Safe Event Management

Theatre Managers Association Conference

Birmingham 10th June 2008

Dr Mick Upton
Centre for Crowd Management & Security Studies
(Buckinghamshire New University)

Abstract
This paper argues that safe event management planning is only fully effective when an
in-depth event risk analysis underpins a risk assessment for crowd safety. Furthermore, a
purely quantitative approach to risk assessment can divert the practitioner’s attention
away from important staff training needs. A current popular likelihood x consequence
model for risk assessment is reviewed and a critique of the limits of mathematical
modelling is considered. Finally an overview of the level of fatal accidents at concert
events is considered as a basis for risk analysis.

Introduction
The introduction of legislation in the form of the Health and Safety at Work Act (1974)
made it mandatory for all employers to conduct a risk assessment for their work activity.
At the time there was perhaps a popular belief that this new legislation was intended for
industrial and commercial workplaces not entertainment events and venues. The full
impact of the Act was realised however when the Head of the Policy Branch of the
Health and Safety Executive explained in an important seminar paper that a mandatory
requirement for risk assessment was intended for all employers. Naturally this included
event organisers and venue operators (Graham 1993). In response to questions on how a
risk assessment might be conducted, Graham explained that the basic principle of
assessing risks were essentially identify hazards and then evaluate the risks, i.e. the
likelihood of the hazard arising and the harm it could cause. Furthermore, provided the
assessment is carried out in a structured way, taking into account the known and
foreseeable hazards, there should be no difficulty in identifying the significant risks and
establishing the relative priority for action.

Possibly in support of Graham‘s 1993 seminar paper, the Health and Safety Executive
(HSE) published a report that provided a formula for risk assessment (Au et al 1993). The
report advised that a numerical figure could be applied to the likelihood of an accident
occurring and the consequences, in terms of severity of injury, if it did. A risk factor is
then established by multiplying the likelihood figure by the consequence figure and the
result transposed into high medium or low risk. The likelihood x consequence model for
risk assessment remains today as possibly the most popular method.
Critique
On face value the likelihood x consequence model appears to be perfectly logical. It is
possible to measure some risk in this fashion, for example the integrity of structures or
temporary structures. A quantitative approach therefore has a valid place in the risk
assessment process. The big question however is, can a purely quantitative approach
predict human behaviour?

A numerical risk assessment model was initially adapted from a commonly used
industrial quantitative method intended to predict the possible failure rate of mechanical
objects. The model was therefore developed by engineers to satisfy a political need to
introduce a system that would fit health and safety legislation.

When dealing with crowd activity however, awarding a numerical figure becomes the
personal opinion of the assessor rather than a scientific system of measurement. Another
assessor might have a different opinion and therefore award a different number to the
same activity, an assessor may in fact introduce any form of numbering they chose as
there is no mandatory requirement to use the HSE suggested system. Where individual
assessors can, and often do, invent their own numerical system confusion might result. In
a scenario where the assessor and the local authority officer (the recipient) are using
different interpretations of a numerical system it is not difficult to imagine that the
proposed risk assessment might be completely misunderstood by the recipient.

In fairness to the researchers that published the numerical system it should be noted that
at least one of them later changed his mind. Zachary Au was a researcher for the 1993
report and he later published a second report in which he favoured an alphabetical system
(Au 1997). The A,B,C categorisation of risk, with A representing high risk, once again
appears to be a logical system, that is until you consider that some sport events categorise
their events in reverse order. In other words C becomes the highest category of risk not
A. It follows therefore that the assessor needs to explain the methodology used to
construct the assessment if confusion is to be avoided.

One year before Au had published his second report recommending an alphabetical
system, Brian Toft (1996) had in fact published a critique of the limits mathematical
modelling of disasters. Toft made a very important point when he argued that individuals
create their own sets of criteria against which risk is interpreted. Toft was emphasising
the fact that risk perceived by a given society or individuals are not objective but
subjective. In other words, individuals perceive risk differently; consequently some
people enjoy so called dangerous sports while others regard them as high-risk actions
bound to lead to serious injury or even death. This is not to imply that the participant in a
high-risk sport is unaware of risk, rather that they feel that they are in control and
therefore managing risk. The fatal accident record for concert events however illustrates
that crowd members are not always fully in control.

Fatal accidents
My own research into concert crowd related accidents/incidents during the period 1974 –
2007 produced the following data:
• 34 died during ingress
• 42 in front of a stage
• 5 during egress
• 53 ingress related
• 13 fell from balconies (seated areas)
A total of 147 deaths in thirteen countries. The figure of 147 is not claimed to be a
definitive level of the accident rate, rather it is an example of the type of accident
suffered at concert events. The figure of fifty-three ingress related does require further
explanation, it was a single accident in a situation where a crowd leaving a venue entered
directly into a public transportation system. I include it because it raises the question of
boundaries of responsibility.

It has been argued that in percentage terms 147 fatal accidents in thirteen separate
countries over a thirty-year period is not that high. But the figure does not include the
‘near miss‘ occasions that we have all experienced and, it should be remembered that the
figure is merely a snapshot of concert event risks. If sport events and incidents of fire are
added, the fatal accident rate increases by leaps and bounds. In America one hundred and
twenty people died in just two fires at venues in Rhode Island and Chicago. The overall
accident rate has in fact prompted politicians in the UK to call for the regulation of social
events.

Regulation v education
In the UK there has been at least six attempts by politicians to introduce Bills to the
House of Commons aimed at either regulating entertainment events or licensing
promoters and security staff. Only the Private Security Industry Act 2001 became
legislation. The 2001 Act is however aimed primarily at door supervisors; it offers
nothing to aid the crowd safety practitioner. After over thirty years of accidents at concert
and sport events, health and safety legislation supported by guidance still underpins our
approach to crowd safety. It was for this reason that Buckinghamshire New University
(BNU) introduced a two year distance learning Foundation Degree course within what is
the Centre for Crowd Management and Security Studies.

The course is supported by the United Kingdom Crowd Management Association

(UKCMA) and the Production Services Association (PSA). It encourages students to use
case study to underpin legal frameworks and published guidance in an approach to crowd
safety planning. Much use is made of Extreme Value Analysis (EVA) which requires the
assessor to demonstrate a scientific approach to risk analysis. The EVA system considers
accidents and incidents in terms of:
• Low frequency – high intensity
• High frequency – low intensity

A classic example of low frequency – high intensity is the 1989 Hillsborough football
disaster that claimed ninety-six lives. It might be argued that had the EVA system been
applied to the risk assessment for the match it would have revealed a remarkably similar
football accident in 1946, which claimed thirty-three lives at Bolton. The lesson learned
is that time can lull us into a false sense of security. In the forty years between these
accidents the circumstances were simply forgotten.

A high frequency – low intensity incident is one that happens often but does not
necessarily result in a high casualty rate. An example being a large number of distressed
people being extracted from a crowd at the front of a stage at a rock concert and treated
by in site medical teams. This second category might perhaps be referred to as near miss
situations and arguably the best source of learning if incidents are recorded. The key
factor here is that high crowd density situations are being allowed to develop.

The primary aim of crowd safety planning therefore is to control crowd density. The
American safety planner C.B. Berlonghi provided a graphic description of the effects of
high density, as follows:
“ A dense crowd is one in which individual physical
movement is rapidly becoming less likely or possible due
to the density of the crowd. People are attempting to
move but they are either swept along with the movement
of the crowd or are falling on top of each other. The
results of this compression of people are fatalities and
serious injuries due to suffocation”. (Berlonghi 1993)

Reports on the Hillsborough disaster tell of forces so powerful that people were propelled
several metres and in one case a police officer mounted on a horse was actually moved a
metre. In high crowd density it is also common for people to have clothing torn off or
loose their shoes.

Approach to planning
It would be impossible here to fully explain a two-year university course approach event
safety planning. Briefly, the students use EVA to evaluate each of the stages of arrival
and ingress, attendance, egress and emergency evacuation for their chosen venue and
event.

Crowd arrival & queuing

Turning first to crowd arrival and queuing, every crowd safety practitioner should be
familiar with a queue theory that argues that there are five considerations to queue
management. The five considerations are:
1. Pedestrian approach to a queue
2. Standing space within a queue
3. Unrestricted forward movement
4. Passing through an ingress system
5. Dispersal

An assumption is made that where arrangements are made appropriate to the needs of a
queue, ingress will be achieved smoothly. The weakness with queue theory however is
that it implies that a crowd will always act rationally. The key factor is ‘where
arrangements are made appropriate to the needs of a queue‘. If we accept that risk
perceived by a given society or individuals are not objective but subjective then queue
management needs to consider how the psychology of people queuing might affect a risk
assessment. Evidence of the emotional state of some crowd members that attend rock
concerts can be found in an article written by fifteen year old Terri Sigmon, a student at
Western Hills High School, U.S.A. In her school paper Sigmon anticipates a forthcoming
concert event and describes her previous experience at rock concerts. Under the headline
“Concerts, a real trip”, Sigmond wrote:
“You lost sight of the people you came in with as soon as
you entered the crowd, but that doesn’t matter now.
Nothing matters except for your fight to get to the front of
the crowd. But that is everyone else‘s goal and it‘s every
man for himself as you shove, claw, and push your way
forward. Covered with sweat, you feel as though you may
faint at any moment, but you keep going …..
All sense of reality has disappeared in the struggle to
move forward” (Sigmond 1979)

The concert that this student was looking forward to was an appearance by The Who at
the Riverfront Arena, Cincinnati in 1979. This event turned out to be the worst concert
disaster in the history of rock ‘n roll when eleven people died during a queue
management and ingress failure caused by a high-density bulk queuing situation. It is
interesting to note that there were at least three serious bulk queue incidents previously at
the Riverfront. In 1976, 1977 and again prior to the Who disaster in 1979 the police
responded to bulk queue situations that caused high crowd density situations and serious
crushing outside the venue prior to doors opening (Fuller 1985).

Moving forward in time, on the 18th July 1995 a sixteen year old young woman and a
seventeen year old young man died and many others were injured as a result of an ingress
failure at the Atrock Rock Festival in Israel. Two days later a fifteen-year-old young
woman also died as a result of injuries she received during the disaster. The lessons of
1979 had again been forgotten with time.

If we want to conduct comparative studies for queue management and ingress failures,
case studies can easily be found. It might be argued that both Burden Park and
Hillsborough are examples. More recently, in 2005 it was reported that police and
ambulances had to respond to a serious queue management and ingress failure at the
opening of a new IKEA store in North London (BBC News 2005). In 2007 police and
ambulances were called to a serious queue management and ingress failure at a sale at a
Primark shop in Oxford Street, London (Roche & Myall (2007). In both cases it was
reported that security staff had been employed specifically for queue management and
ingress control.

Simply employing contract security staff is clearly not planning for queue management.
The term ‘appropriate to the needs of a queue‘ should take a very broad view. Queue
management is underpinned by the formation of a linear queue system early on.
Sufficient staff trained to recognise and control changes in density should supervise
systems. The correct use of crowd control barriers is also essential, as is efficient
communications. Licensed security staff should be on hand to deal with anti social
problems but it must be remembered that current qualification for a security license does
not include crowd safety management.

Once doors are open an efficient processing system includes clear signage and conditions
of entry well publicised. Surprisingly, current published guidance for concerts and events
does not provide calculations for ingress flow turnstiles. The reader is referred to the
Guide to Safety at Sports Grounds (Green Guide) for sample calculations. The flow rate
quoted for turnstiles is 660 persons per hour. This figure naturally assumes that the
turnstiles in question are well maintained and in excellent working order. It is emphasised
that the rate quoted is maximum flow, where searches or ticket purchasing is taking place
this rate will obviously be reduced dramatically.

The foyer is important, it is the distribution area. People unfamiliar with the building will
pause to read signs to find their way and at large venues signage should be at high level.
Program sellers often want to be immediately in front of an ingress flow, this is where
they can sell the most programs and the practice must not be allowed. A key issue is the
admittance of people in wheelchairs. How do they enter, where do they go and how do
they get there.

Non turnstile systems

For the venue that has a free flow ingress system, supervisors should be trained to
calculate pedestrian flow rates. Flow rates for turnstile and non turnstile venues should be
regularly passed to management and logged for use in the event of a necessary
emergency evacuation being needed before ingress is complete. If you don’t know how
many people in a sudden emergency situation are in how do you know how to evacuate
them? Flow data can also be useful after the event to aid venue merchandisers in future
planning.

Attendance
The data provided above reveals a surprising number of falls from balconies during rock
concert events. The need for good stewarding in these areas is therefore obvious.

At least forty-two people have been the victims of a front of a stage accident. In the
majority of cases victims were crushed as a result of high-energy cultural activity. Where
front of stage pits are used to control these situations it is essential that pit teams receive
specific training for their role and responsibility and training records kept. At the
university we also run a course that offers a certificate in pit training. In America at least
one contract security company has been forced to pay compensation for injury to a young
woman in what was alleged to have been the mishandling of a front of stage incident.

Density
Current guidance for a ‘safe density‘ for standing audiences is 0.5m2 or two persons per
square metre, with an acceptable level of three persons per m2 directly in front of a stage.
Controlling density within a standing audience is however very difficult at rock concerts.
Current cultural behaviour can range for pop hysteria to rock cultural behaviour that can
easily create intolerable pressure loads. Activities such as mosh pits and wall of death
make a mockery of 0.5m2 as people move out of the way or people migrate from seated
areas to join in. Natural laws of crowd dynamics then produce lateral or dynamic surges
that can become static loads in a crowd collapse. Activities such as surfing, stage diving
and skanking can also cause a crowd collapse in which case response teams have a
maximum time frame of three minutes to extract people in distress. Any longer and
victims can suffer brain damage or death.

The known actions of performers should be researched and levels of responsibility

established. Particularly where a performer(s) has a reputation for encouraging high-
energy crowd activity that might result in a serious accident. Who conducts a risk
assessment for such activities? My experience has been that lawyers acting for the
performer will try to pin the blame on either the local security team or venue
management or both. In which case the venue operator will need to ensure that they can
produce unambiguous documented evidence of levels of responsibility, risk assessment
and risk management strategies and staff training records.

Egress
This afternoon Keith Still will make a presentation on crowd flow systems that includes
exit capabilities therefore I will restrict my observations on egress to that of emergency
evacuation.

Egress planning is always based on the ability of a venue to evacuate its total capacity
within an emergency evacuation time frame stipulated in published fire safety guidance.
Recently the old `yellow guide` guidance for fire safety was upgraded to become a series
of venue specific guides. The reason for the new guides is a requirement now for venue
operators to conduct a fire risk assessment for their venue or place of work. Evacuation
times for entertainment venues however remain unchanged.

The notion that an emergency evacuation can rely on a theoretical calculation that a given
number of people can exit within an agreed time is however a very simplistic approach in
my opinion. In practical terms there are at least five key questions that need to be
answered:
1. How will the event be stopped and stage sound cut in order to announce the need for
an orderly evacuation?
2. At what stage of a potential alert is the crowd informed?
3. How will the crowd be informed?
4. At what point are lighting levels raised?
5. Are your staff trained to deal with the situation?

Stopping a rock concert when it is in full flight is no easy matter, it requires the full co-
operation of the artiste and the production crew. Even the most placid performer is
unlikely to welcome someone simply walking onto the stage and demanding the show be
stopped. Research into serious incidents and fires such the Stardust Disco fire and the Big
Day Out crowd collapse indicate that that stopping an event and taking decisive action
can be very difficult.
In his study of human behaviour in emergency evacuation situations Jonathan Sime
(1993) argued that venue emergency planers needs to address the relationship between (a)
design and engineering x (b) communications technology x (c) crowd management x (d)
crowd behaviour and movement. Sime‘s research led him to conclude that the time to
escape (T) is a function of T1 (time to start to move) + T2 (time to move to and pass
through exits) rather than T=2 alone.

Similarly, Au et al (1993) modelled four factors affecting likely human behaviour in an

emergency evacuation scenario, each factor having a number of variables.

1.1 Environment
1.2 Availability and clarity of information 1.Sense

2.1 Content of information

2.2 Information source 2.Interpret
2.3 Knowledge, Experience &
Expectations

3.1 Knowledge & Experience

3.2 Goals & objectives 3. Decide
3.3 Alternative & Choice
3.4 Consequence
3.5 Mental Condition & Emotion

4.1 Physical Condition of the individual

4.2 Venue Physical Characteristics 4. Act
4.3 Venue Characteristics at any time

Summary
At an inquiry into a crowd-related accident lawyers for the victims will seek to establish
blame because blame equals compensation. A submission of a risk assessment based
solely on a quantitative method that argued that the event was low risk on the basis that it
scored a risk factor of 24 is fundamentally flawed. A consultant is likely to be called to
explain that a remarkably similar accident had occurred previously. The key question
then asked is did you know of it?

The key to risk assessment is a risk analysis process that considers the phases of arrival
and queuing, ingress, attendance, egress and evacuation as separate activities. The
potential for accidents during each activity should be acknowledged and the management
actions taken to eliminate risk shown. A risk assessment should also acknowledge the
risks associated with engineering. In the past there have been accidents at venues that
include: the collapse of a seating block; a fractured hot water pipe in an auditorium; pieces
of ceiling falling onto an audience due to high volume noise and people falling into the
orchestra pit when performers invited the audience onto the stage.

A crowd management plan should include venue ingress and egress crowd flow
calculations and staff training levels of qualification which should be based on National
Occupational Standards (NOS). In the event of a front of stage accident insurance
companies will also want to see an acceptable qualification for pit team training. The plan
should also provide details of emergency action drills for fire, bomb threat, crowd surge
and crowd collapse and evacuation. It should also be shown that regular exercises had
been carried out to test and upgrade planning.

Finally the question of who is responsible for a risk assessment for the performer and the
production crew needs to be addressed, particularly with regard to one off concert events.
It is very unwise to ignore the performer that has a reputation for encouraging high-
energy crowd activity but it often happens.

Bibliography
Au Z. et al 1993: Managing Crowd Safety in Public Venues: HSE Contract Research
Report No 53/1993: Risk Prioritization p5.16 – 5.18: H.S.M.O.
Au Z 1997: Prototype Methodology for the Assessment of Risk to Crowds in Public
Venues HMSO
BBC News 2005: Television News Report, Crush chaos at IKEA store opening: 10th
February 2005.
Berlonghi E. B. 1993: published seminar paper Understanding and Planning for
Different Spectator Crowds: in Engineering for Crowd Safety: Ed R.A. Smith
& J.F. Dickie: Elsevier Science Publishers B.V.
Fuller G.J. 1981: Are The Kids All Right? A review of the Cincinnati disaster: Times
Books
Graham P. 1993: Published seminar paper The Health and Safety at Work Act 1974 and
Crowd Safety: in Engineering for Crowd Safety: Ed R. A. Smith & J. F.
Dickie: Elsevier Science Publishers B.V.
Roche E. & Myall L. 2007: Newspaper report ‘Battle of Primark‘: Daily Express 6th
April 2007
Sigmon T 1979: Quoted by Fuller in Are The Kids All Right? A review of the Cincinnati
disaster p21: Times Books
Sime J.D. 1993: Published seminar paper, Crowd Psychology and engineering:
Designing for People or Ballbearings?:In Engineering for Crowd Safety: Ed R.
A. Smith & J. F. Dickie: Elsevier Science Publishers B.V.
Toft B. 1996: Limits to The Modelling of Disasters, p 101 Implicit assumptions about
risk: in Accident and Design, Ed Hood C. & Jones D.K.C.: UCL Press.

Dr Mick Upton
Following military service Mick began working as a close protection bodyguard for a
wide range of clients. In the mid sixties he began to specialise in celebrity protection
when he accepted work with the Beatles and then the Rolling Stones. Subsequently, as
Head of Security at Artistes Services Ltd he worked with celebrity clients that ranged
from ABBA to Led Zeppelin. From 1974 on he took a keen interest in crowd safety
management at events following the death of a young woman at a David Cassidy concert
at which he was working.

In 1982 Mick formed ShowSec International Ltd, a company that specialised in celebrity
protection and event crowd management at international level. Just some of the events
that Mick has managed include all European Monsters of Rock events, Live Aid 1985,
the Moscow Peace Festival, the V.E. Commemoration Hyde Park 1995 and Royal Ascot.
He was also a consultant to the open space queue plan for the 2000 Sydney Olympic
Games.

Mick has acted as a consultant to Bramshill Police Staff College and he has devised and
delivered crowd safety training for U.K. police services, local authorities and foreign
agencies. He has also acted as consultant to UK official published guidance on concert
event crowd management. In addition he has served on government sponsored Lead
Bodies set up to introduce NVQ qualifications for events, door supervisors and VIP
protection.

During his career Mick was the recipient of a Silver award from the Event Services
Association, three times winner of the Live Gold Award for crowd management
planning, and the recipient of a Police award for designing and delivering training for the
police service.

On his retirement from ShowSec in January 2002 Mick was presented with: a Lifetime
Achievement Award by the Event Services Association, a Lifetime Contribution to
Concert Safety Standards by Total Production Magazine and an award from the Mojo
Barrier Company for his outstanding achievements in the field of Crowd Management.
In 2006 he was presented with the National Outdoor Event Association (NOEA)
Presidents Award and Lifetime membership of the association in recognition of his work
in the development of crowd safety planning and training.

Mick continued his interest in crowd safety and security issues in retirement by working
with Buckinghamshire Chilterns University College (now Buckinghamshire New
University) where he worked with Chris Kemp and Iain Hill on the development the first
ever Foundation Degree course in crowd safety management. Students on the first course
graduated in 2005. In 2005 Mick was awarded a Doctorate for his work and he was
appointed as Head of the Centre for Crowd Management & Security Studies at BNU. He
retired from that position in September 2007. He is currently the Hon Chairman of the
Event Industry Forum.