You are on page 1of 7

Journal of Safety Research 41 (2010) 1–7

Contents lists available at ScienceDirect

Journal of Safety Research


j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / j s r

Analysis of work related accidents in the Spanish mining sector from 1982-2006
Lluís Sanmiquel a,⁎, Modesto Freijo b, Joaquín Edo a, Josep M. Rossell c
a
Dpto. de Ingeniería Minera y Recursos Naturales, Universitat Politècnica de Catalunya (UPC), Avda. Bases de Manresa, 61-73, 08242-Manresa (Barcelona), Spain
b
Dpto. de Ingeniería Eléctrica, Universitat Politècnica de Catalunya (UPC), Avda. Bases de Manresa, 61-73, 08242-Manresa (Barcelona), Spain
c
Dpto. de Matemática Aplicada III, Universitat Politècnica de Catalunya (UPC), Avda. Bases de Manresa, 61-73, 08242-Manresa (Barcelona), Spain

a r t i c l e i n f o a b s t r a c t

Available online 27 January 2010 Introduction: The rate for work related accidents in the Spanish mining sector is notably higher than in other
countries such as the United States. It produces a very negative impact on the mining industry. This paper is the
Keywords: report of a study on serious and fatal accidents in Spanish mining from 1982-2006. It is based on the reports of
Mining 212 accidents (serious or fatal) carried out by the General Management of Energy and Mining of Catalonia
Incident index (Spain). Method: The high work-related accident rate in the Spanish mining sector makes it necessary to carry
Event
out an analysis and research that can shed light on the causes of this high rate; this is the only way that a
Contributing factor
Skill-based
solution can be found. The study is based on Feyer and Williamson's analysis of accident causes, as they apply to
212 accidents. The types and causes of the accidents are coded according to the coding system used by the
Spanish National Institute for Safety and Hygiene in the Workplace, which allows us to identify a series of direct
causes and contributing factors in different accidents. Results If all the causes and factors that are present in the
accidents are known, we are able to look for appropriate solutions to reduce them as much as possible. In short,
we are able to come up with a series of conclusions that expose the weak links in the management of accident
prevention in companies. This is helpful in the struggle to reduce work injuries in the Spanish mining sector.
© 2010 National Safety Council and Elsevier Ltd. All rights reserved.

1. Introduction the last 10 years (Rubio, 2007) have taken action in order to reduce the
high rate of work injuries. The incidence index of the total economic
Despite improvements in work-related safety in the last decade, sectors in Spain regarding workplace accidents was approximately
about 5,500 people die every year in work accidents in the European 6,000 accidents per 100,000 workers in 2006. This is according to the
Union, and more than 75,000 are so seriously disabled that they Spanish National Institute for Safety and Hygiene in the Workplace
cannot work again (Saari, 2001). A basic human right is to be able to (INSHT), while in the 15 state EU, in 2004, the amount was 3,200
go home from work safe and sound; nobody should die or be injured (EUROSTAT). In the United States the amount was 5,000 (National
in a workplace accident. However, in order to achieve this, there is still Institute for Occupational Safety and Health, USA).
much work to be done in the field of accident prevention. People have In the total number of Spanish economic sectors, mining is one of
more physical problems in the workplace than ever before, and this those with the highest incidence in number of accidents per 100,000
goes against the widespread belief that new technology and workers. Thus, in 2005, mining had an incidence index that was 5.5
automation have eradicated difficulties such as manual lifting of times higher than the total economic sectors (National Institute of
heavy objects (Crichton, 2005). Lack of information and little Safety and Hygiene in Spanish Work (INSHT)).
technological training contribute to mistakes that cause accidents If the indexes of work accidents in Spanish mining are compared
and work incidents (Cox & Tait, 1991). with those of other countries, we can see that the values are also much
Work fatalities are generally higher in Spain than in other parts of higher. The incidence indexes per 100,000 workers in the Spain
the European Union (EU). Spain is, in fact, second to Portugal regarding mining industry, U.S mining industry, and in the state of Queensland
the number of injuries per work accident. This includes the total (Australia) are shown in Table 1.
amount of accidents with sick leave and fatalities (EUROSTAT). There Table 1 shows how work accidents in Spanish mining have
are approximately three deaths and 3,000 injuries with sick leave in decreased continuously since 2000. However, the 2005 rate is still
Spain every day, which means 20 million lost workdays per year much higher than the United States and the State of Queensland
(Santamaría, Catot, & Benevides, 2006). The Spanish governments of (Australia). They are specifically 9.6 times higher than those of the
United States and 27.6 times higher than those of Queensland.
The high work related accident rate in the Spanish mining sector
⁎ Corresponding author.
implies that research should be carried out in this field in order to
E-mail addresses: sanmi@emrn.upc.edu (L. Sanmiquel), freijo@ee.upc.edu show which aspects of the mining companies' prevention systems
(M. Freijo), edo@emrn.upc.edu (J. Edo), josep.maria.rossell@upc.edu (J.M. Rossell). have to be improved. That is the basic aim of this study.

0022-4375/$ – see front matter © 2010 National Safety Council and Elsevier Ltd. All rights reserved.
doi:10.1016/j.jsr.2009.09.008
2 L. Sanmiquel et al. / Journal of Safety Research 41 (2010) 1–7

Table 1 accidents with n2 = 139 cases. Now, by applying the previous formula,
Incidence Index of the mining sector of Spain, The United States and the state of we obtain: For n1 = 73 ⇒ ε = 11.62%, and for n2 = 139 ⇒ ε = 7.06%,
Queensland (Australia).
approximately. These percentages can be considered not high
Year Incidence Incidence Index Incidence Index estimation errors in the samples.
Index Spain United States Queensland (Australia)

1996 42978.4 3927.8 4222.7 2.2. The Classification System


1997 43952.0 3717.3 3536.7
1998 38762.1 3801.7 2927.3
Full details of the classification system are described elsewhere
1999 43484.6 3489.7 2951.3
2000 41802.1 3493.2 2557.3 (Feyer & Williamson, 1991). Briefly speaking, the system was
2001 40461.3 3206.1 2136.4 designed to allow coding of a sequence of up to three events that
2002 35683.7 3183.5 1497.2 immediately preceded the accident leading to the fatality. These were
2003 33646.5 2913.9 1406.8 called Precursor Events (abbreviated to PE1, PE2, and PE3). Each of
2004 32508.6 2770.0 1029.9
these events could be classified in one of four possible ways:
2005 25847.0 2689.3 935.7

Source: National Institute for Occupational Safety and Health of the U.S.A and Queensland 1. Environmental events: events or conditions resulting from the
Mines and Quarries Safety Performance and Health Report. The Spanish values calculated location of the accident; these conditions could not have been
from the annual number of mild, serious and fatal mining accidents and the number of
changed at that point in time (e.g., low lighting, wet floor, or
workers in the mining sector. The accident data comes from the INSHT and the number of
annual workers of the Mining Statistics of the Spanish Ministry of Industry, Tourism and cramped conditions).
Commerce. The office employees have been excluded. 2. Equipment events: events resulting from breakage or malfunction
of machinery or tools that occur at that point in time.
3. Medical events: events resulting from the person's current state of
Once the accident has taken place, the investigation of what physical well-being (e.g., heart attack or diabetic or epileptic
happened is a very important source of information. There is usually episode).
not just one reason for an accident, but a series of causes that are 4. Behavioral events: events resulting directly from human involve-
closely related. An adequate investigation of an accident allows us to ment (e.g., leaning too far into the path of machinery, touching an
determine all of its causes and their influence. Thus, if the causes of an electrically charged object).
accident are known, methods of protection and prevention can be
Behavioral events were coded further into whether they constituted
designed and applied to eliminate them, and in this way the
an error or not (i.e., the incorrect performance of standard operating
possibility of recurrence would be none to very low (Piqué, 1991).
procedure). Errors were classified into each of two well-known
There are different ways to classify the causes of workplace accidents,
classification systems. The first coded error into Omissions (things not
one of them is the method described by Feyer and Williamson (1991).
done) or Commissions (things done incorrectly). The second coding
This system was designed to allow the coding of a sequence of up to
method coded error into skill-based errors, or errors during routine
three events that predict the serious or fatal accident.
behavior, rule-based errors, or errors involving the application of
learned rules (i.e., if X occurs then do Y), and knowledge-based errors, or
2. Methods errors in troubleshooting or solving problems.
The nature of contributing factors was also coded into eight
2.1. Study Population possible categories:

This paper analyzes work related accidents in the Spanish mining 1. Environmental: factors resulting from conditions occurring at an
sector in the years 1982-2006. The reports on the 212 serious and fatal earlier time at the location of the accident.
accidents that took place in Spanish mining between 1982 and 2006 2. Equipment: factors associated with the design or upkeep of
are the starting point. The accidents described in this paper are those machinery, tools, personal protective equipment, or safety
that took place in mining work centers, within regular work hours equipment.
(the so-called “in itinere” accidents are not taken into account), and 3. Work practice: factors involving risky standard operating proce-
that are classified as serious or fatal by competent medical staff. dures accepted by management and/or personnel (this included
The serious accidents always imply serious injuries for the affected categories of poor upkeep or misuse of equipment).
worker. This means that the worker loses many workdays. Thus, in the 4. Supervision: factors relating to inadequate charge of workers.
non-energetic Spanish mining industry, between 1999 and 2006, the 5. Training: factors relating to inadequate training of workers.
average number of missed workdays from serious accidents was 104 6. Task error: factors relating to incorrect performance of duty.
(INSHT), while in the energetic mining industry the average time lost 7. Medical: factors involving physical well-being at an earlier time.
for serious accidents was 89 days (INSHT). 8. Other: factors such as alcohol/drug involvement, delays in reach-
Mild accidents include those that caused the injured worker to ing medical attention, and social aspects.
miss at least one workday. The average duration of mild accidents in
non-energetic and energetic mining was, respectively, 24 and 25 2.3. Reliability
missed workdays (INSHT) for the time period between 1999 and
2006. First, we should note that the reports were made by different
Statistical research shows that to determine the sample size of a coroners, for different kinds of mines, at different times, and written
finite population and working with 95% certainty (i.e., with the in a large variety of report styles. Under these considerations, to
significance level α = 0.05), the relation among the total population obtain an objective pattern to classify mining accidents is not a trivial
(N), the sample size (n), and the estimation error (ε) is given by task.
N
From the information extracted in the coroners' reports, a main
n= ε2 ðN−1Þ + 1
. Since the total population was N = 5000 and the coder proposed two coding instruments in order to codify the
available sample size was n = 212, the estimation error is approximately accidents. To test these coding instruments, three qualified coders
ε = 6.72% (Box, Hunter, & Hunter, 2005; Oncins de Frutos, 1991). (the main coder and two coworkers) analyzed a random sample of 35
The initial sample size, n = 212, will be divided into two subgroups: serious and fatal mining accidents, which were obtained from the 212
underground mining accidents, with n1 = 73 cases and surface registered accidents. The results were discussed and two classification
L. Sanmiquel et al. / Journal of Safety Research 41 (2010) 1–7 3

patterns were established. The first one corresponds to the Precursors The last line in Table 2 gives us the error committed in coding one
Events and the second corresponds to the Contributed Factors. Since it specific accident by two independent coders (225 scores in the
was necessary to prove if the coding instruments had good reliability, example). Taking into account this criterion, the possible maximum
a test evaluating the extent of the lack of concordance among error between two coders, for each analyzed accident, is 600 scores.
independent coders was made. Obviously, it was desired that Therefore, the highest error in coding the 212 registered accidents
someone else who codifies the same data would obtain similar results could be 127,200 scores.
and conclusions. Example: Table 2 shows the coding of precursor events
In the first classification pattern, three possible precursor events corresponding to a fatal accident that has been analyzed by two
were considered. For each precursor event, four possible categories independent coders C1 and C2.
with an intensity level were used (see Table 2). The intensity levels Having these ideas in mind, when the 212 accidents were analyzed
were evaluated with 100, 75, and 50 scores. The classification method and coded by three coders (main coder, C1 and C2), the percentage of
penalizes the difference (error) among coders. At the end of this agreement between C1 and C2 was 83% and the percentages of
process, the total error between two coders gives the degree of agreement of the coders C1 and C2 with the main coder were 86% and
agreement-disagreement achieved. In this case, the maximum 83%, respectively. Thus, the codification into event category sequences
possible error was 21,000 scores. The percentage of concordance allows an evaluation of the reliability of the proposed classification
between the main coder and the coworkers was 86% and 90%, system.
approximately, because the corresponding errors were 2,934 and
2,115 scores, respectively. The details of this first classification pattern
are given below. 2.5. Contributing factors
The second classification method, for contributing factors, follows
a similar reasoning way, establishing eight possible causes to be A similar process was followed for the contributing factors. The
considered, with the same intensity values: 100, 75, and 50 (see way to establish the error between two coders is similar to the
Table 3). The maximum possible error was 28,000 scores. In this case, precursor events codification. Here, the maximum possible error for
the percentage of concordance was approximately 85%, 87%, which each analyzed accident is 800 scores. Therefore, the maxim total error
correspond with differences (error) given by 4,210 and 3,638 scores, for the 212 analyzed accidents is 169,600 scores.
respectively. More details of this second classification method are Example: Table 3 shows the codification of the contributing factors
given below. corresponding to an accident which has been analyzed by the coders
In summary, a strategy to measure the degree of concordance C1 and C2.
among coders has been used. Since the level of concordance was high In our study, the percentage of agreement between both coders C1
for both cases, we find that the coding instruments for the precursors and C2 was 81% and the percentages of agreement of the coders C1 and
events and contributed factors provide a reliable method to codify the C2 with the main coder were 82% and 87%, respectively.
mining accidents. Finally, and taking both causal factors into account, Thus, the obtained reliability was high, both for coding precursor
the total number of accidents (212) was coded independently by the events and contributing factors. We can note that the coding
coders. The obtained results will be presented in the next section. instrument for the contributing factors provides a good method to
rank the possible factors as a cause of an accident. At the end of the
coding process, we can find the most common factor as a prime causal
2.4. Precursor events factor of an accident up to the minor important contributing factor.

Each category (E= Environmental, Eq = Equipment, M = Medical,


and B = Behavioral) was coded in terms of the certainty of evidence for 3. Results and discussion
its existence, associated to a percentage. Thus, 100% = yes, certainly;
75% = yes, probably; 50% = yes, possible. The pattern used to codify The analyzed 212 serious and fatal accidents only take into account
precursor events can be observed in Table 2, which shows a hypothetical two conditions: (a) it has to be an accident that was classified as
accident codified by two independent coders. The boxes numbered as 1, serious or fatal by the competent medical staff, and (b) the possibility
2, and 3 have been associated with 100, 75, and 50 scores, respectively. of having access to the accident report that was made by the General
The following rule was imposed: A coder had to mark, as a Management of Energy and Mining in its day.
maximum, one box for the precursor event PE1, one box for PE2, and The reports on serious and fatal accidents in the mining industry in
only one box for PE3. Consider, for instance, the precursor event PE1. Spain are not digitalized or available on the Internet. One has to
Now, if a coder ticks one category in the box 1 (“Environmental” in our request them from the competent faculty in order to examine the
example), it means that this category is “yes, certainly” in his/her information. This is not often possible. The 212 accidents that were
opinion, having the corresponding 100 scores. But if another coder, for analyzed took place between 1982 and 2006 in the Spanish
the same accident, ticks the same category for PE1 but in the box 2, he/ Autonomous Community of Catalonia. In this period of time, there
she places that accident in the category “yes, probably,” with 75 were 244 serious and fatal accidents in the mining sector in Catalonia
scores. Then, in this case, the error (in absolute value) between both (INSHT). This is why the sample that was studied made up 86.1% of
coders is 25 scores. An analogous procedure is applied to PE2 and PE3. the total accidents in this Spanish community.

Table 2

PE1 PE2 PE3

E Eq M B E Eq M B E Eq M B

1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3

C1 X X X
C2 X X X
Diff. 25 0 0 0 0 100 75 0 0 0 0 25 T = 225

Diff. = Difference (error) between C1 and C2 for a hypothetical accident.


T = Total error between C1 and C2 corresponding to a hypothetical accident.
4 L. Sanmiquel et al. / Journal of Safety Research 41 (2010) 1–7

Table 3

E Eq WP S T TE M O

1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3

C1 X X X
C2 X
Diff. 0 0 25 0 75 100 T = 200

Diff. = Difference (error) between C1 and C2 for a hypothetical accident.


T = Total error between C1 and C2 corresponding to a hypothetical accident.

Out of the 212 serious and fatal accidents that were analyzed, 139 behavior events have also occurred in 33.1% (n = 46) and 21.9%
took place during surface mining jobs, and 73 during underground (n = 16) for surface mining and underground mining, respectively.
mining jobs. The two main age groups of the workers who had Regarding the second event (PE2, which is the event that took place
accidents in the mining total were 30-39 (26.9%) and 45-54 (26.4%). immediately before the first event, PE1), we can observe that the
Moreover, 42.5% of the workers who had accidents had a labor opposite of the first event occurs in the two kinds of mining. That is to
experience of less than 4 years in the position. This would apply to say, the events that predominate as accident precursors are behavior
both surface and underground mining. Regarding the way in which related, which are more frequent than those of environment. Thus, we
the serious or fatal accidents were produced, 25.5% of them were can say that behavior events for surface mining, as a second event
caused by “being trapped by or between objects.” There were 19.3% (PE2), are present in 40.3% (n = 56) and 30.1% (n = 22) for surface
caused by “objects falling down or collapsing,” and 13.2% caused by mining and underground mining, respectively. However, the environ-
“people falling down at different levels.” For these two types of ment events are present only as a second event (PE2) in 5% (n = 7) and
mining, the results are the following: 5.5% (n = 4) of surface mining and underground mining respectively.
Regarding the third event (PE3), behavior events are predominant in
• For surface mining: 30.2% were caused by “being trapped by or surface mining as well as underground mining and are 7.2% (n= 10)
between objects,” 17.3% were caused by “people falling down at and 4.1% (n= 3), respectively.
different levels,” and 7.9% were caused by “being trapped by Equipment events only make up 12.9% (n = 18) and 11.0% (n = 8)
overturned machines or vehicles.” as a first event (PE1) of the surface and underground mining
• For underground mining: 48.0% were caused by “Objects falling accidents. As a second event (PE2), the percentage is 6.5% (n = 9)
down or collapsing,” 16.4% were caused by “being trapped by or and 1.4% (n = 1).
between objects,” and 13.7% were caused by “being run over or hit Out of the 212 accidents that were analyzed, 99.5% of them
by vehicles.” produced a first event (PE1), 46.7% produced a second event (PE2),
and only 9.0% of the accidents produced a third event (PE3). Thus,
Of the accidents, 18.4% took place within the first two hours of only 0.5% of the accidents were not identified with any precursor
work. In surface mining, the percentage rose to 21.6% of the cases; event. The behavior events occurred in 72.1% (n = 153) of the total
while in underground mining it was only 12.3%, and the 4th hour of mining accidents, the environmental ones in 63.2% (n = 134), and the
work was when most accidents occurred (17.8%). equipment ones in 18.4% (n = 39).
Six patterns of event sequences accounted for 74.8% (n = 104),
3.1. Patterns of accident causation 87.7% (n = 64), and 79.2% (n = 168) of the surface, underground, and
total mining accidents, respectively (see Fig. 2).
Fig. 1 shows how environmental events predominate in both Four of the six patterns involved an environmental event just prior
surface and underground mining in the first precursor event (PE1); to the serious or fatal accident. Behavior events appear in five
51.1% (n = 71) refer to surface mining and 67.1% (n = 49) to patterns, and only in one as a first event (PE1). They are also the
underground mining. We can observe that in order for immediate events that appear most often as a third event (PE3). The three
accidents to occur (PE1 first event), workplace conditions have sequence patterns that add up to the highest number of cases are for
influenced underground mining more than surface mining. Human the total of mining: (a) “E-o-o,” that is, accidents in which only one

Fig. 1. Types of events that preceded serious and fatal accidents in surface, underground and total mining (where PE1 indicates the first event that took place immediately before the
accident, PE2 is the event that took place before PE1, and PE3 is the event prior to PE2).
L. Sanmiquel et al. / Journal of Safety Research 41 (2010) 1–7 5

Fig. 2. Patterns of precursor events.

event appears (PE1), and also the environmental type, 24.1% (n = 51); predominant factor is that of the equipment, 21.7% (n = 46). The third
(b) “E-B-o,” accidents in which two preceding events occurred, a first most predominant contributing factor is that of equipment, 20.1%
event (PE1) that is the environmental kind, and a second event (PE2) (n = 28) for surface mining and earlier task error, 17.8% (n = 13) for
involving behavior, 23.1% (n = 49); and (c) “B-o-o,” accidents in underground mining. For the mining total the third predominant
which only one preceding event occurred (PE1), which was of the factor is that of training, 19.3% (n = 41).
behavior kind, 22.2% (n = 47). It has been observed that the contributing factors due to unsafe
All of the behavior events that were identified in the analyzed work practice are clearly the predominant factors in both surface and
accidents implied some kind of error (see Table 4). Behavior events in underground mining. They are also the main factors or causes in a
both surface and underground mining were largely action related higher proportion. These contributing factors, which are caused by
errors. These took place in the three events, with the exception of the unsafe work practice, can be classified in 11 types (Feyer & Williamson,
third event (PE3) in underground mining, where errors by omission 1997; see Fig. 3).
are more frequent than those caused by actions. If one uses the The predominant work practice is “Unsafe procedures by man-
alternative method 2 for the classification of behavior errors and agement,” which are 32% (n = 33) of the work practice factors
which is based on the level of behavior (Rasmussen, 1982), one will identified in surface mining, 52.5% (n = 21) of underground mining,
see that “skill-based” errors predominate in surface and underground and 37.8% (n = 54) of total mining. The second work practice that
mining, as well as in the three events (PE1, PE2, PE3; see Table 4). occurs most frequently is “Safety equip not provided,” 16.5% (n = 17)
These errors took place in 48.2% (n = 54), in 63.4% (n = 26), and in in surface mining, and “Equipment upkeep,” 20% (n = 8) for
52.3% (n = 80) of surface mining, underground mining, and total underground mining. For total mining, the second work practice
mining, respectively. The “rule based” errors, after the “skill-based” that is most often repeated is that of “Safety equip not provided” and
ones, are the kind of errors that take place most frequently in the first “Equipment upkeep,” 14% (n = 20) each one.
event (PE1) in the two types of mining. This is different for the second
event (PE2), where “knowledge based” errors are the second kind of
predominant behavior type error. In all the three events (PE1, PE2, 4. Conclusions
and PE3), “rule based” and “knowledge based” behavior errors
occurred in 51.8% (n = 58), in 36.6% (n = 15), and in 47.7% (n = 73) The analysis of the 212 serious and fatal accidents in Spanish
of surface, underground, and total mining, respectively. mining has allowed the identification of a series of causes that
Regarding the factors that contributed to the causes of the influenced in their occurrence. The classification method of accident
analyzed serious and fatal accidents, the factors that came about causes that was used (Feyer & Williamson, 1991) has exposed some
due to unsafe work practice (see Table 5) are clearly predominant. accident patterns or characteristics that are grouped into surface and
They took place in 74.1% (n = 103) of surface mining accidents, in underground mining and the total of both.
54.8% (n = 40) of underground mining, and in 67.5% (n = 143) of total The first event (PE1) was mostly produced by an environment
mining accidents. The second most predominant contributing factor is event in both surface and underground mining. Environment events
that of training; 23.7% (n = 33) for surface and equipment mining and are essentially due to deficiencies in the preventive system of
24.7% (n = 18) for underground mining. For total mining the second companies, since working conditions depend mainly on whether
companies have established adequate supervision and maintenance

Table 4
Clasification of behavioral events. Table 5
The role of different types of contributing factors.
% Cases % Cases % Cases
Surface Underground Mining Contributing Surface mining Underground mining Mining total
mining mining total Factor
% of % of cases as % of % of cases as % of % of cases as
PE1 PE2 PE3 PE1 PE2 PE3 PE1 PE2 PE3 cases prime cause cases prime cause cases prime cause

Cases with error 33.1 40.3 7.2 21.9 30.1 4.1 29.2 36.8 6.1 Environmental 16.5 39.1 16.4 50.0 16.5 42.9
Clasification method 1 Equipment 20.1 75.0 24.7 55.6 21.7 67.4
Omisión 23.9 12.5 10.0 12.5 18.2 66.7 21.0 14.1 23.1 Work practice 74.1 68.0 54.8 75.0 67.5 69.9
Comminssion 76.1 87.5 90.0 87.5 81.8 33.3 79.0 85.9 76.9 Supervisión 2.9 0.0 16.4 41.7 7.5 31.3
Clasification method 2 Training 23.7 30.3 11.0 25.0 19.3 29.3
Skill-based 52.2 44.6 50.0 68.8 54.5 100.0 56.5 47.4 61.5 Earlier task 19.4 51.9 17.8 61.5 18.9 55.0
Rule-based 32.6 21.4 30.0 18.8 18.2 0.0 29.0 20.5 23.1 error
Knowledge-based 15.2 33.9 20.0 12.5 27.3 0.0 14.5 32.1 15.4 Medical 2.2 100.0 0.0 0.0 1.4 100.0
Unknown 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Other 13.7 21.1 0.0 0.0 9.0 21.1
6 L. Sanmiquel et al. / Journal of Safety Research 41 (2010) 1–7

Fig. 3. Porcentage of cases with each type of unsafe work practice.

mechanisms. If these kinds of underground mining accidents are that 60.0% (n = 21) of these accidents took place while jobs like
analyzed, we see that an event preceding the first event (PE2) was drilling, scaling, or cutting with a continuous miner were being
produced in 30.6% (n=15). carried out near the site of the accident.
A series of contributing factors were identified in 75.5% (n = 37) One possible explanation is that these machines transmit vibra-
and were directly caused by the companies' preventive system, which tions to the roofs or walls of the tunnel while they are functioning and
malfunctioned where those accidents took place. this contributes to material falling down. In these cases, safety
These contributing factors are those of equipment, supervision, protocols and preventive measures should be extreme.
and training; as well as the contributing work practice factors “Unsafe In the case of surface mining, the 71 accidents where a first event
procedures by management,” “Equipment upkeep,” “PPE not provid- (PE1) of the environment type took place are classified by 19.7%
ed,” or “Safety equipment not provided.” (n = 14) that did not have adequate protection against falls at different
In the 35 accidents that occurred in underground mining (i.e., levels, 18.3% (n = 13) had limited space for the worker to carry out his/
48.0% of the accidents in which an environment first event (PE1) took her task correctly, and 18.3% (n = 13) had dangerous environmental
place), the accidents were also caused by “objects falling down or conditions (presence of dust, smoke, gases or projections).
collapsing.” Inadequate workplace conditions were the main cause of all of
The people who had the accidents were directly affected by these accidents. Of the 71 accidents, 59.2% (n = 42) mentioned had a
material falling from the roofs or walls of the tunnels where they behavior event that preceded the first kind (PE2), which is different to
happened to be. Of these accidents, 34.3% were fatal; the rest of them underground mining where only 30.6% of this kind of accident had a
seriously injured the workers involved. Of the 35 accidents, 65.7% preceding cause that was attributed to human behavior. As also occurs
(n = 23) indicated involved one of the indicated contributing factors with underground mining, accidents were also caused by contributing
that was directly caused by a malfunction in the company's preventive factors that can be attributed directly to malfunctioning of the
system. For example, several accidents of this kind have been noted company's preventive systems. These contributing factors appear in
where, if the worker had had access to work equipment or a vehicle 90.1% (n = 64) of the accidents where a first event (PE1) of the
with an adequate cover, he/she would not have been injured, or his/ environmental kind took place. These are equipment, supervision and
her injuries would not have been so serious. Another observation is training factors, as well as contributing factors of work practice such
L. Sanmiquel et al. / Journal of Safety Research 41 (2010) 1–7 7

as “Unsafe procedures by management,” “Equipment upkeep,” “PPE general archives, where we were able to look up the serious and fatal
not provided,” or “Safety equip not provided.” accident reports in Catalonian mining from 1982 to 2006. We would
Concerning the behavior events that occurred in the 212 accidents also like to thank the Instituto Nacional de Seguridad e Higiene en el
that were analyzed, these events are those that occurred most often if Trabajo Español, for allowing us to access their database of all the
we take into account three events (PE1, PE2, and PE3) of the sequence accidents that took place in Spain from 1999-2006.
of the accidents. This coincides with the common belief that the The authors would like to thank the anonymous reviewer for his/
causes of the accidents are mainly attributed to human error her helpful comments paper.
(Murphy, 1994). There have been a total of 153 behavior events in
the mining total; 112 of them occurred in surface mining and 41 in References
underground mining. As was mentioned in the previous paragraph,
Box, G. E. P., Hunter, J. S., & Hunter, W. G. (2005). Statistics for Experimenters: Design,
skill-based behavior errors predominated surface and underground Innovation, and Discovery, 2nd edition. New York: Wiley.
mining, as well as in the case of the three events (PE1, PE2, PE3). Cox, S. J., & Tait, N. R. S. (1991). Reliability, Safety and Risk Management: An Integrated
Behavior errors that are based on rules and knowledge can be avoided Approach. London: Buttenvorth-Heinemann.
Crichton, M. (2005). Attitudes to teamwork, leardership, and stress in oil industry
more easily with good prevention than those that are skill based drilling teams. Safety Science, 43(9), 679−696.
behavior errors (Williamsom & Feyer, 1998). This is partly due to skill- Feyer, A. -M., & Williamson, A. M. (1991). A classification system for causes of
based errors being the result of the worker's experience and training. occupational accidents for use in preventive strategies. Scandinavian Journal of
Work and Environmental Health, 17, 302−311.
This occurs when a worker who has had a lot of experience or is Feyer, A. -M., & Williamson, A. -M. (1997). The involvement of human behavior in
skilled at his/her task does things incorrectly. Self-confidence, routine, occupational accidents : Errors in context. Safety Science, 25(1–3), 55−66.
and so forth all contribute to the worker not paying attention to safety Murphy, J. N. (1994). Coal mine health and safety research in the USA-the achievements
of the US Bureau of Mines. Coal International, 242(6), 219−226.
aspects and this can cause accidents. There was a notably higher
Oncins de Frutos, M. (1991). NTP 283: Encuestas: metodología para su utilización.
proportion of skill based behavior errors in the 73 underground Madrid: Instituto Nacional de Seguridad e Higiene en el Trabajo.
mining accidents than in the case of surface mining (see Table 2); Piqué, T. (1991). NTP 274: Accidents investigation: Causes Tree Analysis. Madrid:
63.4% occurred specifically in underground mining as opposed to Instituto Nacional de Seguridad e Higiene en el Trabajo.
Rasmussen, J. (1982). Human errors: A taxonomy for describing human malfunction in
48.2% in surface mining. industrial installations. Journal of Occupational Accidents, 4, 311−335.
We have said that environment based events and some specific Rubio, M. (2007). De vital importancia. Técnica Industrial, N° 267, 17−21.
contributing factors are mainly caused by the malfunctioning of a Saari, J. (2001). La prevención de accidentes hoy en día. Magazine n° 4 de la Agencia
Europea para la seguridad y salud en el trabajo (pp. 3−5).
company's preventive systems. We have also mentioned that Santamaría, N., Catot, N., & Benavides, F. G. (2006). Time trends in fatal traumatic
behavior type errors that are based on rules and knowledge are occupational injuries in Spain (1992-2002). Gaceta Sanitaria, 20(4), 280−286.
more easily avoided by having a good preventive policy than in the Williamsom, A., & Feyer, A. -M. (1998). The Causes of Electrical Fatalities at Work.
Journal of Safety Research, 29(3), 187−196.
case of skill-based behavior type errors.
In all of the 212 accidents that were analyzed, 80 skill-based
Lluís Sanmiquel Pera: Holds a PhD from the Polytechnical University of Catalonia
behavior type events were identified as being involved in 73 different (UPC) and a Superior Technician Degree in Prevention of Labour Risks in the
accidents. There were also 343 contributing factors identified, of specialties of Security, Hygiene, Ergonomic and Psycho-Sociology. Between 1990 and
which 237 (69.1%) are in the category of factors that were directly 2004 he was a mining Inspector for the Main directorate of Mines and Energy of
Catalonia, (Spain). Since 1990, he is a professor of the department of Mining
caused by errors in the company's preventive systems. The remaining Engineering and Natural Resources of the Polytechnical University of Catalonia (UPC).
106 factors (close tasks errors, medical errors, and others), as well as He has collaborated in different training programs in subjects related to labor security
specific bad work practice (equipment misuse; PPE provided, not for both underground and opencast quarries and exploitation plants.
used/misused; PPE not used, reason unknown; safety equipment
provided, not used/misused; safety equipment not used, reason Modesto Freijó Alvarez: Holds a PhD from the Polytechnical University of Catalonia
(UPC) and a Superior Technician Degree in Prevention of Labour Risks in the
unknown; unsafe procedures by individual; unsafe procedures specialties of Security, Hygiene, Ergonomic and Psycho-Sociology. He has been a Senior
unknown origin), cannot be directly attributed to a bad functioning Professor at the Polytechnical University of Catalonia for the last 22 years. He has
of the preventive policy. The misuse or non-use of individual 9 years professional experience in the industry and has published several articles in
Spanish journals on labour security in mining. He has written communications,
protection equipment (PPE), the misuse of work equipment, and so attended and spoken at International mining Congresses.
forth by the worker could be caused by skill-based behavior.
In all the accidents that were analyzed there has been an event or Joaquín Edo Tomás: Mining Engineer since 1968. Senior Professor of the Polytechnical
contributing factor that can be clearly avoided with a rigorous and University of Catalonia (UPC) for the last 25 years. He has had positions and
efficient preventive policy. This means that the index of work related responsibilities in different mining operations for 21 years. He has published articles
on labour security in mining in Spanish journals. He has written communications,
fatalities in Spanish mining can be greatly reduced and brought to the attended and spoken at International mining Congresses.
level of countries like the United States. In order to do this, and based
on the types of causes and factors identified in the accidents that were Josep M. Rossell Garriga: Recieved his Ph.D. degree in Mathematics from the
analyzed, it is essential to carry out the following actions: improve- Polytechnical University of Catalonia (UPC), Spain, in 1998. He is now a professor in
the same University, where he teaches courses in the area of Mathematics. His
ment of the workers' training and information; adoption of a
research interests include decentralized control, robust control, and statistics methods.
maintenance system and enough preventive and adequate inspec- He leads an important research line of the CoDALab research group. He is the author
tions of work equipment, personal protection equipment (PPE), and and co-author of many papers indexed in ISI JCR and he has participated in prestigious
work areas; adequate supervision by supervisors and those respon- international conferences. Moreover, he has written books and publications on
academic mathematical topics. He has also participated in several National and
sible for jobs that involve serious risk; the election of work equipment International Research Projects.
that is better suited to the task; and the creation of viable and efficient
rules and procedures for safe work.

Acknowledgements

We would especially like to thank the Dirección General de Energía


y Minas de Cataluña (Spain) for allowing us to have access to its

You might also like