Statistics from Altmetric.com
Team physicians travelling with elite athletes to international competitions are not only responsible for the prevention and treatment of injuries, but also the protection of the health of the athletes. Data indicate that up to 50% of all medical consultations in these settings are for illness in athletes.1 ,2 However, in comparison to injury, illness has been less commonly studied during major competitions. Illness surveillance studies have only recently been conducted during major international competitions such as the 2009 IAAF World Athletic championship,3 the 2009 FINA (aquatic) World championships,4 the 2009 Confederations Cup Football tournament,5 the 2010 Vancouver Winter Olympic Games,6 the 2010 Super 14 Rugby Union tournament,7 ,8 the 2010 FIFA World Cup9 the 2010 Asian Beach Games10 and most recently, the 2012 Youth Olympic Games.11
There remains no firm consensus on research methodology for the monitoring and reporting of patterns of medical conditions and illness during sports tournaments. In most studies to date, data have been reported as the proportion (%) of athletes affected by illness (incidence proportion, IP) during tournaments, which were often varying durations.3 ,4 ,6 ,10 ,11 These studies show an IP of illness during these sports tournaments ranging from 6.7% to 12.1%. Although there is consistency in the manner in which the IP of illness has been reported in these studies, the results cannot be compared across studies where the duration of the tournament varies. This makes it difficult to plan and test strategies to prevent illness in athletes during tournaments.12
In contrast, there have been a few studies where the illness rates have been reported as an IR (illness per 1000 athlete-days). This methodology, which takes into account athlete exposure days, allows for comparison between tournaments of different durations, different sports and other factors that may affect illness. In a few studies, the IR (per 1000 athlete-days) for single-code sports have been reported as follows: World Cup football—7.79; Confederations Cup football—16.95 and Rugby Union—20.7.8 This methodology has also been used successfully to identify factors that are associated with illness. In particular, rugby players had an increased IR of illness (32.6/1000 athlete-days) following travel across multiple time zones when compared with competing in their home country (15.4/1000 athlete-days).7 However, the IR of illness has not been reported in large multicoded competitions in able-bodied athletes, and IR of illness as well as factors that may be associated with increased illness rates during the Paralympic Games have never been studied.
Therefore, the main aim of this study was to document the incidence of illness in athletes who participated in the London 2012 Paralympic Games. Furthermore, we wanted to determine if factors such as the competition period, sport type, gender and age were associated with increased rates of illness in these athletes. These data would also form the basis for determining the efficacy of any future intervention strategies to reduce illness during this competition.
Type of study
This study was a component of a large prospective cohort study on the epidemiology of injury and illness conducted over a 14-day period during the London 2012 Paralympic Games. The general methodology used in this study has been described.13 In this section, the general methodology is summarised, and a more detailed description of the methods used for the analysis of the illness data is provided.
A total of 3565 athletes from 160 of 164 countries (97.6% of all countries; 85.3% of all athletes) participated in this 14-day study. Before the onset of the Games, research ethics approval for the study was obtained from the University of Brighton (FREGS/ES/12/11) and the University of Cape Town Health Sciences Research Ethics Committee (HREC/REF 436/2012). Consent to utilise their deidentified medical data for research purposes were obtained from all athletes prior to the Games.
Data collection was carried out on a daily basis during the 3 days before the start of the Games (precompetition period) and ended on the last day of the 11-day Games (competition period). For this prospective study, three data sources were utilised. A comprehensive database was obtained from the International Paralympic Committee (IPC), the IPC athlete database. This database contained the following important data fields in a deidentified format: accreditation number, country code, sports code, gender and age. The second data source was from an electronic medical data capture system (EMDCS) (ATOS, France) where Local Organizing Committee (LOCOG) sports physicians and medical staff were requested to enter all illness encounters (date, athletes accreditation number and the main system affected by illness) at both the Paralympic Village polyclinic and at the sports venues, wherever the athlete reported for care.
The third data source was from medical staff that provided medical care to their own teams (78 delegations). Chief Medical Officers or their designated staff entered daily illness encounter data on a novel web-based electronic injury and illness capturing system (WEB-IISS) that was specifically developed for this purpose.13 In addition to the data on illness that was available on the EMDCS, the WEB-ISS allowed us to capture more clinical details, including details that specifically apply to Paralympic athletes, as well as the capturing of accurate exposure data (athlete-days). For the purposes of this manuscript, only data related to the main system that was affected by the illness can be reported, as this was the only clinical information on the EMDCS.
A medical illness was defined as ‘any newly acquired illness as well as exacerbations of pre-existing illness that occurred during training or competition, and during or immediately before the London 2012 Paralympic Games’. Illness data from teams without their own medical support (n=82 countries; 236 athletes) were collected only through the EMDCS. An assumption was made that all illness encounters in athletes from these teams would be reported by the polyclinic and venue medical staff using the fields described with the EMDCS.
Illness data from the EMDCS and WEB-IISS were exported in separate Excel spread sheets. The date on which medical staff registered the illness or injury was used as the date of occurrence of the illness. A credible algorithm was developed to identify and delete duplicate records in the data. Records having the same dates and same clinical characteristics were considered duplicates. The required definitions were made for the variables to be analysed. Although medical staff was requested to report all illness in their athletes on the WEB-IISS, there were instances where athletes reported to the LOCOG services directly, underwent further investigation using the LOCOG medical services or had second opinions from LOCOG medical staff and thus were reported on both systems or only on the EMDCS. In the case of duplicate illness encounters (defined as recorded on the same date with same clinical characteristics), only the data from the WEB-IISS were used. Data from the two systems (EMDCS and WEB-IISS) were then combined, identifying and aligning the variables that occurred in both. Therefore, data were collected over the 14-day study period (3-day precompetition period, 11-day competition period) in 3565 athletes from 160 countries.
Calculation of the IR and IP
The details of the calculation of athlete-days have been described.13 An analysis of the data on teams with their own medical support (WEB-ISS) showed that there was a negligible variation between reported number of athletes in each delegation and the total number of athletes, as published in the IPC athlete database.13 Therefore, total athlete-days for each country were calculated as follows: total team days (period in days) × daily team size (for each day). The incidence of illness was calculated as illness per 1000 athlete-days and will be referred to as the IR.
The IP refers to the percentage of athletes reporting an illness by sport, gender or age group. IP was calculated as the number of illnesses per 100 athletes in the subgroup(s) (%). During the precompetition , the competition and the total periods, illness data were recorded for a total of 49 910 athlete-days (table 1).
Athlete participation in different sports and by gender
There were 2347 male athletes (65.8% of all athletes) and 1218 female athletes (34.2%). The number of male and female athletes participating in each sport is depicted in table 2.
Athlete participation in different age groups
The age of an athlete was taken as the age on 27 August 2012. The mean (+SD) age of the athletes in this study was 30.9±9.2 years (minimum=13 years, maximum=67 years). In order to determine the relationship between age on illness IR, athletes were divided into three age group tertiles as follows: ≤25 (n=1142), 26–34 (n=1249) and ≥35 years (n=1174).
Statistical analysis of data
Data were in the form of counts, that is, the number of illnesses each athlete contracted. An athlete could sustain an illness in either the precompetition or competition period of the games or in both. Athletes could participate in more than one sport and/or more than one event, and some athletes reported more than one illness for the same period or for a different period.
Standard descriptive statistical analyses were conducted. These include numbers, proportions/percentages (including 95% CIs) and incidences (including exact 95% CIs) of illnesses in the total sample as well as during the precompetition and competition periods, genders, age groups, sport types and affected systems.
The counts were assumed to have a Poisson distribution and were therefore modelled using generalised linear modelling. The data were overdispersed (the variance was 1.4 times the mean) and this was accounted for in the modelling. The model accounted for the correlated data in terms of the countries (clustering) and the correlation owing to repeated observations over the two periods using generalised estimating equations.
Overall incidence of illness (total, precompetition and competition periods)
During the total period, 657 illnesses were reported in 505 athletes (IP=14.2%; 95% CI 13.0 to 15.3). During the precompetition period, 156 illnesses were reported in 140 athletes (IP=3.9%; 95% CI 3.3 to 4.6). During the competition period, 501 illnesses were reported in 401 athletes (IP=10.2%; 95% CI 9.2 to 11.2).
The IR of illness in the total period was 13.2/1000 athlete-days (95% CI 12.2 to 14.2). The IR of illness in the precompetition period was 14.6/1000 athlete-days (95% CI 12.4 to 17.1), while the IR in the competition period was 12.8/1000 athlete-days (95% CI 11.7 to 13.9).
The IP of illness in the precompetition period was significantly lower (p<0.0001) than that of the competition period (3.9% vs 11.2%), but the IR of illness was not significantly different (p=0.187).
Incidence of illness by system affected
The IP and IR of illness, in each system, in the total period are depicted in Otable 3. The highest IR was in the respiratory system (3.61), followed by the skin and subcutaneous tissue (2.40), digestive system (1.90), nervous system (1.26), genitourinary system (1.12) and the ears and mastoid (0.88). If illness in the ears and mastoid are combined with illness in the respiratory system, these account for 34.1% of all the illnesses reported.
Incidence of illness in the precompetition and the competition periods
The IP and IR of illness, in each system, in the precompetition and the competition periods are depicted in table 4. The highest IR in the precompetition period was in the respiratory system (3.93), followed by the skin and subcutaneous tissue (2.71), digestive system (1.96), nervous system (1.78), genitourinary system (1.68) and the ears and mastoid (1.12).
During the competition period, the highest IR was in the respiratory system (3.52), followed by the skin and subcutaneous tissue (2.32), digestive system (1.89), nervous system (1.12), genitourinary system (0.97) and the ears and mastoid (0.82).
There was no significant difference in the IR in the competition period compared with the precompetition period in all the common systems.
Incidence of illness in different sports
The IP and IR of illness in different sports for all athletes during the total period are depicted in Otable 5. The sport with the highest IR was equestrian sports (20.7), followed by powerlifting (15.8), athletics (15.4) and table tennis (15.2). In all these sports, more than 20% of athletes were affected by illness during the total period. In only two sports, less than 10% of athletes were affected by illness: football 7-a-side (3.1%) and shooting (6.1%).
Incidence of illness by gender and in different sports
The overall IR of illness in female athletes was 14.4 (95% CI 12.6 to 16.3) and in male athletes 12.5 (95% CI 11.4 to 13.8; NS; p=0.158) and the IP of illness in all female athletes was 20.1% and in male athletes 17.6% (NS; p=0.373). The IP and IR of illness in female and male athletes participating in different sports during the total period are depicted in table 6. For females, the sport with the highest IR of illness during the total period was athletics (19.1), followed by archery (19), equestrian (19) and judo (18.3). In these sports, as well as rowing, cycling-track, wheelchair basketball, powerlifting and table tennis, more than 20% of female athletes were affected by illness during the total period. Only in goalball were less than 10% of female athletes affected by illness.
For males, the sport with the highest IR of illness was equestrian (25), followed by table tennis (16.9), powerlifting (16.8), and cycling-road (16.2). In these sports, as well as wheelchair tennis, more than 20% of male athletes were affected by illness during the total period. In four sports, less than 10% of male athletes were affected by illness: football 7-a-side (3.1%), judo (6.6%), shooting (8.7%) and archery (9.6%).
Incidence of illness in different age groups and sports
The IP and IR of illness in different age groups (13–25, 26–34 and 35–67 years) of athletes participating in different sports during the total period are depicted in table 7. The IR (95% CI) of illness in the three age groups (13–25, 26–34 and 35–67 years), respectively, was as follows: 12.4 (10.7 to 14.2), 12.8 (11.1 to 14.5) and 14.4 (12.6 to 16.3; NS; p=0.365). The overall IP of illness in the three age groups was 17.3%, 17.9% and 20.1%, respectively (NS; p=0.334). In the youngest age group (13–25 years) the sport with the highest IR of illness was boccia (17.1; 95% CI 6 to 3–37.3), followed by sitting volleyball (16.8; 95% CI 7 to 3–33), judo (16.5; 95% CI 7 to 5–31.5) and wheelchair basketball (16.2; 95% CI 7 to 8–29.9). In these sports, as well as wheelchair fencing, wheelchair rugby and athletics more than 20% of the athletes were affected by illness during the total period. In seven sports, less than 10% of athletes in the 13-year to 25-year age group were affected by illness: archery (0%), sailing (0%), shooting (0%), football 7-a-side (4.2%), cycling (track) (7.1%), cycling (road) (7.4%) and football 5-a-side (7.7%).
In the 26-year to 35-year age group, the sport with the highest IR of illness was cycling (road) (19.3; 95% CI 10 to 3–33.1), followed by cycling (track) (19.0; 95% CI 8 to 2–37.5) and powerlifting (17.1; 95% CI 9 to 7–27.7). In these sports as well as archery, boccia, sailing, swimming, wheelchair fencing, wheelchair rugby and wheelchair tennis, more than 20% of athletes were affected by illness during the period. In three sports, less than 10% of athletes in the 13-year to 25-year age group were affected by illness: shooting (0%), judo (2%) and football 7-a-side (2.2%).
In the 35-year to 67-year age group, the sport with the highest IR of illness (per 1000 athlete-days) was equestrian (27.5; 95% CI 15 to 4–45.3), followed by rowing (20.4; 95% CI 9 to 8–37.5) and table tennis (19.0; 95% CI 12 to 7–27.5). In these sports as well as athletics (24.7%), powerlifting (23.9%) and wheelchair basketball (20.6%), more than 20% of athletes were affected by illness during the period. In six sports, less than 10% of athletes in the 35-year to 67-year age group were affected by illness: football 7-a-side (0%), boccia (5.9%), wheelchair rugby (6.5%), football 5-a-side (9.1%), shooting (9.1%) and sitting volleyball (9.7%).
Incidence of illness in different sports by affected system
The IR of illness is the most commonly affected systems in the larger sports (>100 athletes per sport) during the total period is depicted in Otable 8. The highest IR of respiratory system illness was observed in cycling (road) (7.8; 95% CI 4 to 8–12.1) followed by table tennis (5.4; 95% CI 3 to 1–8.6), swimming (4.7; 95% CI 3 to 3–6.6) and athletics (4; 95% CI 3 to 0–5.2). Skin and subcutaneous illness was most common in powerlifting (3.9; 95% CI 1 to 8–7.5), followed by wheelchair basketball (3.5; 95% CI 1 to 7–6.5) and table tennis (3.2; 95% CI 1 to 5–5.8). Digestive illness was most common in swimming (2.7; 95% CI 1 to 9–3.7), followed by athletics (2.7; 95% CI 1 to 9–3.7) and powerlifting (2.6; 95% CI 1 to 0–5.7). Illness in the genitourinary tract was common in wheelchair tennis (2.7; 95% CI 0 to 7–6.9), powerlifting (2.2; 95% CI 0 to 7–5.1) and wheelchair basketball (2.1; 95% CI 0 to 8–4.6). Finally, illness in the ears and mastoid was most common in wheelchair tennis (2; 95% CI 0 to 4–5.9), followed by swimming (1.72; 95% CI 0 to 9–3) and sitting volleyball (1.4; 95% CI 0 to 3–4.1).
Independent factors associated with illness (Poisson regression model)
Covariates that were included in the Poisson regression model were gender, age groups, precompetition/competition periods and an indicator variable for the six most common sport types (>150 athletes per sport; athletics, swimming, table tennis, wheelchair basketball, road cycling, power lifting and sitting volleyball).
Gender, age (tertiles) and period (precompetition vs competition period) were not associated with an increased risk of illness in Paralympic athletes. However, participation in the sport of athletics (independent of age and gender) was associated with a higher risk of illness compared with participation in other sports (p=0.01). Some sports (equestrian and powerlifting) had a higher IR for illness than athletics, but there were too few athletes participating in these sports (<40 athletes) for the inclusion in the regression analysis.
This study is the first to report on the epidemiology of illness in the Paralympic Games. In the total study period, which consisted of the precompetition and the competition periods, the IP was 14.2%, and the overall IR was 13.2/1000 athlete-days. In this study, a number of factors that may affect the incidence of illness were also examined. The main findings of this study were that (1) there was no significant difference in the IR of illness in the precompetition period, compared with the competition period, (2) the IR of illness was significantly higher in athletics compared with other sports, (3) the IR of illness was similar in male compared with female athletes, and across age groups and (4) that the IR of illness in different systems varied between sports.
The overall IP and IR in this study, compared with that reported in other single or multicoded sports events, has already been discussed.13 In general, the IP of illness in this group of Paralympic athletes (14.2%) was higher than those reported in able-bodied athletes (6.7–12.1%).3 ,4 ,6 ,10 ,11 However, a comparison of these IP data should be carried out with caution, as exposure data were not reported in these studies and the duration of tournaments differ. In our study, we report the IR of illness and, therefore, our data can be compared with other studies where IR was also reported. The IR of illness in Paralympic athletes is higher than that reported for some,3 ,9 but not all studies.5 ,8 It should, however, be noted that these data are from able-bodied athletes participating in single sports codes (football and rugby union). In the future, we suggest a uniform approach to the reporting of incidence of illness (preferably the IR), particularly for multicoded sports events.
The documentation of illness rates in the precompetition period is important because a higher rate of illness in athletes may occur owing to factors such as: precompetitions anxiety, effects of intercontinental travel,7 easier access to medical care (own team physician or LOCOG services) and acclimation to a new environment including the exposure to new pathogens or allergens. In our study, we show that although the IP was lower in the precompetition period, the IR was not significantly higher during the precompetition period, compared with the competition period. However, our data should also be interpreted with some caution, because we could only obtain illness data for the 3-day precompetition period. In reality, most teams arrived at the athletes’ village >3 days before the start of competition, or even spent time in the host country in the weeks preceding the Games. Therefore, we recommend that a longer precompetition period be investigated in future studies by documenting illness rates.
In our study, we showed that illness rates differed between sports. In particular, in the regression model, we show that athletes in the sport of athletics (independent of age and gender) have a significantly higher risk of illness compared with Paralympic athletes participating in other sports (p=0.012). The precise reasons for the higher IR in athletics are not clear, but increased competition stress from participation in multiple events including heats, quarter, semi and finals may lead to increased risk for illness; however, this requires future study. To our knowledge, there are no other similar published studies on illness rates in this group of Paralympic athletes to which we can compare our data. However, this group is at a higher risk for illness and team physicians can target it for specific interventions to reduce this risk.
The third main finding of our study was that neither gender nor age was an independent factor associated with illness in Paralympic athletes. In able-bodied athletes participating in the Olympic Winter Games and the IAAF World Athletics Championships, significantly higher IP of illness have been reported in both female and older athletes.6 ,14 In certain aquatic sports during the 2009 FINA World Championships, higher IP of illness were also reported in women, compared with male able-bodied athletes.15 However, in each of these studies the effect of age and gender on illness was reported using IP and not IR.
Finally, in our study we describe the incidence of IR in different sports by the affected system. These data are important and highlight the fact that the risk of illness in different systems in each Paralympic sport is variable. We show that the respiratory system is the most commonly affected system in most, but not all sports. This confirms the finding of the majority of previously published studies in single and multiple sports codes.3–11 However, we note that the skin and subcutaneous tissue were the most commonly affected system in wheelchair basketball, powerlifting and sitting volleyball, while the genitourinary system was commonly affected in judo athletes. The clinical relevance of these data is that team physicians are now aware of the differences in illness profiles in Paralympic athletes participating in different sports. Unique intervention strategies can now be planned to reduce the illness risk in different Paralympic sports.
The strengths of this study are that it is, to our knowledge, the largest prospective cohort study of illness in Paralympic athletes. We had a very high compliance rate from the participating countries (97%) and were able to monitor >85% of all the athletes participating in the London 2012 Paralympic Games.13 Furthermore, in this study we measured exposure (athlete-days) and were, therefore, able to report IR and not only IP. This methodological consideration is very important in epidemiological studies of this nature because it allows for comparison with other sports events and codes. The main limitations of this study have also been described,13 and relate mainly to reliance on team physician and medical staff to report all illness encounters, and the limitations in the data recorded in the EMDCS database. As a result, information on time-loss illnesses, types of impairments and more detailed clinical information on illness was not available for analysis using this cohort.
In summary, this study showed that the overall illness rate of athletes participating in the Paralympic Games was not higher in the immediate precompetition period, compared with the competition period. Age and gender were not independent risk factors for illness in this cohort of athletes. However, participating in athletics was associated with a significantly higher risk of illness compared with other sports. Perhaps most importantly, these data on illness in Paralympic athletes can now form the basis for planning and testing intervention strategies to reduce the risk of illness, and these strategies may well be different among sports.
What are the new findings
There is a high incidence of illness in Paralympic athletes, mainly affecting the respiratory, skin, digestive, nervous and genitourinary systems.
The incidence of illness was similar during the precompetition and the competition periods.
There was a significantly higher incidence of illness in athletics compared with other sports.
Age and gender were not independent predictors of illness in Paralympic athletes.
How might it impact clinical practice in the near future
These data on illness in Paralympic athletes can now form the basis for planning and testing intervention strategies to reduce the risk of illness, and these strategies may well be different among sports.
Team physicians are now aware of the differences in illness profiles in Paralympic athletes participating in different sports, and can plan unique intervention strategies to reduce illness risk in different Paralympic sports.
The authors would like to acknowledge the significant contributions of the following individuals, who contributed to the work of the research team during the London 2012 Paralympic Games: Dr Richard Budgett, Dr Stuart Miller, Dr Harry Benjamin Laing, Mr Greg Vice and Ms Jane Orr. The authors also thank the team physicians and their medical staff who gave their time to collect the data for this project on a daily basis. They also thank the athletes and the administrators for their support with the project. The authors thank Dr Gabrielle Prinsloo for her assistance with this project, particularly the data management and co-ordination of the submission of the manuscript. This study was approved and supported by the International Paralympic Committee. Financial support for this study was received from the IOC Research Centre in Cape Town. Finally, they thank Acer for the donation of the tablets.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.