., Ramnath, T., Chaudhuri, R.N., Saiyed, H.N., 2004. An accident risk assessment study of temporary piece rated workers. Indian J. Med. Sci. 42 (2), 240–245. Saha, A., Kulkarni, P.K., Chaudhuri, R., Saiyed, H., 2005. Occupational injuries: is job security a factor? Indian J. Med. Sci. 59 (9), 375– 381.