Wetterhall SF, Coulombier DM, Herndon JM, Zaza S, Cantwell JD, for the Centers for Disease Control and Prevention Olympics Surveillance
Unit . Medical Care Delivery at the 1996 Olympic Games. JAMA. 1998;279(18):1463-1468. doi:10.1001/jama.279.18.1463
From the Office of Program Planning and Evaluation (Dr Wetterhall) and the Epidemiology Program Office (Ms Herndon and Dr Zaza), Centers for Disease Control and Prevention, Atlanta, Ga; and the Unité Systèmes d'Information et de Communication, Réseau National de Santé Publique, Saint-Maurice, France (Dr Coulombier). Dr Cantwell was the Chief Medical Officer for the 1996 Olympic Games, Atlanta.
Context.— Mass gatherings like the 1996 Olympic Games require medical services
for large populations assembled under unusual circumstances.
Objective.— To examine delivery of medical services and to provide data for planning
Design.— Observational cohort study, with review of medical records at Olympics
Setting.— One large multipurpose clinic and 128 medical aid stations operating
at Olympics-sponsored sites in the vicinity of Atlanta, Ga.
Participants.— A total of 10 715 patients, including 1804 athletes, 890 officials,
480 Olympic dignitaries, 3280 volunteers, 3482 spectators, and 779 others
who received medical care from a physician at an Olympic medical station.
Main Outcome Measures.— Number of injuries and cases of heat-related illness among participant
categories, medical use rates among participants with official Games credentials,
and use rates per 10000 persons attending athletic competitions.
Results.— Injuries, accounting for 35% of all medical visits, were more common
among athletes (51.9% of their visits, P<.001)
than among other groups. Injuries accounted for 31.4% of all other groups
combined. Spectators and volunteers accounted for most (88.9%, P<.001) of the 1059 visits for heat-related illness. The rates for
number of medical encounters treated by a physician were highest for athletes
(16.2 per 100 persons, P<.001) and lowest for
volunteers (2.0 per 100). Overall physician treatment rate was 4.2 per 10000
in attendance (range, 1.6-30.1 per 10000). A total of 432 patients were transferred
Conclusions.— Organizers used these data during the Games to monitor the health of
participants and to redirect medical and other resources to areas of increased
need. These data should be useful for planning medical services for future
MASS GATHERINGS require the provision of medical services for large
populations who have assembled under unusual circumstances. Mass gatherings,
including scheduled events in sports facilities, air shows, rock concerts,
outdoor celebrations, and visits by dignitaries, vary in their complexity
and demand for medical services.1- 5
The 1996 Centennial Olympic Games was a mass gathering that posed unique challenges
for ensuring the public health and medical safety needs of its participants.6 During the 17 days of the Olympics, an estimated 5
million people (including visitors and residents of metropolitan Atlanta)
from 197 different countries gathered in a confined area under subtropical
environmental conditions.7 The duration, size,
and location of the gathering, likelihood of hot weather, and concerns about
terrorism influenced the planning of medical services for participants.8- 11 The
official organizer, the Atlanta Committee for the Olympic Games (ACOG), was
responsible for providing medical services at all Olympic residential, training,
and competition sites. The Centers for Disease Control and Prevention (CDC)
developed a health surveillance system that monitored daily the health and
safety of Olympics participants. In this article we examine medical care delivery
during the 1996 Olympics and provide data that should be useful for planning
future Summer Olympics and other mass gatherings.
We categorized persons on the basis of accreditation status, the system
of classification used by ACOG in issuing identification passes. Because of
concerns about security, all persons who were working or otherwise participating
in the Olympics (other than spectators) were required to wear identification
Athletes were persons engaged in competition; officials included team
officials (eg, coaches and trainers), referees, and judges; the Olympic family
comprised members (and their dependents) of the International Olympic Committee
and the National Olympic Committees and other dignitaries; volunteers included
full-time and part-time staff of ACOG; and spectators were persons without
Medical care for the 1996 Olympics was coordinated by the Medical Services
Department of ACOG. There were 3346 volunteers who provided medical care.12 Volunteers included 664 physicians, 474 nurses, 246
paramedics, 242 emergency medical technicians, 411 American Red Cross first
responders, 796 trainers, 128 physical therapists and massage therapists,
30 dentists, 20 podiatrists, and 335 administrators. Medical services were
responsible for providing first aid and emergency care for all athletes, spectators,
staff, and volunteers who sought care at any Olympic site. All persons were
required to provide written consent before receiving medical care. Sites included
the polyclinic (located in the Olympic Village), 24 competition venues (eg,
the Olympic Stadium) and 11 noncompetition venues (eg, Centennial Olympic
Park and the main press center).
Beginning July 6, the polyclinic, a 2250-m2
facility located in the Student Health Center of Georgia Institute of Technology,
Atlanta, provided medical care for athletes and the Olympic family residing
in the Olympic Village. The polyclinic was open 24 hours a day, provided primary
care and emergency medical care, and was equipped with an emergency assessment
area, an observation unit, and 8 examination rooms. Other clinical services
with capacity for minor surgical procedures were available in orthopedics,
otolaryngology, ophthalmology, gynecology, psychiatry, and oral health. Support
services included radiography (with teleradiography capacity), diagnostic
ultrasound, mobile magnetic resonance imaging (MRI), pharmacy, and clinical
laboratory. Although the polyclinic was available to all athletes, teams from
some countries had their own physicians and other medical personnel. These
medical teams were provided separate space within their residential areas.
At each of the sites of athletic competition, medical care was organized
into sports medicine stations and spectator care stations. Athletes received
care during competition at sports medicine stations, which were located near
the competition site and were staffed by at least 1 physician, 1 nurse, and
1 trainer. Athletes also could seek care (eg, massage therapy and athletic
trainers) at a sports medicine station located in the Olympic Village.
Eighty-four spectator care stations operated at the 24 competition venues.
Each venue supported 1 to 3 stations. At the Olympic Stadium, the largest
venue, there were 8 spectator care stations. At these stations, physicians
and other medical personnel provided care for spectators, staff, and volunteers.
Each care station, which opened 2 hours before each day's competition began,
was equipped to respond to cardiac arrests and to provide care for minor and
less urgent medical problems. In addition to the spectator care stations,
mobile aid teams (comprising 1 emergency medical technician and 1 paramedic)
at each Olympic site circulated among spectators within the venue and provided
minor aid (eg, dispensed aspirin or sunscreen). There was 1 mobile aid team
for each 20000 spectators.
Spectators and others requiring nonemergency care beyond that available
at the venue were referred to an outpatient clinic or designated referral
hospital. For emergencies, the patient was transported by ambulance to the
closest hospital. Emergency medical services and transportation plans were
developed to provide service at each venue. At least 1 ambulance was assigned
for every 20000 spectators expected. As soon as an ambulance crew began to
provide care for a patient, a replacement ambulance was dispatched to that
Nine hospitals in the metropolitan Atlanta area and 8 in outlying areas
formed a network of Olympic hospitals that provided care to patients who were
referred from ACOG medical facilities. In the Atlanta area, each venue had
a designated hospital to which patients requiring emergency treatment were
The cost of establishing and operating the medical care system was $4.36
million (polyclinic, $1.46 million; sports medicine and spectator care stations,
$2.9 million). Because the building that housed the polyclinic had recently
undergone extensive renovation, physical refurbishments for the Olympics were
limited to creating 4 dental assessment and treatment rooms. Sources of funds
included $450000 from ACOG's operating budget and approximately $4 million
value-in-kind donated products, including $600000 for major equipment (eg,
defibrillators, hemodynamic monitoring devices, use of MRI). No tax dollars
were used to support medical service delivery. Nearly all of the equipment
and unused products were returned to the original donor source (eg, hospitals,
equipment manufacturers). At the end of the Games, some of the monitors and
defibrillators were donated to community groups.
Medical care at the polyclinic, sports medicine stations, and spectator
care stations was provided without charge to the patient. For additional medical
care, ACOG, through an insurance carrier, provided coverage for medical expenses
(eg, for emergency illness and unintentional injury) to athletes, officials,
and members of the Olympic family. Spectators requiring additional care beyond
the spectator care stations assumed responsibility for those medical expenses.
A health information system was developed to monitor the health status
of athletes, staff, and spectators; to document service delivery; and to identify
unusual patterns of illness or injury requiring further investigation. Reports
summarizing the previous 24 hours were submitted each evening to the Medical
Commission of the International Olympic Committee. The 2 sources of information
for these reports were logbook entries (completed for all patients seeking
care) and medical encounter forms (completed for patients who were evaluated
and treated by a physician).
The mobile aid teams recorded in a logbook the name and the nature of
the treatment of each person who received care. For any athlete seeking care
at a sports medicine station, an entry was listed in a clinic logbook. The
name of any person seeking treatment at a spectator care station was listed
in a medical services logbook. For a minor condition (eg, headache) that could
be treated by nursing or other staff, the listing in the medical services
logbook was the only documentation of the visit to the aid station. For persons
with more substantial problems that were treated by a physician, medical information
was recorded on a standard encounter form.
The medical encounter form was a legal-size, single-page form that was
designed and pretested at 4 world-class sporting events held in the Atlanta
area during 1995 and 1996. Medical encounter forms were used in sports medicine
stations and spectator care stations. The practitioner used the upper portion
of the form to record written notes. A listing of clinical diagnoses (ie,
primary reason for visit) that the practitioner could check was printed on
the lower portion of the form.
Every 2 hours, medical clerks from each Olympic venue faxed all logsheets
(from mobile aid teams, sports medicine stations, and spectator care stations)
and all completed medical encounter forms to a central location. Data from
the logsheets and medical encounter forms were entered into a computer, results
were analyzed, and daily reports were generated.
We used data from the logbooks and medical encounter forms to determine
medical service use. Logbook data were used to calculate overall service delivery
(ie, number of persons seeking any treatment). Data regarding patient characteristics
(eg, age, sex, accreditation status) and diagnoses were available only from
the subset of patients who had medical encounter forms. Data were available
from July 6, when the Olympic Village and the polyclinic opened, through August
4, the end of the Olympics. Unless otherwise specified, we did not separate
data for the preparation period (July 6-18) from data collected during the
official dates of the Olympics (July 19-August 4).
To calculate physician treatment rates for persons who had been accredited,
we obtained denominator data supplied by ACOG from their accreditation database.
The accreditation database was a computerized listing of the names and limited
demographic data (eg, age and sex) of all persons who were issued identification
badges. Rates (ie, cumulative incidence) were calculated as the number of
medical encounters (treated by a physician and generating a medical encounter
form) per 100 persons in a given accreditation category. Dental and routine
eye examinations were excluded from analysis.
To calculate use rates at competition venues, we used ticket attendance
data supplied by ACOG. Analysis was restricted to the 16 venues located in
the greater Atlanta metropolitan area for which complete medical records were
available for inspection at the end of the Games. Medical care use rates were
calculated by determining the number of logbook entries at spectator care
stations, divided by total number of attendees. Physician treatment rates
were calculated by determining the number of patients (excluding athletes)
with medical encounter forms completed by a physician divided by the total
number of attendees.
We analyzed data using Epi Info Version 6.04 (CDC, Atlanta) software,
comparing means using analysis of variance and proportions using χ2 statistics.
From July 6 through August 4, the ACOG Medical Services Department provided
care for 44142 persons whose medical encounters were recorded in logbooks.
Visits to spectator care stations accounted for 33643 (76.2%) encounters,
mobile aid teams provided care to 3451 (7.8%) persons, and visits to sports
medicine stations accounted for 7048 (16.%) encounters. From July 6 through
July 18, the daily number of encounters increased gradually (Figure 1), and after the opening ceremony, the number increased
rapidly. From July 19 through August 4, spectator care stations averaged 2363
daily encounters; sports medicine stations, 311; and mobile aid teams, 178.
From July 6 through August 4, a total of 10715 persons who visited sports
medicine or spectator care stations were treated by a physician and had a
medical encounter form completed. The temporal pattern for medical encounters
treated by physicians paralleled overall medical service use. Most patients
(n=9134, 85.2%) were treated from July 19 through August 4, when the daily
average was 537 patients (daily range, 233-676).
Among patients treated by physicians, spectators accounted for the largest
number of visits (n=3482, 32.5%), followed by volunteers (n=3280, 30.6%),
athletes (n=1804, 16.8%), and other groups (Table 1). Repeat visits accounted for 893 (8.3%) of the total. Repeat
visits were more common for the Olympic family, officials, and athletes; such
visits were infrequent for spectators.
Characteristics of persons receiving medical care varied by accreditation
status. Overall mean age was 33.9 years; the mean age of athletes was 26.0
years (P<.001), whereas the mean age of the media,
Olympic family, and officials was older (42.1 years, P<.001).
Among all patients, there were more males (53.4%) than females. Within accreditation
subgroups, athletes, officials, the Olympic family, and the media were predominately
male (P<.001). Only among volunteers and spectators
was the majority of persons who received care female.
Regarding specific diagnoses, sprain/strain was the most common reason
for a visit (n=1450, 13%), followed by upper respiratory tract infection (n=922,
9%), heat cramps/dehydration (n=801, 8%), contusion/abrasion (n=777, 7%),
other injury (n=520, 5%), laceration (n=487, 4%), and nausea/vomiting (n=387,
The distribution of reasons for receiving medical care varied by accreditation
status (Table 1). Injuries predominated
among athletes, accounting for 51.9% of their visits (P<.001). Among members of the Olympic family, upper respiratory tract
infection was the most frequently listed reason for the visit, and 2 chronic
conditions, diabetes and hypertension, were among the 10 most common reasons
for physician care.
Heat-related illness was the most common condition among spectators
(n=752, 21.6%). Spectators and volunteers accounted for most (88.9%) of the
visits for heat-related illness (P<.001). Heat-related
illness was rare among officials, members of the Olympic family, and the media.
Most of the patients treated by physicians were released, either to
their own physicians (n=9275) or for follow-up in another ACOG medical facility
(n=808). A few patients (n=63, 0.6%) left before being seen or against medical
advice. A total of 569 patients were transferred to the polyclinic, another
outpatient facility, or a hospital. Of the 432 patients transferred to hospitals,
271 required emergency (ie, ambulance) transport. Most emergency transports
involved spectators (n=155, 57.2%) or volunteers (n=61, 22.5%). Diagnoses
among the 271 persons who required emergency transport included heat-related
illness (n=27, 10%), injuries (n=62, 22.9%), and other medical illness (n=182,
67.2%). Of those persons with medical illness, 98 received emergency transport
for cardiac disorders (ie, diagnoses of chest pain, ischemic heart disease,
or hypertension). Three patients sustained cardiac arrest; 2 were successfully
resuscitated and transferred to a hospital and 1 died. Forty-one patients
were transported via ambulance for gastrointestinal tract disorders (ie, abdominal
Patient outcome data were available for the 3 major Atlanta hospitals
responsible for providing care to athletes, the Olympic family, the media,
residents of the Olympic Village, and 10 large competition venues (eg, Olympic
stadium). Transfers to these hospitals accounted for 183 (67.6%) of all emergency
Of the 101 persons admitted to these hospitals, 23 were admitted for
chest pain, 8 of whom had confirmed myocardial infarctions, and 2 of whom
had subsequent coronary angioplasty. Two underwent coronary artery bypass
graft surgery. Other reasons for hospitalization include injuries (n=19, mostly
athletes), gastrointestinal tract disorders (n=11, eg, 5 with upper gastrointestinal
tract bleeding), renal disorders (n=8, eg, 3 with renal stones), infectious
diseases (n=7, eg, 2 with malaria), pulmonary conditions (n=7, eg, 3 with
pulmonary embolus), and central nervous system disorders (n=5, eg, 1 with
seizures). Two persons were admitted for heat exhaustion. There were no deaths
among persons admitted to hospitals.
The clinical services and diagnostic tests provided at the polyclinic
are summarized in Table 2. Most
of the MRI and ultrasound tests were for musculoskeletal conditions. Use of
these tests increased steadily during the Games.
A total of 178367 persons were accredited to participate in the Olympics
(Table 3). The majority of accredited
persons were volunteers (n=136660, 76.6%). The overall medical use rate (based
on data from the 6690 persons for whom complete information on age and sex
were available) was 3.8 encounters per 100 persons. Use rates varied markedly
by accreditation status. Unadjusted use rates were highest for athletes (16.2
per 100, P<.001) and lowest for volunteers (2.0
per 100). Use rates did not vary by sex, except for volunteers, for whom the
rate was higher for females than for males (2.6 vs 1.6 medical encounters
per 100 accredited persons). Use rates adjusted for age and sex were similar
to unadjusted rates, except among members of the Olympic family. Adjustment
for age and sex lowered the Olympic family use rates from 7.7 to 5.8 per 100.
Use rates (ie, number of medical encounters treated by a physician per
100 accredited persons) varied by age and displayed distinct patterns for
athletes compared with other accredited persons (Table 3). For athletes, use rates gradually increase with age, except
for a decline among persons aged 50 to 59 years (which is based on more limited
data). For officials, the rate declined with increasing age. For the Olympic
family and volunteers, the distribution of use rates is U-shaped, with higher
rates among younger and older persons and the lowest rates among persons aged
20 to 39 years. Rates among members of the media did not exhibit a pattern.
From July 19 through August 4, the total number of patients (recorded
in spectator care logbooks) treated at competition venues ranged from 75 at
Clark Atlanta University (a small stadium where field hockey was played) to
4840 at the Olympic Stadium (site of the track and field competition and the
opening and closing ceremonies) (Table 4). The overall use rate was 22.9 patients per 10000 in attendance
(range, 8.4-130 per 10000). The number of patient visits that required treatment
by a physician at each venue ranged from 14 to 742. The overall physician
treatment rate for this period was 4.2 patients per 10000 in attendance (range,
1.6-30.1 per 10000).
The 1996 Summer Olympics was a mass gathering with unique characteristics
that created complex demands on medical service delivery. These characteristics
include the large number of participants and spectators (the largest peacetime
gathering in history), long duration (31 days, beginning with the opening
of the Olympic Village), and areas of high population density (with 12 of
the 24 venues located within a 5.2-km-diameter circle in downtown Atlanta).
The Olympics medical services had to be fully coordinated with emergency medical
services, disaster planners, and local, state, and federal health officials.
Planning for medical services at the Olympics began in 1991.13 Because of Atlanta's climate, heat-related illness
was expected to pose substantial risk. Injuries were expected to produce a
major proportion of medical visits.14,15
Outbreaks of infectious diseases, such as common-source foodborne outbreaks,
were possible, as were person-to-person transmission of disease among athletes
and airborne transmission of diseases (eg, measles) within the confines of
indoor venues.16,17 Finally, planning
efforts also recognized that the highly publicized Olympics might attract
a terrorist attack with a biological or chemical weapon.11
In contrast with previous studies of medical care at Olympics held in
North America, we analyzed data by accreditation status of the participants.14,15,18 We found differences
in the types of medical problems and use rates across categories of participants.
The variability in use probably reflects the different types of activities
and exposures that members of each group experienced. For example, volunteers
and spectators experienced most of the heat-related illness, and generally
relied on walking or using public transportation, whereas members of the other
groups usually were provided with special transportation. For the Olympic
family and spectators, the higher use rates among younger and older persons
reflect a pattern commonly seen in rates for visits to outpatient and emergency
Three cardiac arrests occurred at the venues, and most emergency transports
from venues to hospitals were for cardiac conditions. These findings underscore
the importance of having advanced cardiac life support capability within each
venue and of coordinating ACOG medical services with experienced emergency
medical services, transportation systems, and local hospitals. The bombing
incident in Centennial Olympic Park occurred in the early morning of July
27, after the spectator care station had closed for the night. The number
of casualties was so large that local first responders (ie, firefighters and
police) immediately assumed responsibility, as planned, for evacuating injured
persons to local hospitals. Our study does not include data on immediate treatment
for bombing victims.
Operation of the ACOG medical care system was fully integrated with
the activities of public health and other government officials. During the
Olympics, CDC operated the health information system that generated daily
reports (in English and French) and summarized medical activity for review
by the International Olympic Committee. Local, state, and federal public health
officials reviewed daily medical service data.
The primary objective of our system was to monitor the health and safety
of athletes, staff, and spectators who sought care at any Olympic medical
site. The system was not designed to monitor patterns of use of existing medical
facilities such as emergency departments. During the Olympics, the Georgia
Department of Human Resources monitored emergency department encounters in
8 sentinel hospitals. Overall use of these facilities did not markedly change
from the baseline period preceding the Games, and the number of emergency
department visits for heat-related illness actually decreased during the Olympic
period.20 A decrease in the number of visits
to hospital emergency facilities during the Olympics has been previously described.15
During the Games, ACOG used medical service data to redirect resources
(eg, to assign additional medical personnel to the spectator care stations
in Centennial Olympic Park, a popular gathering spot), to identify persons
whose diagnoses (eg, bloody diarrhea or jaundice) had potential public health
impact, and to correct environmental hazards at the venues.
The 1984 Olympic Games in Los Angeles, Calif, was the most recent summer
Olympics held in North America. Compared with those Games,14
the 1996 Summer Olympics were considerably larger, with nearly 8 million spectator
tickets, double that in Los Angeles. Data from logbooks show that although
the total number of persons treated was 44142 in Atlanta and 5516 in Los Angeles,
total use rates were comparable in Atlanta (22.9 visits per 10000 in attendance)
and Los Angeles (16 per 10000 in attendance). Moreover, the proportion of
total patient visits requiring treatment by a physician was comparable (26.3%
in Atlanta vs 29% in Los Angeles).
We calculated use rates using 2 different methods. Among accredited
persons, we calculated the number of medical encounters as a proportion (ie,
cumulative incidence) of the number of persons within a given accreditation
category. Thus, in planning similar events, the anticipated number of persons
within a given category (eg, number of athletes) can be used to calculate
expected number of medical encounters. For use rates at venues, we used total
attendance in the denominator. Organizers, knowing the expected attendance
at their venue, can use these data to predict patient loads.
Actual use of ACOG physicians and medical services, particularly for
athletes, can be influenced by several factors. The presence of team physicians
from other countries decreases the demand for services from ACOG, although
US specialists who were volunteering at the polyclinic were frequently sought
for consultations. Furthermore, the rules of each sports federation that governed
a particular sport often determined whether an ACOG volunteer physician or
a physician working for the federation would provide initial care to an athlete
during competition. Finally, the athletes commonly seek the services provided
in the polyclinic, such as dental and ophthalmic care and diagnostic imaging
tests, that may be less readily available in their own countries.
Several potential limitations in these data should be noted. With the
large number (n=664) and high turnover of volunteer physicians involved in
the Olympics, the recording of diagnoses may not have been standard across
sites. Because our interest is in aggregated categories (eg, injuries or heat-related
illness), however, such variability should not substantially affect the results.
Incomplete reporting of medical encounters can produce biased results.
Data on age and sex were missing on 5.2% and 2.9% of forms, respectively.
We encouraged timely and complete reporting of data from the venues by training
clerical and professional staff, conducting frequent site visits, and providing
daily feedback reports. Thus, although underreporting was an ongoing concern,
this potential bias did not substantially influence our summary results. At
the completion of the Centennial Olympic Games, original records were received
from nearly all venues and were checked for completeness. Reporting was probably
less complete from the 4 soccer venues located in other states (and 4 venues
in the Atlanta metropolitan area), from which final medical records were not
available to verify completeness. Because reports from these sites represent
a small proportion of the number of visits, such underreporting is unlikely
to have substantially influenced our results.
In planning for mass gatherings, medical personnel and other organizers
need to identify comparable events and experiences that can provide the best
guidance for estimating the number of patients, types of medical problems,
and estimated use rates. The medical service data from the 1996 Olympics provide
valuable information for planning upcoming mass gatherings (eg, the 1998 Goodwill
Games in New York, NY and the Summer Olympics in Sydney, Australia, in 2000
and Athens, Greece, in 2004). The surveillance methods that we developed and
implemented to monitor medical care delivery for the Games illustrate the
evolving challenges and benefits of establishing flexible and useful health
information systems. These systems must satisfy multiple needs to operate
effectively at the complex interface between the practices of clinical medicine
and the assurances of public health.