Journal of
J Occup Health 2007; 49: 482–492
Evaluation of the Control Banding Method—Comparison with
Measurement-based Comprehensive Risk Assessment
Haruo HASHIMOTO1, Toshiaki GOTO1, Nobutoyo NAKACHI1, Hidetaka SUZUKI1,
Toru TAKEBAYASHI2, Shigeyuki KAJIKI3 and Koji MORI3
1
Medicine and Occupational Health, ExxonMobil Yugen Kaisha, 2Department of Preventive Medicine and Public
Health, School of Medicine, Keio University and 3Occupational Health Training Center, University of Occupational
and Environmental Health, Japan
Abstract: Evaluation of the Control Banding
Method—Comparison with Measurement-based
Comprehensive Risk Assessment: Haruo
HASHIMOTO, et al. Medicine and Occupational Health,
ExxonMobil Yugen Kaisha—The control banding
method, or “control banding”, is a simplified risk
assessment system for chemical handling tasks. This
system is supposed to provide assessment results of
reasonable quality without expert involvement. The
objective of this study was to evaluate the
appropriateness of control banding judgment on the
basis of workplace safety. A common approach for
assessing workplace risk, which is called
“comprehensive risk assessment” in this study, is to
measure workers’ exposure and compare it with
relevant occupational exposure limits. Risk
assessment was performed with control banding
(COSHH Essentials, UK) at 12 workplaces of a
petroleum company in Japan, where health risks had
already been assessed separately through
comprehensive risk assessment by experts and control
technologies implemented accordingly. The obtained
control banding judgments were then examined with
regard to their adequacy by comparing them with
existing control technologies. There was majority of
cases (seven) where judgments by control banding
were identified as “over-controlled”; there was no
judgments identified as “under-controlled”. Control
banding also requested the seeking of expert advice
in the majority of cases (eight). Thus, it was
demonstrated that control banding tends to provide
safe-sided judgment. A possible interpretation of this
is that control banding is inherently designed to secure
Received Jun 28, 2007; Accepted Aug 13, 2007
Correspondence to: H. Hashimoto, Medicine and Occupational
Health, ExxonMobil Yugen Kaisha, 1–8–15 Kohnan, Minato-Ku,
Tokyo 108-8005, Japan
(e-mail: haruo.hashimoto@exxonmobil.com)
Occupational Health
workplace safety by compensating for its insufficient
exposure information with safe-sided judgment criteria
and by requiring experts’ intervention in high-risk cases.
Control banding could be widely and effectively utilized
in Japan, especially by employers in small enterprises,
provided that the above characteristics are pre-
acknowledged and health experts are made available.
To this aim, it is essential to develop new local health
experts and establish institutional mechanisms for
facilitating employers’ access to expert advice. It should
however be noted that the number of workplaces
evaluated in this study was small.
(J Occup Health 2007; 49: 482–492)
Key words: Control banding, COSHH Essentials, Risk
assessment, Exposure, Chemicals, Hazard, Control,
Small enterprises, Expert
It has been reported that the number of chemicals
currently used in Japan is more than fifty thousand. This
makes the appropriate control of the health risks of these
chemicals a current occupational health issue1). While
occupational health control of chemicals in Japan has been
performed according to the Industrial Safety and Health
Law, which defines specific requirements for workplaces,
employers are supposed to autonomously perform risk
assessment and management of chemicals that are not
specified in regulatory clauses, since the number of
chemicals regulated by the law is limited. With this
background, the Ministry of Health, Labour and Welfare,
Japan, promulgated the “Guideline for risk assessment
of chemicals and other hazards” in March 2006, which
recommends employers to perform risk assessments for
all chemicals used by them and to take risk mitigation
measures as needed2).
In order to perform an accurate risk assessment of a
workplace, collection of detailed hazard information and
measurements of workers’exposure are commonly
 Haruo HASHIMOTO, et al.: Evaluation of the Control Banding Method
required, and a health expert who has sufficient
knowledge of chemicals management is usually needed
to perform them. However, there are often cases where
an expert is not readily available, especially among small
and medium enterprises. Thus, a simple risk assessment
method is required which does not require expert
involvement.
The control banding method, or “control banding”, is
a risk assessment system for chemical handling tasks
which was originally developed by the HSE (Health and
Safety Executive, UK) and is now being globally
promoted by the ILO (International Labour Organization).
Control banding recommends necessary control
technologies based on only three characteristics of a
chemical handled: these are “hazard of the chemical”,
“scale of use”, and “ability to become airborne”3). Since
control banding is designed to follow a standardized
process without direct exposure monitoring, this method
is supposed to provide assessment results with reasonable
quality under circumstances in which an expert is not
available. Several published studies have investigated
the appropriateness of risk assessment judgments of
control banding. These studies have attempted to validate
the two exposure models on which the risk decision
scheme of control banding is grounded3–5). It was partially
demonstrated that control banding tends to provide
appropriate or safe-sided (over-controlled) judgment in
general. However, the results of these studies are
insufficient to draw an overall conclusion regarding the
appropriateness of control banding for workplace safety,
because the actual exposures of workplaces controlled in
accordance with control banding judgment were not
directly compared with the relevant occupational
exposure limits.
Occupational exposure limits have been common
judgment criteria for workplace safety. In order to
accomplish accurate risk assessment of chemical handling
tasks, it is usually considered as a standard approach to
evaluate the exposure level and to compare it with
occupational exposure limit of the chemical; this approach
will hereafter be called “comprehensive risk assessment”
in this study. In this study, risk assessment was performed
with control banding at workplaces where health risk had
already been assessed separately by experienced experts
through comprehensive exposure assessment and control
technologies had been implemented, as needed,
accordingly. Then, the appropriateness of control banding
was examined by comparing the obtained control banding
judgment with the already existing controls.
Methods
Control banding
The risk levels of 12 chemical handling tasks, which
are performed in refineries, a petrochemical plant, oil
terminals and a research laboratory of a petroleum
483
company in Japan, were assessed with control banding
(COSHH Essentials) available on the HSE website6). The
details of each task, its process, duration, the chemical
handled and its amount, are shown in Table 1.
Characteristics of each task relevant to control banding
assessment, such as the chemical species assessed, its
Risk-phrases (R-phrases) and hazard group, scale of use,
and ability to become airborne, are shown in Table 2. R-
phrases corresponding to a chemical were identified
through the chemicals database carried on the homepage
of the European Chemicals Bureau7). When the handled
chemical was a mixture of multiple chemical species, R-
phrases assigned to that mixture (example: gasoline) were
used for assessment, and R-phrases assigned to
components of the mixture (example: benzene) were not
used. This is because it was assumed that it would not
necessarily be easy for small and medium enterprises,
the chief target employers of control banding, to obtain
specific information regarding components of a chemical
mixture. However, the chemical species assessed for task
# 4 was selected as benzene, and not petroleum products,
since it was well known by workers that the petroleum
products handled in this task contained appreciable
amounts of benzene (1 to 5%). For respective tasks, the
set of above information was input into control banding
online interface, and then one of the four risk levels—
the “control approach criteria” which range from level 1
(lowest risk) to level 4 (highest risk)—was judged and
presented (Table 2).
Comprehensive risk assessment
The following risk assessments and managements were
performed by workplace health experts (a Certified
Industrial Hygienist of U.S.A. or equivalent) with more
than several years of experience. A series of assessment
steps was used for this risk assessment. First, personal
exposure monitoring was performed regarding the most
hazardous or most representative component of the
chemical handled. The monitored chemical species were
benzene, lubricating oil (monitored as total hydrocarbons)
and n-heptane (Table 1). The exposure monitoring was
carried out with passive samplers (Traceair OVM-1 from
K&M Environmental Co.) which were then analyzed at
the ExxonMobil Biomedical Sciences Inc. (New Jersey,
U.S.A.) which is an AIHA (American Industrial Hygiene
Association) accredited industrial hygiene laboratory8).
Second, a set of “exposure rating” was defined, in order
to classify the degree of exposure for workers. This is
shown in Table 3: there are five levels of exposure rating
from A (highest exposure) to E (lowest exposure) which
are based on the relative relationship between the
magnitude of exposure and the corresponding
occupational exposure limit of the chemical monitored.
For a task with a duration of more than 15 min (i.e. with
monitoring time more than 15 min), the 8-h time-weighted
 484
Table 1. Chemical handling tasks and their risk assessment results by comprehensive exposure assessment

# Task Description Task Chemical Quantity Monitoring RV *2 Exp. HER Risk Control Note
Duration Handled Handled Chemical n Av. TWA (ppm) rating *3 Level
(min) (l) (ppm) (ppm) *1 Requested? Implemented

1 Disconnecting and cleaning a 15 Gasoline 1 Benzene 6 0.18 0.18 2.5 D I 2 N –

strainer (height=30 cm) attached to *4
gasoline transfer piping
2 Filling naphtha into 200 l drums 10 Naphtha 1,000 Benzene 4 2.08 1.37 2.5 B I 1 Y RPE-1 *6
while holding the charging nozzle *5
3 Shoveling and cleaning residual oily 480 Oily 1,000 Benzene 6 6.84 6.84 0.5 A I 1 Y RPE-2 *6
sludge inside a huge crude oil tank Sludge
4 Testing benzene-containing 480 Oil with 1 (Benzene) – – – 0.5 A I 1 Y Ventilation *7, *8
petroleum products in laboratory benzene hood
5 Washing used glass sample bottles 480 Petroleum 0.75 (Toluene) – – – 50 A III 1 Y Ventilation *7, *8
of petroleum products in laboratory Products hood
6 Drawing off hot motor oil from a 90 Lubricating 4 Total 2 <0.35 <0.07 100 E II 3 N – *8
test automobile engine oil *9 hydrocarbon *9
7 Washing disassembled engine 215 Petroleum 1 n-Heptane 2 6.60 2.96 400 E III 3 N – *8
parts, with solvent solvent
8 Refilling gasoline into a test 45 Gasoline 0.01 Benzene 3 0.16 0.02 0.5 D I 2 N – *8
automobile engine, with small
amount of leakage
9 Rinsing laboratory glassware with 10 Acetone 0.05 (Acetone) – – – 750 C III 2 N – *7, *8
acetone
10 Rinsing laboratory glassware with 50 Toluene 0.5 (Toluene) – – – 50 A III 1 Y Ventilation *7, *8
toluene hood
11 Loading a tank lorry with gasoline 15 Gasoline 8,000 Benzene 33 0.40 0.40 2.5 C I 2 N –
product, while handling the
charging nozzle on the lorry roof
12 Sampling small amount of naphtha 3 Naphtha 2 Benzene 2 <0.00 <0.00 2.5 D I 2 N –
from a faucet attached to process
piping

*1: Time weighted average (TWA) exposure: 15 min TWA for task duration ≤15 min, 8 h TWA for task duration >15 min, *2: Reference value: ACGIH-TLV-STEL for task duration ≤15 min, ACGIH-
TLV-TWA for task duration >15 min, *3: Health effect rating, *4: Major gasoline ingredients: benzene(0.5%), toluene(6%), n-hexane(5%), *5: Major naphtha ingredients: benzene(1.2%), toluene(1.5%),
n-hexane(12%), *6: Respiratory protective equipment: RPE-1=half face mask, RPE-2=full face mask, *7: Exposure rating judged by observation, *8: General exhaust ventilation originally equipped, *9:
Monitored as total hydrocarbons; the RV (100 ppm) is a company standard of ExxonMobil Co.
J Occup Health, Vol. 49, 2007
 Haruo HASHIMOTO, et al.: Evaluation of the Control Banding Method 485

Table 2. Risk assessment results by control banding

Task Chemical evaluated Scale of use Ability to Become Control Approach

# Name R-phrase Assigned Hazard Group *1 *2 Airborne *3 Criterion *4

1 Gasoline 45, 65 E M M 4
2 Naphtha 45, 65 E L M 4
3 Crude Oil 45 E L M 4
4 Benzene 45, 46, 36/38, 48/23/24/25, 65 E M M 4, S
5 Toluene 38, 48/20, 63, 65, 67 D S M 3, S
6 Lubricating oils 45 E M L 4
7 n-Heptane 38, 50/53, 65, 67 A M M 1, S
8 Gasoline 45, 65 E S M 4
9 Acetone 36, 66, 67 A S M 1, S
10 Toluene 38, 48/20, 63, 65, 67 D S M 3, S
11 Gasoline 45, 65 E L M 4
12 Naphtha 45, 65 E M M 4

*1: A=Least hazardous, E=Most hazardous, *2: S=Small, M=Medium, L=Large, *3: L=Low, M=Medium, H=High, *4:
Corresponding control recommendation: 1=General ventilation, 2=Local exhaust ventilation, 3=Enclosure, 4=Seek expert advice,
S=Skin and eye precautions.

Table 3. Definition of “exposure rating”
Exposure rating Definition
A >RV*
B 50–100 % of RV
C 10–50 % of RV
D 1–10% of RV
E No exposure
*Reference value: ACGIH-TLV-TWA for tasks >15 min,
or ACGIH-TLV-STEL for tasks ≤15 min.
average (TWA) exposure was calculated from the
analyzed airborne concentration; the average 8-h TWA
was then obtained by averaging the relevant sample data.
Then, the exposure rating of this task was assigned based
on that average TWA. For a task with a duration of equal
to or less than 15 min, the average 15-min TWA was
calculated similarly, and the exposure rating of the task
was then assigned.
The exposure rating of tasks #4, 5, 9 and 10 were
assigned based on qualitative judgment: the experts
carefully observed the respective tasks and estimated
exposure levels with knowledge of past exposure
monitoring results for similar tasks within the worksite.
This type of qualitative judgment has been acknowledged
as an effective screening method before quantitative
measurements in the Occupational Exposure Sampling
Strategy Manual by NIOSH9). Specifically, the exposure
levels were estimated to be relatively high for tasks #4, 5
and 10, since the amount of chemical handled was
relatively large, more than 0.5 l. On the other hand, the
exposure level was estimated to be low for task #9, since
the amount of chemical handled was low, 0.05 l, and
ACGIH-TLV-STEL of acetone is exceptionally high, 750
ppm. Consequently, the exposure rating of the 12 tasks
was respectively assigned to one of the levels from A to
E, as shown in Table 1.
Third, according to the scheme utilized by Booher et
al.10), hazard levels of handled chemicals were classified,
based on the EU labeling classification 11), into four
levels—from I (most hazardous) to IV (least hazardous)—
hereafter called “health effect rating” (HER). The
labeling classification of each chemical was assigned in
accordance with R-phrases of the chemical. Specifically,
benzene was assigned to HER I since it is a “category 1
carcinogen”; lubricating oil was assigned to HER II since
it is a “category 2 carcinogen”; and n-heptane, toluene
and acetone were assigned to HER III since they are
“irritant”(Table 1).
Finally, a matrix was constructed as shown in Table 4.
This matrix defines three risk levels from 1 (highest) to 3
(lowest) for a chemical handling task, based on a
combination of the exposure rating and the health effect
rating of the task. The risk levels of the 12 tasks were
individually determined by adopting the exposure rating
and the health effect rating of each task in this matrix;
the result is shown in Table 1. For tasks with a determined
risk level of 1 (highest), a new exposure control
technology, such as engineering or administrative control,
was executed. The actual design or content of the control
was determined by the experts. For tasks with risk levels
of 2 or 3, no exposure control technology was newly
implemented.
 486 J Occup Health, Vol. 49, 2007

Table 4. Matrix for “Risk level*” assignment in comprehensive exposure assessment

HER** Exposure rating

A B C D E

I 1 1 2 2 3
II 1 1 2 3 3
III 1 2 2 3 3
IV 2 2 3 3 3

*Risk level: 1=Highest risk, 3=Lowest risk, **Health effect rating: I=Most hazardous,
IV=Least hazardous

Table 5. Comparison scheme of risk assessment results

Control banding: Comprehensive exp. “Appropriateness” of judgment

Control approach criteriona assessment: Risk levelb by control banding

1, 2, or 3 1 Over-controlled c
Appropriate c
Under-controlled c
1, 2, or 3 2 or 3 Over-controlled
4 1 Not comparable
4 2 or 3 Over-controlled
a1=General ventilation, 2=Local exhaust ventilation, 3=Enclosure, 4=Seek expert advice,
b1=Control requested, 2=Control not requested, 3=Control not requested,
cThe actual control executed through comprehensive exposure assessment was compared with the control

recommended by the control approach criterion.

The comprehensive risk assessment utilized here,
which represents a series of risk assessment steps
described above, is a way of quantifying magnitude of
risk from a two-dimensional matrix. These two
dimensions are the hazard and exposure levels of a
chemical, the latter being determined by the relative
relationship between the observed exposure and the
occupational exposure limit of the chemical. These two
dimensions respectively represent “magnitude of a
hazard” and “probability of occurrence of that hazard”
which are key components of “risk” by definition.
Comparison of control banding and the comprehensive
risk assessment
Let us make an assumption that the comprehensive
risk assessment method utilized in this study is capable
of assessing magnitude of risk with reasonable accuracy
and that the control measures actually implemented were
just enough in regard to their scale and quality; the basis
of this assumption is that this assessment method is based
on direct or indirect exposure measurements, and that
experts are involved throughout the process. The
appropriateness of the control banding judgment was
examined for 12 tasks, on the basis that risk judgment by
the comprehensive risk assessment is a standard. The
results were classified into four criteria—“over-
controlled”, “appropriate”, “under-controlled” and “not
comparable”, as shown in Table 5. Namely, the control
approach criterion given by control banding was
compared with the judgment of the comprehensive risk
assessment whether new control technology was
requested or not (i.e. whether the judged risk level was
“1” or “2 or 3”). In cases where the control approach
criterion was level 1, 2 or 3, and also control technology
was requested by the comprehensive risk assessment (i.e.
judged risk level was 1), the actual control ececuted was
compared with the specific type of the control
recommended by the control approach criterion.
Results
Risk assessment results by means of control banding
are shown in Table 2. The judged control approach
criteria of 12 tasks ranged from 1 to 4. Each control
approach criterion, from 1 to 4, represents a specific
control recommendation. These are “general ventilation”,
“local exhaust ventilation”, “enclosure”, and “seeking
expert advice”, respectively. Five tasks which involved
chemicals hazardous to the skin and eyes were
 Haruo HASHIMOTO, et al.: Evaluation of the Control Banding Method 487

Table 6. Repeat risk assessment by comprehensive exposure assessment after implementation of control technologies

# Monitoring TWA *1 RV *2 Exp. HER Risk Note

Chemical n Av. TWA Assumed (ppm) rating *3 level
(ppm) (ppm)*1 (ppm)

2 (Benzene) – – – <0.14 2.5 D I 2 Assumption based on APF (*4)

of a half face mask (10)
3 (Benzene) – – – <0.14 0.5 C I 2 Assumption based on APF (*4)
of a full face mask (50)
4 Benzene 9 0.07 0.07 – 0.5 C I 2
5 Toluene 2 6.10 6.10 – 50 C III 2
10 (Toluene) – – – – 50 D III 3 Judged by observation

*1: 8 h time weighted average exposure, *2: Reference value: ACGIH-TLV-STEL for task duration >15 min, ACGIH-TLV-TWA for task
duration >15 min, *3: Health effect rating, *4: Assigned protection factor (NIOSH, yielding 1/10 or 1/50 decrease in exposure for a half or
full face mask respectively.)

Table 7. Comparison of risk assessment results and appropriateness of control banding judgment

Task # Control banding: Comprehensive exp. assessment:

Control approach Risk level Appropriateness of control
criterion Risk level Control done? banding judgment

1 4 2 N Over-Controlled
2 4 1 Y Not Comparable
3 4 1 Y Not Comparable
4 4 1 Y Not Comparable
5 3 1 Y, Ventilation hood Over-Controlled
6 4 3 N Over-Controlled
7 1 3 N* Appropriate
8 4 2 N Over-Controlled
9 1 2 N* Appropriate
10 3 1 Y, Ventilation hood Over-Controlled
11 4 2 N Over-Controlled
12 4 2 N Over-Controlled

*General exhaust ventilation was originally equipped.

additionally classified “S” criterion. The control approach
criterion having the largest number of tasks (eight tasks)
is criterion 4. This represents the highest risk level and
requests expert advice.
Risk assessment results by means of comprehensive
risk assessment are shown in Table 1. The exposure rating
and the health effect rating of the 12 tasks ranged from A
to E, and I to III respectively. The risk levels determined
from the combination of these two dimensions ranged
from 1 to 3. Among them, there were five tasks with risk
level 1, to which engineering or administrative controls
were then executed: local exhaust ventilation (laboratory
enclosing hoods) was installed for tasks #4, 5 and 10,
and respiratory protective equipment was provided for
tasks #2 and 3.
In order to evaluate the effectiveness of the controls
executed, a repeat risk assessment was performed for the
tasks after implementation of the controls; the results are
shown in Table 6. For tasks #2 and 3, exposure levels at
breathing zones were estimated based on an assumption
that a half-face or full-face air-purifying mask yields a 1/
10 or 1/50 decrease in exposure, respectively, compared
with the original exposures previously measured (shown
in Table 1), since the assigned protection factors defined
by NIOSH for half-face and full-face masks are 10 and
50, respectively 12). For tasks #4 and 5, exposure
monitoring was carried out. For task #10, the expert
carefully observed the task and qualitatively assessed the
exposure level. Consequently, the risk levels of these
tasks went down to either level 2 or 3 after implementation
 488
of controls: effectiveness of the controls was thus verified.
The comparison of the results of the risk assessments
performed with control banding and those with the
comprehensive risk assessment are shown in Table 7. For
tasks #2, 3 and 4, while control banding assigned them
to control approach criterion 4 (requesting expert advice),
the comprehensive risk assessment determined their risk
as level 1 and controls were executed, accordingly.
Control banding judgment was classified as “not
comparable” for these tasks, since it was impossible to
forecast the actual control advice to be provided by the
expert requested by control banding. For tasks #1, 6, 8,
11 and 12, while control banding assigned them to control
approach criterion 4 (requesting expert advice), the
comprehensive risk assessment determined their risks as
levels 2 or 3, and therefore did not request controls.
Control banding judgment was classified as “over-
controlled” for these tasks, since intervention of an expert
was deemed to be unnecessary. For tasks #5 and 10,
control banding assigned them to control approach
criterion 3, and requested “enclosure” of workplaces.
Local exhaust ventilation had actually been installed at
these workplaces according to the comprehensive risk
assessment results, and it was later confirmed that
exposure risk had been successfully mitigated. Thus,
control banding judgment was classified as “over-
controlled” for these tasks, since “enclosure” was not
necessarily needed. For tasks #7 and 9, control banding
assigned them to control approach criterion 1, and
requested general ventilation. The comprehensive risk
assessment had not requested control technologies.
However, both workplaces were already equipped with
general ventilation systems for these tasks, and the
comprehensive risk assessment was performed with
general ventilation functioning; tasks #7 and 9 were
carried out in a research and testing laboratory room,
respectively. Thus, control banding judgment was
classified as “appropriate” for these tasks, since the
exposure risk was appropriately controlled with general
ventilation. In summary, judgments by control banding
were “over-controlled” for the majority of tasks, seven;
three tasks were “not comparable”, and two were
“appropriate”; no task was “under-controlled”.
Discussion
Characteristics of control banding
The following characteristics of control banding can
be pointed out based on the results obtained.
(a) There is an enhanced tendency to provide safe-sided
judgment.
(b)There is an enhanced tendency to recommend seeking
expert advice; these cases are represented by control
approach criterion 4.
The following characteristics were also identified.
(c) It is valuable in practice that the need for skin/eye
J Occup Health, Vol. 49, 2007
protection is explicitly recommended for skin/eye
hazardous chemicals without omission.
(d)It should be cautioned that the risk assessment outcome
can be different depending on what species of
chemicals is selected as assessment objects for a
chemical mixture. For example, the risk assessment
of task #1 was performed for gasoline, a common
name of the mixture. If it were performed for toluene,
a gasoline component with typical concentration of
several percent, a risk phrase of R-38 would be
additionally identified (in the same way as task #5 in
Table 2) and control approach criteria “S” would be
supplemented accordingly. Control banding requires
that all components of a chemical mixture should be
evaluated and the highest (most demanding) control
approach criterion among all criteria presented should
be followed. However, such a comprehensive
approach is often not realizable, a typical obstacle
being that there are MSDS’s with insufficient
component information. There is also an issue of cut-
off criterion of mixture components, the percent
composition above which a chemical species is defined
as a component. For example, the handling of a
chemical species with a concentration in the range of
0.1% to 1.0% in a mixture may be controversial.
While this issue can also be problematic for
comprehensive risk assessment, extra caution should
be taken in case of control banding assessments when
experts are not involed.
(e) The risk of a task can not be re-assessed after the
implementation of control technology since facility
information is not incorporated into control banding
assessment steps. In contrast, re-assessment can be
readily performed by means of the comprehensive risk
assessment.
(f) The COSHH on-line interface is very user-friendly.
Comparison of this study with preceding studies
Control banding judges control approach criteria based
on three factors: “hazard of a chemical”, “scale of use”,
and “ability to become airborne”. It can be interpreted
that the judgment scheme of control banding has its
grounds on two exposure models3). The first model is:
“the exposure level is uniformly predicted, given ‘scale
of use’, ‘ability to become airborne’, and ‘control
approach criterion’”. Here, the exposure level is called
an “exposure band”, and it represents a range of exposure
and not a single value. The second model is: “the safe
exposure limit of a chemical is defined in correspondence
to the respective hazard group (‘A’ to ‘D’) to which the
chemical belongs, and is not directly related to each
chemical species”. This exposure limit represents a range
of exposure and not a single value, and it actually
coincides with the “exposure band”. When performing
risk assessment regarding a chemical handling task, one
 Haruo HASHIMOTO, et al.: Evaluation of the Control Banding Method
first identifies the hazard group to which the chemical
belongs. This then provides the target safe exposure limit
(i.e. “exposure band”) of the chemical. With this
“exposure band”, along with “scale of use” and “ability
to become airborne” of the chemical, one can determine
the “control approach criterion” through back-tracking
the logic of the first exposure model; this is the risk
judgment scheme of control banding. Chemicals in
hazard group “E” do not have a corresponding exposure
band, and they are always assigned to control approach
criterion 4.
Tischer et al. researched existing databases of 18
different industry operations in Germany regarding their
work conditions and monitored exposure levels3). They
identified the exposure band of each chemical handling
task based on the first exposure model of control banding,
and then compared the obtained exposure band with the
measured exposure. They found that actual exposure
levels were equivalent to or lower than corresponding
exposure bands for most of the tasks, which suggested
that the first exposure model was tuned to an appropriate
level or a rather safe-sided level.
Jones et al. compared exposure bands to measured
exposures with regard to solvent vapor degreasing and
powder bag filling operations, utilizing monitoring data
accumulated by the National Institute for Occupational
Safety and Health (NIOSH)4). Their study also represents
a validation attempt for the first exposure model of control
banding. They defined an “over-controlled” error as an
instance where conditions of chemical use prompted
control banding to recommend newly implementing a
control technology, although the monitored airborne
concentration was within or below the chemical’s
exposure band in the absence of any control technology.
This error was observed for as much as 61% of relevant
vapor degreasing operations. They also defined an
“under-controlled” error in which the monitored airborne
concentration exceeded the upper limit of the chemical’s
exposure band although there was a control technology
implemented according to control banding
recommendation. This error was observed for as much
as 78% of the vapor degreasing operations, and for as
much as 48% of the bag filling operations. Based on
these results, Jones et al. concluded that their analysis
did not support the view that control banding would be
able to accurately identify appropriate control
technologies and that the recommended control
technologies were capable of adequately controlling
exposures. This is a different conclusion from that of
Tischer et al4). Caution should be noted for the “under-
controlled” errors in Jones’ study, since the monitored
exposure levels had not been compared with the
occupational exposure limits of the chemicals concerned.
Therefore, an “under-controlled” error does not
necessarily suggest a workplace condition deemed to be
489
unsafe for workers.
Brooke compared the occupational exposure limits of
the UK with exposure bands for 111 chemicals. His study
represents a validation attempt for the second exposure
model of control banding, and it identified only 2 cases
out of 111 chemicals where the occupational exposure
limit was lower than the exposure band5). This means
that the exposure bands have been designed to be more
conservative than the occupational exposure limits for
most of chemicals. Thus, it was demonstrated that the
judgment scheme of control banding is tuned to provide
safe-sided decisions.
The studies referred to above attempted to validate the
appropriateness of the two exposure models of control
banding. It was partially demonstrated that control
banding tends to provide appropriate or safe-sided (over-
controlled) judgment in general. However, these results
are insufficient to draw overall conclusions regarding the
appropriateness of control banding from the point of view
of workplace safety. This is because the actual workplace
exposures and occupational exposure limits were
compared only indirectly, and not directly, within the
studies. Here, let us suppose a chemical handling task
which is controlled under a control approach criterion
provided by control banding judgment. When we follow
the general conclusion reached by Tischer and Brooke,
we can assume that the exposure of this task is either
equal to or lower than the exposure band, and that this
exposure band is lower than the occupational exposure
limit of the chemical. This results in a situation where
the exposure is lower than the occupational exposure
limit, and hence, we can conclude that the workplace is
surely safe and that control banding judgment is safe-
sided. On the other hand, Jones et al. identified a
significant number of “under-controlled” cases where
workplace exposures exceeded the relevant exposure
bands. For these cases, even after taking account of
Brooke’s general conclusion, it is difficult to predict
whether or not workplace exposure is lower than the
exposure band in general. Thus, it can be summarized
that verifying the appropriateness of respective exposure
models has limited effectiveness in evaluating control
banding system.
The most appropriate approach would therefore be to
compare the measured exposure of a chemical, used in a
workplace where the risk assessment had been performed
with control banding and the control technology advised
by it implemented, with the occupational exposure limit
of the chemical, in order to verify control banding system
on the basis of actual workplace safety. However, there
is a concern in terms of workers’ safety if we first assess
a workplace with control banding, then implement control
technology accordingly, and lastly monitor exposure of
the workplace to examine its environment (i.e. a
prospective study). In our study, risk assessment was
 490
performed with control banding at workplaces where the
risk had already been assessed by means of a
comprehensive exposure assessment (that is, exposures
were monitored and compared with occupational
exposure limits) and control technologies had been
implemented as needed according to experts’ professional
judgment. Then, the appropriateness of control banding
was evaluated by comparing the obtained control banding
judgment with existing control technologies (i.e. a
retrospective study).
The findings of this study demonstrate that control
banding tends to make safe-sided judgments. Namely,
control banding requested additional exposure protection
measures for workplaces where the actual exposures were
lower than the relevant occupational exposure limits.
Such cases were observed in more than half of the tasks
evaluated. Thus, this study confirms that control banding
judgment is safe-sided in reality on the basis of workplace
safety.
It was also observed in this study that there were a
number of cases where control banding requested expert
consultation (i.e. control approach criterion 4). This is
because many tasks involved petroleum products such
as gasoline or naphtha which contain benzene
(carcinogen, R45). In the Jones’ paper, “control approach
criterion 4” was assigned to 13 chemicals among 26
chemical substances (liquid or powder) evaluated4); the
frequency of “control approach criterion 4” was not
specifically discussed in the other studies referred to
earlier. In addition, all chemicals assigned to hazard group
E and some of the chemicals assigned to hazard groups
C and D are judged as “control approach criterion 4”
within the decision scheme of control banding6). On the
whole, we can predict that there will be, in general,
moderately frequent occurrence of control banding
judgment requesting expert consultation.
Position and limitation of this study
It must be noted that the scope of tasks evaluated in
this study was limited. The number of tasks was 12, and
the task processes were mainly liquid transfer and
washing within one petroleum company. More than half
of the chemicals handled in these tasks were benzene-
containing petroleum materials, such as gasoline or
naphtha. In addition, most of the chemicals evaluated
are classified as “medium” for their “ability to become
airborne”, and the number of exposure measurement data
utilized was rather small—about 70.
Tischer et al. mention in their study that they were
able to validate the appropriateness of COSHH Essentials
scheme only partially, despite the fact that they evaluated
as many as 18 kinds of tasks within a wide variety of
industries such as textile, printing, chemical and furniture
manufacturing, and that the number of exposure data
utilized was substantial—about 9603). They attributed
J Occup Health, Vol. 49, 2007
the reason for the imperfection of their validation as being
due to the fact that while all combinations of task
conditions amounted to a total of 54 exposure scenarios
in control banding, their study covered only 8 scenarios
and further tended to concentrate on “medium scale” and
“medium volatility/dustiness” conditions. Jones’s study
contains a large database as well. The numbers of
workplaces evaluated were 33 for vapor degreasing and
22 for bag filling operations, and the numbers of chemical
species measured were 7 and 19, respectively, while the
number of exposure measurements was about 710 4).
The above two studies suggest that substantially large-
scale investigation will be necessary to evaluate control
banding system with accuracy and objectiveness. Thus,
since the scope of tasks and chemicals evaluated in our
study was limited, it is appropriate to acknowledge this
work as a pilot study which compared control banding
with actual workplace management performed by means
of comprehensive risk assessment, and it should be
cautioned that the generalizability of this study will be
limited.
Utilization of control banding in Japan
It is desirable that health risk assessment of chemicals
be performed by experts having an appropriate level of
expert knowledge, since health risk is not often readily
perceivable by workers and specialized knowledge such
as toxicology is usually required for health risk
assessment and management. However, the availability
of experts is significantly different in practice from one
enterprise to another. Therefore, an employer will
implement control technology complying with an expert’s
risk assessment result, as long as an expert is available
and the quality of the collected hazard information is
good. In contrast, an employer will rather carry out safe-
sided (over-controlled) control technology if an expert is
unavailable or the quality of the hazard information is
poor, considering uncertainty factors within the risk
assessment system employed13). An example of such
choices between the two is the comprehensive risk
assessment and control banding investigated in this study.
These two methods can represent, in short, choices of a
“tailored method with accuracy” or a “safe-sided method
compensating for scarcity of information”. It can be
interpreted that control banding is inherently designed to
compensate for insufficient exposure information by a
safe-sided judgment and to secure the safety of high-risk
workplaces by requesting expert intervention.
Control banding could be effectively utilized for risk
assessment of chemicals in Japan, provided that its safe-
sided characteristics are pre-acknowledged by users and
that channels for expert intervention are secured. This
system could also be utilized for screening purposes
before performing more accurate risk assessments such
as the comprehensive risk assessment. In such cases, the
 Haruo HASHIMOTO, et al.: Evaluation of the Control Banding Method
degree of safety of a workplace would be expected to be
reasonably high if control banding had judged the present
status of the workplace as “safe”.
Money et al., as members of the COSHH Essentials
working group of the HSE, reported that control banding
is designed as inherently conservative and that it needs
to be operated in a “screening mode” which presupposes
seeking access to experts as needed14). They admitted
that there have been criticisms regarding artifacts of
control banding methodology and cases where control
banding “advice” is overly protective for some tasks when
compared with recommendations derived from empirical
occupational hygiene approaches. However, they stressed
that the most important consideration is the extent to
which reasonable risk control advice can be accessed and
implemented by users in practice. They also mentioned
that control banding has a far greater likelihood of meeting
users’ needs in a context where risk assessment experts
can not readily be accessed. Their opinion can be
interpreted as actively supporting the conclusion of this
study, that the control banding judgment is safe-sided,
and that control banding has potential for future utilization
including screening use in Japan.
When control banding provides a truly safe-sided (over-
controlling) judgment, the employer will bear a wasteful
burden if they obediently implement control technology
according to the judgment. In reality, it is predictable
that there will be frequent cases where employers actively
seek expert advice after suspecting the accuracy of the
judgment in attempts to avoid potential waste of
resources. Also, there can be other cases where employers
neglect seemingly over-controlling judgment at their own
discretion because they are cautious about the excessive
cost burden. Considering such situations, and also
considering the fact that control banding often advises
control approach criteria 4, it can be concluded that the
necessity of providing health experts for control banding
is substantially high in reality, although the control
banding system is often perceived as not requiring experts.
The COSHH Essentials on-line interface actually
recommends users to consult certified occupational
hygienists when it assigns a case as control approach
criteria 4; on-line linkage to the British Occupational
Hygiene Society is also provided 6). The existence of
established occupational hygienists is thus one of the key
infrastructures for smooth functioning of control banding
system. Therefore, in order to aim for encouraging
utilization of control banding in Japan, it would be
essential to establish institutional and social mechanisms
for facilitating employers’ access to expert advice with
ease and at low cost. It would also be indispensable to
develop new local health experts such as those available
in the U.S.A. or European countries.
491
Conclusion
The appropriateness of control banding system was
evaluated based on workplace safety by assessing risks
of chemical handling tasks with control banding and then
comparing the results with a practical risk assessment
outcome performed with the comprehensive risk
assessment. It was demonstrated that control banding
tends to provide a safe-sided judgment. A possible
interpretation of this is that control banding is inherently
designed to secure workplace safety by compensating for
insufficient exposure information with safe-sided
judgment criteria and by requiring expert advice in high-
risk cases. Control banding could be widely and
effectively utilized by employers in Japan, provided that
the above characteristics are pre-acknowledged by users.
To this aim, it will be essential to establish institutional
mechanisms which facilitate development and utilization
of new local health experts. However, it should be noted
that the scope of tasks and chemicals evaluated in this
study was limited.
Acknowledgments: This study was supported by the
Labor Science Research Fund (the Program for
Comprehensive Occupational Safety and Health
Research) under supervision of the Ministry of Health,
Labour and Welfare, Japan: subject number of H16-
roudou-6.
References
1) Mori K: Gendai sangyou-hoken no kadai to korekara.
[The issues and the perspectives of today’s occupational
health; JP]. Hataraku Hitono Anzen to Kenko 5, 50–
55 (2004) (in Japanese)
2) Guidelines for risk assessment of chemicals and other
hazards. Tokyo: Ministry of Health, Labor and Welfare
Japan, 2006.
3) Tischer M, Bredendiek-Kamper S and Poppek U:
Evaluation of the HSE COSHH Essentials exposure
predictive model on the basis of BAuA studies and
existing substances exposure data. Ann Occup Hyg 47,
557–569 (2003)
4) Jones RM and Nicas M: Evaluation of the COSHH
Essentials foe vapor degreasing and bag filling
operations. Ann Occup Hyg 50, 137–147 (2006)
5) Brooke I: A UK scheme to help small firms control
health risks from chemicals: toxicological
considerations. Ann Occup Hyg 42, 377–390 (1998)
6) Health and Safety Executive. UK: COSHH Essentials,
Easy steps to control health risks from chemicals.
(online), available from <http://www.coshh-
essentials.org.uk/>, (accessed 2006-09-10).
7) European Chemicals Bureau: Classification and
labeling, Search Classlab, Search Annex 1. (online),
available from <http://ecb.jrc.it/classification-
labelling>, (accessed 2006-11-05).
8) American Industrial Hygiene Association: Laboratory
 492
Programs, Laboratory Accreditation Programs,
Accredited Labs. (online), available from <http://
www.aiha.org/Content/LQAP/lqap-splash.htm>,
(accessed 2006-10-24).
9) Nelson AL, Kenneth AB and Jeremiah RL:
Occupational exposure sampling strategy manual,
NIOSH Publication No. 77-173. Cincinnati: NIOSH;
1977 Jan. (online), available from <http://
www.cdc.gov/niosh/docs/77-173>, (accessed 2006-10-
25).
10) Booher LE, Gelatt RH, Barone NJ and Lockwood MC:
A global approach to exposure/health risk assessment.
Proceedings of the 6th IOHA conference; 2005.
(online), available from <http://www.saioh.org/
ioha2005/Proceedings/Papers/SSC/PaperC5web.pdf>,
(accessed 2006-10-10).
11) European Chemicals Bureau: Classification and
J Occup Health, Vol. 49, 2007
labeling, Annex VI, General classification and labeling
requirements for dangerous substances and
preparations. 2001. (online), available from <http://
ecb.jrc.it/classification-labelling/>, (accessed 2006-10-
10).
12) DiNardi SR: The occupational environment—its
evaluation and control. 2nd ed. Fairfax: AIHA, 2003:
943.
13) Mori K and Takebayashi T: Interview study on
autonomous chemical management system and the
contribution of occupational health specialists in
companies, JP. San Ei Shi 46, 181–187 (2004) (in
Japanese with English abstract)
14) Money C, Bailey S, Smith M, Hay A, Hudspith B,
Tolley D, Dobbie J and Jackson H: Evaluation of the
utility and reliability of COSHH Essentials. Ann Occup
Hyg 50, 642–644 (2006)