study of glass injuries among bar staff found risk assessment poor and health
and safety standards lacking. By Alison
Warburton and Jonathan Shepherd
per cent of assaults treated in UK emergency units are caused by bar
glassware1, 2. These usually lead to permanent, disfiguring facial scars3.
an occupational health viewpoint, bar workers are at high risk of injury from
bar glassware, many involving deep structures such as ligaments, that are
study revealed that 41 per cent of bar workers in South Wales had suffered such
an injury5. Fifteen per cent of these had sustained five or more separate
"sharp" injuries from broken glassware. In another study6, 74 per
cent of bar workers reported lacerations from broken glassware at work, and of
these, 18 per cent had been injured on more than 10 occasions.
addition, most of those interviewed and examined for glass scars had had some
contact with body fluids such as blood, vomit, urine, and faeces, increasing
the potential risk of hepatitis B transmission6.
levels are unacceptably high, and might be reduced by the training of this
occupational group, the introduction of glass manufacturing and handling
standards and improved bar management.
aim of this study, therefore, was to develop a bar worker’s perspective of
assault in bars and to test the hypothesis that bar managers’ estimation of
risk of violence reflects actual risk.
study was designed to relate perceived risk with injury experience of bar
workers, which is important because the degree of concordance of perceived risk
with actual risk helps define objectives in health and safety training in this
occupational group. Of particular interest was the incidence of violence, the
extent of the use of glasses and bottles as weapons, and the site and severity
of injuries sustained. A further aim was to investigate the risk and
circumstances of accidental glass injury.
study of predictors of bar glass injury was performed as a field study in the context
of a randomised-controlled trial of toughened glassware, which has been
reported separately7. The subjects were 1,229 bar workers recruited from a
random sample of 57 bars in South Wales, Bristol, and the West Midlands.
were briefed on the aims of the research, and their role was explained. They
then completed an initial structured questionnaire designed to ascertain the
extent of glass-related injuries and the exact nature and type of bar operated.
Licensees were asked to categorise their bar in terms of type (bar, club,
restaurant), clientele (age range), and popularity (busy, steady, quiet). To
assess the risk of violence, licensees were asked to rate risk on a visual
was continued until the number of eligible bars was sufficient to yield at
least 600 bar workers. Over a six-month period injury was recorded using
questionnaires distributed monthly via licensees to all bar staff. Workers were
asked to record details of violent incidents that had occurred during the
previous month, and the nature, seriousness, and use of glasses and/or bottles
as weapons. With reference to accidental glass injuries, staff were asked to
record details of the type, cause and seriousness of their injuries.
Throughout, the term "glasses" is taken to mean drinking glasses, as
distinct from bottles that, although made of glass, are categorised separately
and distinguishable by shape.
the six-month period 1,229 questionnaires were completed and returned by bar
workers (782 females and 445 males). A cross-section of all bar types was
included in the survey: bars with standard opening hours, bars-come-clubs with
late licenses, family orientated bars, bars with restaurant facilities, modern,
traditional, and sparsely furnished bars, and from town, city and country
locations. Most catered for a mixture of age-ranges. The average age of staff was 32 years (range 14-74), and the
average length of service in the bar surveyed was 41 months (range: 0.25-360).
unstructured interviews, licensees spoke freely about incidents: fights that
ceased before damage and/or injury occurred, or were perfunctory. Some
licensees and their staff considered these incidents to be normal and
unavoidable, while others considered minor scuffles to be violent incidents.
asked formally if there had been any "trouble caused by drunkenness while
working in the bar", staff reported 199 incidents. Of those, 88 (44 per
cent) led to injury. Customers had been injured in 60 (69 per cent) of these
assaults. On 26 occasions (30) injury had been inflicted with glasses, on 13
occasions (15) with bottles, and on other occasions by body parts. Taking all
violent incidents into account – including those that did not lead to injury –
50 (25) involved glasses and bottles.
if they had ever been "stressed by the drunken behaviour of
customers", 148 (12 per cent) staff said they had. These stresses were
incurred by customers being verbally abusive; using bad language, and harassing
incidents were categorised in terms of licensee-assessed risk factor (bar
dangerousness). There were close similarities between bars in all risk
categories both in relation to the number of violent incidents, and in the
proportion of these in which injuries were sustained. When the number of
violent incidents is expressed as a ratio of the number of bars, one in four
bars in low- and medium-risk categories (1:5 in the high-risk category)
experienced violence. Of the incidents, 60 per cent led to injury in the
medium-risk category, while only 38 per cent and 37 per cent of incidents led
to injury in the low- and high-risk categories respectively.
statistical analysis, the high-risk and medium-risk bars were merged. The
number of incidents per bar was seven per year in the low risk group, and nine
per year in the medium/high group, giving a risk ratio of 1.2 (95 per cent
confidence interval (CI) 0.9-1.7, NS).
disregards the precise distribution of incidents across bars, but does suggest
that there may be a relationship, although there was no statistical
significance. For incidents leading to injury, the rate ratio is 2.4/1.4 or 1.8
(95 per cent CI 1.1 to 2.8 p<0.05).
licensees accepted violence as an issue in bars but not all exhibited
responsible attitudes. Those with responsible attitudes took precautions to
avert or calm violence when necessary, and tended to be either tenants or
owners of the bar, as opposed to managers.
precautionary methods were employed, both singly or in combination, and at the
discretion of the licensee. These included employing door staff and providing
training for bar workers, using alternatives to standard glassware (for
example, toughened glass or plasticware), and banning drinks served in bottles.
Additionally, some licensees controlled troublesome clientele strictly. Only
one bar had installed CCTV.
average of 0.7 injuries per bar worker per year was estimated. Over the
six-month period 413 incidents were recorded; 107 staff said they had incurred
injury on at least one occasion, while 31 had been injured on five or more
occasions in a single month. Bottles had caused injury on 139 (34 per cent)
occasions, glasses on 176 occasions. Of the glass injuries, 114 (28) were
caused by one-pint glasses, 37 (9) by half-pint glasses, and the remainder by
other glass types.
1,229 completed questionnaires, 198 bar workers (16 per cent) reported an
accidental injury during the survey period. That is 16 per cent of staff
sustaining injury on 413 occasions, suggesting that some bar workers are more
at risk of injury than others are. The mean number of injuries per injured bar
worker was six per year. No apparent relation was found between number of
injuries and hours of work, age, experience, or licensee-assessed dangerousness
of the bar.
bottle injuries were recorded in detail; 43 were hand injuries, 42 bar workers
were treated in the workplace with first aid, and two required sutures in
hospital. The majority did not require time off work, and injuries caused only
minor inconvenience. Five injuries (10 per cent) severely affected lifestyle,
but only one resulted in time off work (four weeks). One injury was reportedly
sustained in an assault.
bar managers kept large containers behind the bar to store empty bottles to be
sorted for recycling. With this practice 29 per cent of injuries occurred while
emptying the containers.
176 injuries were reported in which glasses had caused injury, 127 bar workers
were treated in the workplace with first aid, and five required sutures in
hospital. The majority did not require time off work, and injury caused only
minor inconvenience. Thirteen injuries (7 per cent) affected lifestyle
severely, and eight resulted in time off work (one day to four weeks). Injuries
were sustained while collecting glasses (81), washing up (57), or performing
other tasks (39), such as clearing away broken glass. Two bar workers were
injured in assaults.
was expected the licensee would be able to make a realistic assessment of the
likelihood of violence, and consequently act to safeguard both staff and
customers. However, it is apparent that licensee-assessed risk of violence is
an extremely unreliable measure of dangerousness. No statistically-significant
relationship between actual risk and licensee-assessed risk of violence could
be demonstrated, but the risk of injury in low-risk bars was just significantly
lower than the risk in the medium/high- risk bars.
explanation is that licensee opinion of what constitutes violence varies. Some
categorise "minor scuffles" as violent incidents, while others accept
these and the injuries sustained as routine and of no real cause for concern.
who are managers, as opposed to owners or tenants, tend to take fewer
precautionary measures. Those licensees who own their bar, or are tenants
appear to have the greatest incentive to avoid violence. Bar owners/tenants are
also more likely to have been in the trade for many years, and thus have more
experience than many managers, who, in some cases, have been in the trade for
only a few months. This is reflected in the attitudes of bar owners/tenants –
most are aware of the potential for trouble and take precautionary measures.
Interestingly, although precautionary measures included changing drinking ware,
only toughened glass has been subject to formal evaluation7. Many are also
members of professional organisations and local associations and keep up to
date with developments in the trade.
While the location of a bar and the clientele it attracts may influence
dangerousness, a competent licensee should have the training and processes in
place to detect precursors to violence. Adequate surveillance of customers is
clearly a prerequisite.
universal training in violence prevention and partnerships with the police and
community safety initiatives are clearly needed. Voluntary collaborations have
been established in some areas8. Mandatory training is probably required.
Schemes are becoming available for licensees and those preparing to become
licensees, to obtain relevant qualifications, but these need to be open to all
bar workers9. The regulation of training must be developed with multinational
bar chains and breweries. Greater incentives are needed to invest in this area.
Some firms provide short courses on core skills, which include awareness of
security, drugs and violence9. Such initiatives now need to be generalised and
be subject to regular external review.
It is also evident from this study that bar workers do not record all incidents
of violent behaviour. Thus, the scale of the problem of violence in bars
remains uncertain. Fear of repercussions, the acceptance of such behaviour as
normal or apathy could account for staff silence.
issue of violence is taboo – few licensees admitted to having problems with
violence, especially those in the high-risk category. Many managers of
high-risk bars were asked to participate in this study, but refused. However,
this study and hospital and police data indicate that violence in licensed
premises remains a substantial problem.
particularly promising means of auditing licensed premises violence is through
hospital emergency unit data, linked to Crime and Disorder Act local crime
audits10. Past research shows that only one in nine assaults in licensed
premises resulting in hospital treatment have been recorded by the police: to a
substantial extent, bars are not included in any formal violence surveillance9.
of hospital-derived data on assault in bars would facilitate improved levels of
safety. In particular, the stigma and potential loss of revenue associated with
publication of injury data in local media would provide powerful incentives to
bar management to improve levels of safety.
terms of accidental injuries to UK bar workers there appears to have been a
reduction in the proportion of staff injured on at least one occasion since
previous small scale, localised studies5,6. Similarly, only 28 per cent of all
injuries in this study were reportedly due to pint glasses compared with 66 per
cent previously. This suggests a real reduction in the rate of actual injury,
perhaps because staff have become more aware of potential danger.
Alternatively, it could reflect a change in drinking habits and a shift from
glass to bottle injuries.
safety guidelines: According to the industry, bottles are becoming
increasingly popular. Consistent with this, bottle injuries accounted for 34
per cent of accidental injuries in this survey. Of these, 29 per cent resulted
from the hazardous practice of discarding empty bottles into large containers
to be sorted.
and safety guidelines on bottle safety should therefore be established to
protect staff. Many licensees reported problems with bottles being very thin
and easily broken which suggests that manufacturing impact-resistance standards
should be established to increase and maintain bottle safety.
L Warburton is research fellow and Jonathan P Shepherd is professor of oral and
maxillofacial surgery at University of Wales College of Medicine, Cardiff
Shepherd JP, Shapland M, Pearce NX & Scully C Pattern, Severity, and
Aetiology of Injuries in Victims of Assault. J Roy. Soc. Med. 1990; 83: 75-78
Hocking MA. Assaults in South East London. J Roy. Soc. Med. 1989; 82: 281-4. Shepherd JP, Price M & Shenfine P Glass
Abuse and Urban Licensed Premises. J Roy. Soc. Med. 1990; 83: 276-7.
Evans DM. Hand Injuries Due to Glass. J. Hand Surg. 1987; 123: 284.
Shepherd JP, Brickley MR, Gallagher D et al. Risk of Occupational Glass Injury
in Bar Staff. Injury 1994; 25(4), 219-20.
McLean W, Shepherd JP, Brann CR & Westmoreland D. Risks Associated with
Occupational Glass Injury in Bar Staff with Special Consideration of Hepatitis
B Infection. Occ. Med. 1997; 47:147-150.
Warburton AL & Shepherd JP. The Effectiveness of Toughened Glassware in
Terms of Reducing Injury in Bars: A Randomised Controlled Trial. Injury Prev.
2000, 6: 0-4.
Portman Group In: Keeping the Peace – A Guide to the Prevention of
Alcohol-related Disorder. The Portman Group, London. 1998
Shepherd JP, Shapland M, & Scully C. Recording of violent offences by the
police: an accident and emergency perspective. Med. Sci. Law 1989; 29: 251-257
Shepherd JP. Preventing Violence. Magistrate 1999; 55: 168-169
authors are grateful to the PH Trust and Demaglass for financial support, and to
all the publicans and their staff who participated in this survey. For
statistical advice and analyses we acknowledge the help of Dr Robert Newcombe,
Department of Medical Statistics, University of Wales College of Medicine,