Mass-gathering Health Research Foundational Theory: Part 1 - Population Models for Mass Gatherings

Adam Lund, MD, MEd, FRCPC; Sheila A. Turris, RN, PhD; Ron Bowles, EMA II, PhD; Malinda Steenkamp, PhD;
Alison Hutton, RN, PhD; Jamie Ranse, RN, FRCNA, FACN; Paul Arbon, RN, PhD5

Abstract
Background: The science underpinning the study of mass-gathering health (MGH) is developing rapidly. Current
knowledge fails to adequately inform the understanding of the science of mass gatherings (MGs) because of the
lack of theory development and adequate conceptual analysis. Defining populations of interest in the context of
MGs is required to permit meaningful comparison and meta-analysis between events.
Process: A critique of existing definitions and descriptions of MGs was undertaken. Analyzing gaps in current
knowledge, the authors sought to delineate the populations affected by MGs, employing a consensus approach
to formulating a population model. The proposed conceptual model evolved through face-to-face group
meetings, structured breakout sessions, asynchronous collaboration, and virtual international meetings.
Findings and Interpretation: Reporting on the incidence of health conditions at specific MGs, and comparing
those rates between and across events, requires a common under-standing of the denominators, or the total
populations in question. There are many, nested populations to consider within a MG, such as the population of
patients, the population of medical services providers, the population of attendees/audience/participants, the
crew, contractors, staff, and volunteers, as well as the population of the host community affected by, but not
necessarily attending, the event.
A pictorial representation of a basic population model was generated, followed by a more complex
representation, capturing a global-health perspective, as well as academically- and operationally-relevant
divisions in MG populations.
Conclusions: Consistent definitions of MG populations will support more rigorous data collection. This, in turn,
will support meta-analysis and pooling of data sources internationally, creating a foundation for risk assessment
as well as illness and injury prediction modeling. Ultimately, more rigorous data collection will support
methodology for evaluating health promotion, harm reduction, and clinical-response interventions at MGs.
Delineating MG populations progresses the current body of knowledge of MGs and informs the understanding
of the full scope of their health effects.
Introduction
Background
Mass-gathering health (MGH) is a relatively new field of research. A mass gathering (MG) is a situation or event
in which crowds gather, and where there is the potential for a delayed response to emergencies because of
limited or delayed access or other features of the environment and location. 1 The science underpinning this
body of knowledge is developing rapidly, but there is still a lack of theory maturity and adequate conceptual
analysis to inform the understanding of MGs.2 In December of 2013, the Flinders University World Health
Organization (WHO; Geneva, Switzerland) Collaborating Center for High Risk/Visibility Events hosted a scientific
meeting in Adelaide, South Australia, bringing together international researchers in the MGH field. The team
included members from: Flinders University and the University of Canberra, in Australia; the University of
British Columbia (BC) and the Justice Institute of BC, in Canada; as well as Public Health England (London,
England, United Kingdom) and the WHO.
The MGH Collaborating Team, formed during this meeting, discussed populations of interest to MGH
researchers and clinicians in order to support a common understanding about the human health effects of
MGs. These discussions formed part of a larger, international consensus project aimed at developing a MGH
minimum data set and accompanying data dictionary. Such a data set will support description, measurement,
comparison, evaluation, and reporting on parameters of interest, permitting international collaboration that
to date has been impossible. Collecting data on a cohesive set of common variables during MGs will help
researchers, policy makers, event operations personnel, and clinicians under-stand the health effects of MGs,
both on those attending or participating in events and on the host communities.3-6
1

In keeping with this goal, defining MG events and the subsequent populations is required to obtain
consistent totals, numerators, and denominators in reporting fields. Consistent data collection will permit
meta-analysis and pooling of data sources internationally, and will create a foundation for risk assessment, as
well as illness and injury prediction modeling. Ultimately, more rigorous data collection will support methodology
for evaluating health promotion, harm reduction, and clinical response interventions at MGs.4,7 The
development of a population model is the focus of this report, and it addresses a gap in the current body of
knowledge with regard to MGs and understanding the full scope of their health effects.
A variety of factors may affect the composition of, and dictate the characteristics of, the population at any
specific event. Because there will always be an interaction between event type and the people who attend or
participate in the event, a population and an event model are both required. Accordingly, this manuscript is the
first of two reports. This first report analyzes the variables of interest regarding the various populations of
people involved in MGs and proposes population models for MGH. The second report addresses the
characterization of events that may be reported internationally, ultimately presenting an event model for MGH.
Conceptualizing a MGH Research System
Mass-gathering health populations of interest are part of larger systems. Research in the MGH field has tended
to focus on ‘‘macro’’ elements separately, including: type of event, event population, patient population,
medical operations, and to a lesser extent, community setting and local health infrastructure. Figure 1 depicts
the macro elements as a whole, in relation to each other. In this report, the authors initially present three
population-related variables drawn from this model then examine how selected characteristics and relationships
between these variables may influence their definition and use in MGH research.
Process/Methods
The development of the population model was undertaken using a 2-pronged approach. The first stage involved
a review and

Lund & 2014 Prehospital and Disaster Medicine

Figure 1. Mass Gathering Health Research System.

critique of existing definitions and descriptions of MGs in published literature. The team attempted to describe
the populations involved in, or affected by, MGs of various sizes and forms. Mass gatherings from small,
medium, and large local events, through to international mega- events and compound events, were considered
in turn. When it became apparent that published definitions lack sufficient depth to support researchers and
clinicians in defining, describing, and explaining MGH effects, the need to build on existing work became evident
and the second stage of the process, formulating a population model, was undertaken.
Taking existing definitions as a starting point, the team employed a consensus approach to delineate the
population groups affected, or potentially affected, by MGs, thus formulating the model described below.
Through face-to-face group meetings and structured breakout sessions in Adelaide, and subsequent
asynchronous collaboration as well as virtual meetings, a conceptual model was developed in an iterative
fashion and presented graphically, capturing both academically- and operationally-relevant divisions in MG
populations.
2

Findings and Interpretation
The results of the iterative process are presented below in four sections. First, a proposed MGH Population
Model is presented as applied in local contexts. Second, global considerations are explored in more detail. Third,
analyses of selected characteristics (or facets) of MG events and populations are described over a time axis.
Finally, an analysis of some initial applications, and related metrics arising from the model, is given.
Elements of a Local Population Model for MGH
The starting point for developing a basic population model consists of three ‘‘nested’’ variables: (1) the number
of people in the host community (NHC); (2) the number of people attending the event (NEV); and (3) the number
of patients who present to medical care (NPP; Table 1). Figure 2 presents a basic, graphic representation of the
relationships between these terms and variables.
These three variables form an important part of MGH research. Though each appears self-evident, there are a
number of factors that researchers should consider in defining and analyzing the variables that they will use in
specific MG studies. Each population in Figure 2 exists in relation to the others. Specifically, researchers should
attend to the nesting of the populations, the boundaries between the populations, and the shifting nature of
those boundaries. The circles in Figure 2 are ‘‘nested’’ to emphasize that the total population of the community
during a
Abbreviation
N

Term
No. of People

Definition
Integer representing the total population of interest.

NHC

No. of People in the Host Community

Integer representing the total population of the host community or
communities, on a nonevent day.

NEv

No. of People at the Event

Integer representing the number of people attending the event,
including attendees, spectators, talent, workforce, etc.

NPP

No. of Patient Presentations

Integer representing the number of patients presenting during the event;
usually presenting to on-site health services, but may include direct
presentations to community health services.

NHCEv

No. of People in Host Community
Plus Event Population on Event
Days

Integer representing the total population of the host community plus the
guests who attend the event but live outside of the community.
Lund & 2014 Prehospital and Disaster Medicine

Table 1. Abbreviation and Definition of Terms for the Mass-gathering Health Population Model

3

Lund & 2014 Prehospital and Disaster Medicine

Figure 2. Populations of Interest at a
Mass Gathering Event

MG event. This event contains a set of sub-populations (ie, those attending the event), and sub-sub-populations
(ie, those attending who present with a health-related issue). The boundaries in the figure are represented as
hatched lines, indicating porosity, fluctuation, and expandability. During MGs, populations are in flux and
‘‘boundaries shift;’’ people may move in and out of the community before, during, and after the event, which
will affect the denominator in most population descriptions.
The population of the host community at baseline (ie, nonevent time), or NHC, is represented by a
proportionately large circle. However, NHC will often increase during event dates. In cases where this increase is
significant, an additional variable, NHCEv, may be useful to indicate the value of the host community’s base
population plus the guests who attend the event but live outside of the community. Depending on the size of
the host community and the scope of the event, the total size of the community on an event day may increase
fractionally, or may double, triple, or increase even further the size of the total population during an event. In
the latter case, the NEv may in fact be greater than the NHC. For example, there is a popular music festival in
Western Canada that takes place in a local town with a population of roughly 1,200 people. During the 2008
festival, the more than 40,000 attendees taxed local infrastructure to the extent that the festival was cancelled
the following year.
8

The population attending the event, NEv, is represented by the middle circle, and includes participants,
attendees, spectators, staff, volunteers, and others. Note that the event population will commonly be composed
of a combination of members from the host community as well as visitors from outside of the host community
who are drawn to the area by the event. Therefore, some population of NHC may migrate into NEv.
Another sub-population is represented in Figure 2 by the penumbra, or fringe. This concept acknowledges the
population not directly involved in the event, but attracted to, or affected by, the event due to proximity (ie, just
outside of the event boundaries). These may include crowds that gather outside of event boundaries (ie, no
ticket holders, fans, and protesters), or members of the host community whose baseline health services are
affected by event-related service interruptions. Although not easy to measure, this population requires
consideration in event planning, particularly with regard to security measures and the possible need to deploy
members of the on-site medical team beyond the borders of a given event.
4

Patient presentations, or NPP, are represented by the inner-most circle. This captures members of the event
population who present for a health encounter, whether related to health promotion, illness prevention, selftreatment, infectious disease, illness, or injury. In some cases, patients may come from the fringe or the host
community. Clarity will be required with regard to how these encounters are counted and reported. Patients may
present to on-site care teams, to health resources outside of the event in the host community, or may delay
seeking care until they can access health services in their home community.
Global/Complex MGH Population Models
The WHO perspective on MGH requires a broad view of MG populations, including consideration of the
international and global contexts. For example, at what point do events acquire

Lund & 2014 Prehospital and Disaster Medicine

Figure 4. Overlapping Functional Communities
Lund & 2014 Prehospital and Disaster Medicine

Figure 3. Nested Geographic Communities

international significance? How might risks related to the transmission of infectious diseases, hazards related to the
environment and geography of an event, or security risks inherent in an event such as the Olympic Games be
captured in population models? An expanded population model would therefore be of value (Figure 3).
The addition of the two outer rings is intended to remind the viewer of multiple perspectives on larger,
internationally-relevant MGs. Depending on where one’s responsibility lies, different considerations may be
relevant. For example:
• from a health/medical operations perspective, the primary focus may be on health infrastructure, of the
event and in the host community, to ensure the safety of all event participants;
• from a host community/health authority perspective, the impact of the event population and the
ability to maintain a baseline health-service capacity for the surrounding community may be of
highest relevance;
• from the perspective of the nation-state, considerations like transportation, security, and border control
may be the most important; and
• from an international perspective, global-health and infectious-disease vectors may be primary considerations,
as well as health security, including obligations under the International Health Regulations.
The graphic representation of the basic Population Model (Figure 2) implies that the event population is drawn
from the local host community. However, MG events can involve regional, national, and even international
5

participants and attendees. In the global context, the host community population from which the event
population is drawn must be carefully described, and may even have to be considered as multiple communities
representing the different groups from which attendees are drawn. For example, the event population in a
large marathon may include predominantly local spectators, but call upon event staff from national
organizations and include significant participation from national and international athletes. Alternatively, from a
functional perspective, the event population (NEV) may form from the intersection of the host community,
participant community, and crew community (Figure 4).
Mass-gathering researchers and medical services providers commonly report NEV and NPP to assess risk,
anticipate patient presentations, and predict resource needs. Increasingly, impact on the host community (NHC)
is reported in the literature as an important variable in planning. Researchers must be aware, however, that
even simple population models can be influenced by a number of characteristics and relationships between
these variables. Thus, researchers must clearly identify their choice and definitions of population variables and
articulate the rationale for these decisions.
9-15

Factors to Consider in Modeling MGH Populations Over Time As previously discussed, populations are not static
over time. Operationally, there are relevant times, or phases, when populations are predictably in flux,
including: pre-event, during event, and post event. As argued by Zielioski and Pawlak, an important aspect of
understanding risk vis a vis events depends on analyzing population movement and the contact between
population groups (eg, exposure of host population to the guest population and exposure of guests to hosts).
The authors of the current report propose that the population needs to be modeled for each separate phase of
an event in order to capture the fluctuations of the model as the events unfold (Table 2).
16

Population Metrics for MGH Research
The population model described above, and the variables included in the model, can be used to study a
number of aspects of MG events. The following discussion explores issues that MG researchers should
consider in using the population variables within their calculations.
Determining the Patient Presentation Rate (PPR)—The PPR, expressed as the rate of patient presentations per
1,000 attendees, is a calculation central to MGH research.17 The PPR is generally expressed by the formula:
PPR5NPP/NEV x 1,000. Two values

6

Domains
International

Discussion and Event Phase
Considerations of global public health and spread of infectious disease before, during, and after the
event.

Nation State

Considerations of transportation, security, and border control before, during, and after the event.

Host Community Population

Considerations in the changing value of NHC over the time period under study:
● Pre-Event
o Baseline population stats re: the defined host community population.
● During Event
o The host community may expand due to influx of people to attend or
provide services related to the event.
Some members of the host community may leave the community to avoid
disruptions related to a large event.
● Post-Event
o In most cases, it is expected that the community population would return to baseline
levels.

Event & Participant Population

Considerations of data collection (eg, obtaining accurate values for NEV), impact on statistical
analysis (eg, changing values of NEV during period of study), and planning and provision of services
(eg, meeting ‘‘surge’’ times or providing care post-event):
● Pre-Event
o In terms of data, event population is based on projections or intentions, which may or
may not reflect the reality of the event day populations.
● During Event
o Actual numbers of those attending the event in all categories are relevant, but may
change on a daily or hourly basis.
● Post-Event
o After the event, the participant population will return to their home communities at
variable rates.

Patient Population

Considerations of data collection (eg, how to define and who to include in population values) and
post-event presentations (eg, whether or not to include post-event surveillance and include in
study data):
● Pre-Event
o In terms of data, patient populations in the pre-event space are to predict patient
numbers, acuity, and case mix to facilitate planning.
● During Event
o Patients may present to the on-site medical team, self-refer to ambulance,
emergency, or local clinic, or not present at all, seeking care in a delayed fashion in
their home community (which may be the host community).
● Post-Event
o Consider post-event illness/injury surveillance.

Lund & 2014 Prehospital and Disaster Medicine

Table 2. Domains and Discussion of Event Phase in Mass-gathering Health Population Modeling

Are necessary to determine the PPR: (1) the number of patients (NPP); and (2) the number of attendees at the
event (NEV). This conceives the PPR of an event as a holistic value useful in determining the overall risk for an
event and in planning medical coverage.
The PPR is a useful calculation for both researchers and clinicians, however, in the current form, the concept
7

fails to capture peaks or surges in patient flow. This is of particular importance for clinicians providing medical
care during an event. For example, during a marathon, the absolute number of patient encounters may not be
high, but if the majority of patients present over a 90-minute period, this will affect staffing requirements for
the event. Therefore, the authors propose that, in addition to calculating the PPR as an aggregated value for
an entire event (or series of events), calculating the PPR/hour may also be useful. This will require that medical
teams have the ability to document the time of arrival and discharge for individual patients.
Determining the Population of Patient Presentations (NPP)—Much of the MG medicine literature presents
case reports and analysis of medical services at MGs and, thus, the number of patient presentations (NPP)
are usually expressed in terms of the number of patients who present to the medical facilities at an event.
However, delineating the patient population may be complicated by factors such as whether or not to
include encounters related to ‘‘self-treatment’’ (eg, receiving a band aid or female hygiene products) or
‘‘dispensary’’ requests (eg, acetaminophen), and whether/how to ‘‘count’’ patients with event-related
complaints who seek medical treatment off site (eg, patients who present at local emergency departments,
walk-in clinics, and pharmacies).
In addition, researchers must consider the effects of temporality and geography on populations
attending and supporting MGs. For example, researchers looking at MGs from a public health perspective
may study patient populations for a longer period of time (eg, days and weeks after the events), as is the case
in infectious-disease outbreaks. In the case of infectious disease, due to international travel, the population of
interest may extend far beyond the population of the host community, crossing international boundaries.
Researchers are encouraged to indicate clearly how these considerations are either incorporated into their
study or why they were excluded. Excluding some patient presentations will underestimate the total health
burden of a MG.
Determining the Event Population (NEV)—Defining the event population (NEv) is integral to determining PPR. This
value may also be used in classifying the size or potential risks of an MG event. In its simplest terms, the event
population consists of the attendees to a MG event. Depending on the type of event, the event population
may consist of spectators (most MG events), participants (athletic and/or cultural events), and/or crew
(broadly conceived as those volunteers, contractors, media, and workers who are involved in staging and
supporting the event). Researchers should ensure that their definitions of the NEV address the inclusion or
exclusion of these groups.
Determining consistent numbers for NEv is challenging. For example, at a music festival, the media may report
ticket sales as a proxy for event attendees, which fails to include those who are provided complimentary
access by promoters. Gate numbers may be reported as a proxy for the total population, however, at events
with ‘‘in-and-out’’ privileges, or variable attractions through the day, the actual number of persons on site
may vary (ie, maximum number during headliner attraction, but fewer earlier in day).
Defining the Host Community (NHC)—the host community is most commonly thought of in relation to the
geographic or municipal jurisdiction in which the MG event is held. The variable NHC captures the size of the
host community, or in the case of a multi-jurisdictional event, such as a marathon, host communities.
Understanding and defining the size of the host community is most valuable in evaluating the risks and resources
associated with hosting a particular event. If an event will temporarily quadruple the number of people in a
given community, this will create a burden for local infrastructure in relation to health services, policing,
firefighting, traffic management, and other services. Challenges arise in accurately determining the size of the
host community. The availability of current population statistics may vary. Determining the boundaries of
the host community may require that a range be used (eg, for an event held in an urban setting, should the
population include the entire city or just population in the section of the city through which the event will
pass). Specifically reporting both numbers, or providing a detailed range, would promote consistency and allow
researchers to compare patient populations and PPRs across events.
8

Limitations
The population model presented in this report results from discussions by the international collaborative group.
This report aims to elicit further discussion and developmental input from other experts working in the MGH
field. The models have not yet been tested prospectively in a real-world setting.
Conclusions
The objective of the team assembled in Australia in 2013 was to initiate a consensus process to develop an
internationally-accepted minimum dataset for use in MGH research. The present lack of agreement on the
definitions and classification of MGs makes this work challenging. This report proposes a population model with
definitions and conceptual categories underlying MGH research and operational health planning for MGs. Clarity
in definitions and descriptions of host community, event, and patient populations in the context of MGs will
permit consistent description and reporting in the international literature. Importantly, it will also permit a better
understanding of injury and illness presentations and clarify the health service impacts on the various populations
of interest. Operational planning for emergency response, health promotion, injury and illness prevention, and
surveillance will be measurable against more precisely-defined populations.
Research in Context
Literature Review
The writing team consisted of an international-collaborative group with substantial experience with clinical,
research, and policy development in the context of MGs. Review of the literature and addressing gaps in theory
building identified in published reviews of the literature prompted the consensus process.
Interpretation
This manuscript puts forward theory-building concepts and models that may stimulate further discussion
and consensus building amongst MG researchers in the international community.
Acknowledgements
The authors would like to acknowledge: Dr. Greg Anderson, PhD, Dean, Office of Applied Research, Justice
Institute of British Columbia, New Westminster, BC, Canada; Dr. Katherine Arbuthnott, WHO and Public Health
England; and Dr. Mark Salter, Public Health England.
Authors’ Contributions
Paul Arbon participated in the discussions of the population model at the Flinders University meeting and
on team conference calls discussing the model and the publication; he also commented on the various drafts
of the document. Ron Bowles contributed significantly to the theoretical model, and, in conjunction with the
principle author, expanded the paper into two parts in order to clarify the population of events vs the
populations of people affected by MGs. Dr. Bowles also contributed the formatted figures to the submitted
version of the manuscript. Alison Hutton contributed to the initial draft of this paper; she was also part of the
panel that gave feedback to the initial diagram concepts for this paper. Adam Lund was the principle author on
this paper; Dr. Lund created the initial drafts and the diagram concepts and overviewed the team process in
bringing the paper to completion, including final submission of all document elements to the journal. Jamie
Ranse contributed to the development of the population model at the Flinders University meeting and on team
conference calls; he also contributed to the initial draft of this paper. Malinda Steenkamp participated in
discussions of the population model at the Flinders University meeting and on team conference calls discussing
the model and the publication; she commented on the various drafts of the document. Sheila Turris cow rote
this paper with the principle author; she was also a member of the panel that created the original drafts of the
figures and tables.

9

References
1. Arbon P. The development of conceptual models for mass-gathering health. Prehosp Disaster Med. 2004;
19(2):212-217.
2. Arbon P, Cusack L, Verdonk N. Mass gathering public health and emergency medicine literature
review: levels of evidence. Australasian J of Paramed. 2013; 10(1):1-5.
3. Tam JS, Barbeschi M, Shapovalova N, Briand S, Memish ZA, Kieny MP. Research agenda for mass
gathering: a call to action. Lancet. 2012; 12(2):231-239.
4. Ranse J, Hutton A. Minimum data set for mass gathering health research and evaluation: a
discussion paper. Prehosp Disaster Med. 2012; 27(6):1-8.
5. Steffen R, Bouchama A, Johansson A, et al. Non-communicable health risks during mass gatherings.
Lancet. 2012; 12(2):142-149.
6. Turris SA, Lund A. Minimum data set for mass-gatherings health research and evaluation: a response.
Prehosp Disaster Med. 2013; 28(2):191-193.
7. Abubakar I, Gautret P, Brumnette GW, et al. Global perspectives for prevention of infectious diseases
associated with mass gatherings. Lancet. 2012; 12(2):66-74.
8. CBC News. Popular Pemberton Music Festival Cancelled. http://www.cbc.ca/news/canada/britishcolumbia/popular-pemberton-music-festival-cancelled-for-2009- 1.820216. Accessed March 16, 2014.
9. Medical Risk Classification for Mass Gatherings. Delivering a Healthy WA. http://
www.public.health.wa.gov.au/cproot/2542/2/Medical%20Risk%20Classification.pdf. Accessed March 16,
2014.
10. World Health Organization. 2008. Communicable disease alert and response: key considerations.
http://www.who.int/csr/Mass_gatherings2.pdf. Accessed March 16, 2014.
11. Lund A, Turris SA, Bowles RR. Conceptualizing the impact of special events on community health
service levels: an operational analysis. Prehosp Disaster Med. 2014; 29(5):525-532.
12. Welch TR. Emergency department volumes and ‘‘mass events.’’ Journal of Pediatrics. 2014; 164(2):223225.
13. Botelho-Nevers E, Gautret P. Outbreaks associated to large open air festivals, including music festivals,
1980-2012. Eurosurveill. 2013; 18(11):1-10.
14. Molloy MS, Brady F, Maleady K. Impact of a single large mass gathering music event, from a series of
such events, on a receiving hospital’s emergency department (ED). Prehosp Disaster Med. 2013;
28(1):S186.
15. Heiby MJ, Barnhardt W, Berry T, Welcher M, Brady WJ. The impact of a mass gathering events with an
on-site medical management team on municipal 911 emergency medical services. Am J Emerg Med.
2013; 31(1):256-257.
16. Zielioski A, Pawlak BJ. Toolbox for implementation of surveillance at mass gatherings: surveillance
during mass gatherings. 2013. http://www.rki.de/EN/Content/
Prevention/React/Work/wp4/WP_4_ToolBox.pdf? blob5publicationFile. Accessed March 16, 2014.
17. Arbon P, Bridgewater FHG, Smith C. Mass gathering medicine: a predictive model for patient presentation
and transport rates. Prehosp Disaster Med. 2001; 16(3):150-158.

10