Introduction Despite detailed recommendations for sports injury data capture provided since the mid-1990s, international data collection efforts for sport-related death remains limited in scope. The purpose of this paper was to review the data sources available for studying sport-related death and describe their key features, coverage, accessibility and strengths and limitations.
Methods The outcomes of interest for this review was death occurring as a result of participation in organised sport-related activity. Data sources used to enumerate death in sport were identified, drawing from the authors’ knowledge/experience and review of key references from international organisations. The general purpose, case identification, structure, strengths and limitations of each source in relation to collection of data for sport-related death were summarised, drawing on examples from the international published literature to illustrate this application.
Results Seven types of resources were identified for capturing deaths in sport. Data sources varied considerably in their ability to identify: participant status, sport relatedness of the death, types of sport-related deaths they capture, level of detail provided about the circumstances and medical care received. The most detailed sources were those that were dedicated to sports surveillance. Sport relatedness and type of sport may not be reliably captured by systems not dedicated to sports injury surveillance. Only one source permitted international comparisons and was limited to one sport (soccer).
Conclusion Data on sport-related death are currently collected across a wide variety of data sources. This review highlights the need for robust, comprehensive approaches with standardised methodologies enabling linkage between sources and international comparisons.
Statistics from Altmetric.com
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.
Though rare, sport-related deaths have considerable public significance and impact.1 2 Because of their relative infrequency and because they occur in a seemingly fit and healthy population, these deaths have a major effect on family members, other athletes and sporting communities.3 4 This impact is felt as the immediate shock and grief to families and those directly involved in the event. Sports-related deaths also have major long-term societal impacts in terms of years of life lost due to the death, cost of emergency and medical care, and psychological and sociological costs to the sports community. Accurate enumeration of the number of sport-related deaths, how they occur and other sport-specific details is a crucial first step towards their prevention.5–7
Public health monitoring efforts (surveillance, data collection, data analyses and summarisation), can be prioritised based on those health-related conditions that pose the greatest threat to human life or ability to function.8 9 The purpose of monitoring and surveillance is to control and prevent the occurrence of a condition (sport death).9 However, national injury statistics that are reliant on death register and certificate reporting generally lack specific enumeration of sports-related causes that would enable separate categorisation. For example, in the USA, unintentional injuries are the leading cause of death for both males and females aged 1–34 years.10 Motor vehicle traffic, poisoning, drowning and falls rank highest for both males and females.11 Likewise, in Australia, unintentional injury is the leading cause of death among people aged 1–14 years and the second leading cause of death in people aged 15–44 years, after suicide.12 Motor vehicle traffic, poisoning and drownings are among the top unintentional injury-related causes among these age groups.12 These national injury statistics rely on death register and certificate reporting and thus do not provide enumeration of sports-related causes. Sport injury deaths are not readily identifiable from national vital statistics because they are grouped within non-specific injury categories (eg, falls or other) in standard mortality reports. Therefore, specific national data collection systems for some sports injury deaths have been developed to monitor the number and rate of specific types of deaths and to assess the impacts of prevention efforts. For example, in the USA, the National Center for Catastrophic Sports Injury Research has conducted detailed surveillance of sports-related deaths and severe sports injuries since 198213 and maintains a registry of football-related deaths that extends back into the 1930s.14 From 2006/2007 to 2015/2016, an annual average of 22 deaths (five traumatic injuries) occurred among high school and 5 deaths (<1 traumatic injury) occurred among college-level sport participants.15
While basic monitoring of fatal sport-related events can identify when a death has occurred, just knowing the numbers alone is insufficient. There is a critical need for specific details about the events resulting in death if we are to understand how best to prevent them or control their risk.7 This includes detail on the person, the specific injury/condition that occurred, when, where and under what circumstances the event occurred and the type of emergency services and medical care that was available and provided at the sport event location. For example, it is important to know that there was a death in the sport of cricket for basic monitoring purposes. However, for prevention purposes, information about the type of equipment being used (if any) (eg, helmet), the specific activity at the time of the event (eg, batting or fielding), any rule violations (eg, illegal bowling) and whether a first-aid or medical provider was in attendance. This additional level of detail forms the basis for actionable public health measures to prevent further similar deaths from occurring.
Based on the limited reporting to date, the most common types of sport-related deaths are due to acute trauma to the head, neck, spinal cord, internal organs or blunt chest impacts (commotio cordis); environmental-related and exertional-related events such as heat stroke, exertional sickling and anaphylaxis; and sudden cardiac death.16 17 Data collection and reporting of sport-related deaths is dependent on a range of factors relating to the type of death (eg, spinal cord vs internal organ injury), the sport played, the sponsoring organisation (eg, National Collegiate Athletic Association (NCAA) reporting mechanisms), participation level (ie, professional vs organised vs informal or recreational play) and country of origin. Inconsistent reporting could lead to undercounting and biased data on the participants, sports and levels of play that are at risk. There is a need for a comprehensive and systematic approach if we are to enumerate and understand how best to prevent sport-related death. Despite considerable work and recommendations for sports injury data capture that have occurred since the mid-1990s,18–20 data collection efforts for sport-related death remains largely limited to specific settings (eg, the USA7) and certain high-risk sports (eg, horse racing21). Currently, only one of the published consensus statements on collection and reporting of injury/illness in sport includes any mention of death.7
General injury surveillance guidelines have been developed22 23 to cover a broad array of injury problems such as road traffic24 and occupational injury.25 Theoretically, data that include cases of sport-related death should be available from a variety of existing sources including surveillance systems, registries, administrative records, research publications, government vital statistics and publicly available reports. Data sources have been previously described for studying injury mortality in general9 and for specific types of injury including occupational,26 traffic-related27 and sport-related injuries.19 As a first step towards developing recommendations for data collection of sport-related deaths, the aim of this paper is to describe the characteristic features of available data sources from the perspective of their (actual and potential) capture of sport-related deaths. This overview provides important information about the current state of data availability, identifies gaps in data sources and targets areas for further development.
The outcomes of interest for this review are deaths occurring as a result of participation in a sport-related activity. Injury deaths are defined as those directly associated with the activities of the sport (eg, making a tackle and hit by a ball) or those resulting from environmental exposures (eg, heat stroke and insect bite). Non-injury deaths were defined as physical exertion exposures during participation (eg, exertional sickling and sudden cardiac arrest) that occurred while the participant was engaged in sport activity. Note that environmental-related conditions such as heat stroke are included under International Classification of Disease (ICD) external cause coding as an injury,20 yet sudden cardiac arrest during sport would not be considered an injury.
Data sources that could potentially be used to identify death in sport were identified, drawing from the authors’ knowledge and experience as well as a review of the following key references: the WHO guidelines for fatal injury surveillance,23 the International Collaborative Effort on Injury Statistics sponsored by CDC’s National Center for Health Statistics (http://www.cdc.gov/nchs/injury/advice.htm) and the Australia Minimum Dataset Requirements from the Australian Institute of Health and Welfare (http://220.127.116.11/publication-detail/?id=6442467501&tab=2).28 Surveillance systems and registries were defined as data collections according to established guidelines for that data source, not according to the activities of ‘surveillance’ and ‘registration’ as defined public health activities (ie, notification, collection, investigation and recording).29 30
From each data source, the general purpose, case identification, structure and the strengths and limitations of each source in relation to collection of data for sport-related death were summarised in tables, drawing on examples from the published literature to illustrate this application. A list of key data features determined by the coauthors were assessed for each source (data custodian and coverage, case capture rate, timeliness, accuracy, injury/death classification system, sport-relatedness and number of participants). Table 1 enumerates each data feature examined and the assessment. For example, case capture rate was evaluated according to whether the data source was inclusive of all cases: complete, incomplete or variable.
Seven types of resources were identified as potential data sources for capturing deaths in sport. Descriptions of the sources, including their strengths and weaknesses along with link(s) to published examples of the data source relating to death in sport are detailed in the online supplementary table. Table 1 highlights the key features from each source. The purpose of the data sources varied from specific targeted collection of death in sport (eg, surveillance and registry) to ad hoc collection and reporting (eg, public records and media reports). Ownership of data tended to be organisation specific, for example, whoever collects the data owns the data (surveillance, insurance and registries), with or without government interest (eg, trauma registries that are owned by the state or country that sponsors it). Specific state-owned information may include that relating to hospitals, emergency services and vital statistics. Data coverage was identified as ranging from broad coverage of the entire population of interest (eg, US high school athletes), to targeted subpopulations (eg, geographic region) or type of event (eg, sudden death). The timeliness of data being available was generally more rapid in surveillance systems (with dissemination of information being a defining characteristic of this source31) as well as registries, while coroner and insurance reports involve delays as details of cases are first completed.
Supplementary file 1
The classification system used to identify the type of injury/condition for cases in administrative sources frequently included the diagnosis or cause of death. Administrative sources include death records, hospital records, insurance claims, and workers’ compensation claims. Whenever an individual dies, regardless of whether sport-related or not, a record of death or death certificate is filed by a designated official such as a medical examiner, coroner or other medical provider. Deaths due to ‘natural’ causes usually only have a limited set of details from the death certificate (ie, age, gender, cause and underlying causes of death) and generally do not include specifics like the sport played, sponsoring school, club or organisation or medical services provided. Administrative data from hospital and emergency departments may comprise records from one hospital, a group or system of hospitals or data pooled from several hospitals. Such deaths are representative of the coverage/treatment area for the particular hospital or hospital system, that is, individuals who would go to (or be transferred to) that hospital for treatment after a sport injury. Deaths that occur outside of a hospital are not recorded here (eg, death on the field of play). Like hospital data, individuals covered by insurance plans file claims to cover the costs of medical treatment and other care received. Insurance coverage may be medical related, death benefits or catastrophic coverage for the sport participant or the venue where the sport is played. Claims are filed by the family or next of kin, and the organisation evaluates whether the claim will be covered.
Overall, the level of detail varied greatly and depended on the purpose of the data source. For example, death records include cause of death or underlying cause of death that is less detailed compared with hospital records that tend to emphasise diagnosis and provision of medical care before the death. News and media sources may provide details on the individual (eg, sport, age and gender) and what happened (eg, collapsed on sidelines) but often lack detail on medical care provided and the diagnosis.32 The most detailed sources were those that were dedicated specifically to sports surveillance. Of particular importance to surveillance efforts are the inclusion of sport relatedness and type of sport. These key data items may not be reliably captured, or else not captured at all, by the systems that are not dedicated to sports injury surveillance, such death certificates, hospital records, news and social media reports and forensic death investigations.
Additionally, enumeration of the sport participants or the population at risk enables estimation of the incidence and prevalence of death in the population. Surveillance systems and registries included estimates of populations at risk, whereas the number of participants is often unknown for sport, organisation and club specific collections.
The most comprehensive individual data source identified was the National Center for Catastrophic Sport Injury Research (NCCSIR), a primary source for published1 13 17 and unpublished (http://nccsir.unc.edu/reports/) catastrophic sport-related death information in the USA since 1982. The NCCSIR uses a combination of passive and active surveillance strategies, with data routinely disseminated in annual reports. Historically, data were obtained largely from publicly available news and media sources and from a short report completed by school officials (eg, coach or athletic trainer). News and media sources are an important source of information on sport deaths, which NCCSIR has used since its inception; however, these sources do not represent a complete capture of all sport-related deaths. Expansion to a consortium model in 2013 facilitated development of new methods that include an online reporting site where anyone (parent, athletic trainer, coach, bystander, etc and so on) can report a catastrophic sport injury (https://www.sportinjuryreport.org). Deaths are now captured prospectively through regular and systematic searching of publicly available sources and online reporting. If performed and available, autopsy reports are also collected and used.
It was notable that there was only one data source that spanned multiple nations. The only international source for sport-related death identified was the Federation Internationale de Football Association Sudden Death Registry (FIFA-SDR)33 developed in 2014 by the FIFA to collect sudden deaths among football (soccer) players worldwide. Similar to NCCSIR, the registry is composed of deaths captured via publicly available media sources and deaths reported at their website (http://www.uni-saarland.de/page/fifa.html). Developed primarily to address sudden cardiac death, the system collects details of sudden deaths of any cause including traumatic injury. Athlete names are not collected, but contact information for the reporter is collected along with basic information about the player and event. Other than the FIFA-SDR, there is currently no international standard for sharing data between the countries in this problem.
Discussion and recommendations
This review provides a summary of some data sources available to characterise deaths in sport. In general, data on sport-related death is collected across an unlinked set of data sources. The purpose and scope of many of the data sources have largely been driven by the sponsoring organisation and/or user’s goals. Overall, the data sources vary in their ability to identify: participant status (ie, individual was a member of a sport team or a recreational participant), sport relatedness of the death (ie, death directly related to sport activity or environmental exposure or physical exertion during sport), the types of sport-related deaths they capture (ie, trauma registry may not capture heat stroke; hospital data may not capture deaths not reporting to hospital), the level of detail provided about the circumstances (ie, what was the person doing) and immediate treatment received (ie, emergency medical management and surgery). In addition, data sources may or may not overlap with one another illustrating the limitations of relying on a single data source to capture all relevant sport-related deaths. Although researchers have recommended improved sport injury data collection and approaches for fatalities since the mid-1990s,18–20 it is clear that there is considerable room for improvement.
There is increasing public awareness of the importance of preventing sports injuries. This growth in awareness has been driven by a variety of factors including media interest in the effects of repetitive head impacts such as chronic traumatic encephalopathy. Sports-related fatalities are highly visible and sometimes generate enormous public empathy and awareness of the potential for injury prevention in sport. The death of Australian cricket player Phillip Hughes, after being hit in the neck by a cricket ball, resulted in a major outpouring of emotion and led to a review of the safety of the sport and proposals for improved helmet design.34 35 The time is right to begin to design systematic, ongoing and structured international surveillance efforts that allow preplanned sharing of data and allow statistics and advances in safety improvements to be shared internationally.
Our review identified only one current source that permits international comparisons, and it is limited to one sport (soccer). Clearly, there are a myriad of logistical issues that would need to be resolved into order to organise this kind of information and share it across countries. A robust and comprehensive implementation plan for the monitoring of sport-related deaths would need to be developed. Such a system needs to consider the purpose, types of events and sporting participants and exposures of interest, timeliness versus detail and completeness. Current national surveillance systems have the potential for data pooling across multiple countries through the development of standardised criteria and/or the addition of sport-related fields to existing national data collections. However, substantial collaboration between researchers, sporting bodies and organisations will be required if such a plan is to be developed and implemented.
Despite the magnitude of the task, some preliminary actions could be initiated based on this review. Notably, in order for a data source to be most useful for international monitoring purposes, the source must be national in focus, capture a well-defined population and be ongoing, systematic and timely.29 From all resources considered in this review, the surveillance system operated by the NCCSIR was the only source that captures sport-related deaths on a national scale consistently since 1982. There are both strengths and weaknesses to this system, and the NCCSIR data have not been formally evaluated using CDC’s criteria for evaluating injury surveillance systems,36 nor has its true capture rate been established. Without knowledge of the true capture rate, it is not clear whether increases in the fatality rates over time represent true increases in deaths or better and improved capture/reporting of events.1 20
There are a number of other existing sports injury/illness surveillance systems in the USA (eg, High School Reporting Information Online, NCAA Injury Surveillance Program and Consumer Product Safety Commission’s National Electronic Injury Surveillance System-All Injury Program) that are each comprised of samples of participants. Unfortunately, such systems are not immediately useful for surveillance of sport-related deaths given the rarity of these events. The dedicated sport-related death registries identified either target a specific type of event (US Commotio Cordis Registry37), sport body (FIFA-SDR33), location of medical care (Victorian State Trauma Registry38) or are not focused on timely reporting (US Sudden Death in Athletes Registry39).
While data from both surveillance systems and registries contained information to capture the sport or sport activity, in most cases (excluding the NCCSIR) additional work and/or input is required before these sources would support the comprehensive collection of fatalities in sport. Registries and surveillance systems tended to have similar data fields that could be specific to a population, type of injury or condition or exposure. Registries tend to have high capture and are more narrowly focused compared with surveillance systems. However, registries differ from surveillance systems in the timeliness of reporting and the emphasis of surveillance also being inclusive of dissemination of information.
The administrative data sources we reviewed (ie, death/hospital/insurance records) are valuable for monitoring injury/illnesses in specific populations (eg, occupational health for workers’ compensation40) or specific conditions (eg, brain injuries41). The primary weakness of administrative sources is that it is very problematic to establish, in a reliable manner, sport relatedness and type of sporting activity using these sources. Because of this limitation, these sources are not comprehensive in their scope or coverage, and there is a potential for missing data, including an inability to establish a linkage to prevention recommendations that would be specific to a given sport or sport activity.42 43 Like occupational deaths, studying sport deaths suffers from similar challenges due to the ability to establish ‘work-relatedness’44 and ‘sport-relatedness’ of the death. The quality of data coding in administrative data sources may vary according to the background and training of data coders.45 46 A further challenge with administrative sources is that data may not be available immediately, with time required for the custodian of the data to obtain, process and analyse prior to release to outside entities. Nonetheless, the advantage offered by administrative sources is that reviewing these data retrospectively may provide more detailed information relating to diagnoses and medical care received. Government statistics are often derived from a combination of administrative and national survey sources and therefore face similar issues as administration data. In comparison, organisation or sport body-based data sources may lack information on medical care and diagnoses but are able to identify that the individual was a member of the sport team or organisation as well as the sport activity engaged in at the time of death. For example, a national organisation such as NCAA or USA Lacrosse will enumerate a heat stroke death of a collegiate lacrosse player, but details on medical care provided or exact diagnosis may not be available.
As a general conclusion, considerations that influence which source or combination of sources are most ideal to enumerate and describe sport-related deaths include the type of death, sports relatedness and the availability of information about external exposures and types of sporting activity at the time of injury. When investigating deaths that occurred during sport activity, ideal data systems would enable identification of both participant status as well as actual sport participation and could include surveillance, registries, insurance and work-cover claims and research studies. The usefulness of publicly available and other administrative sources depends on whether participant status can be established and/or the availability of details on the sport activity at the time of the event. The type or cause of the death under investigation is also an important consideration. For example, when investigating traumatic injury death due to blunt force chest impact (commotio cordis), the relevance to sport activity is known since sudden cardiac arrest occurred as a result of the impact. Whereas a sport participant who dies of sudden cardiac arrest an hour after competition, the relevance to sport activity to the sudden cardiac arrest may not be so readily established.
Finally, it is important to note that no single source in this review represented a complete enumeration of all sport-related deaths. Additionally, single source monitoring may only provide part of the picture. Pooling of multiple data sources is one approach for monitoring sport-related deaths and is likely to provide a more comprehensive and complete accounting of the burden of sport-related fatality. Capture–recapture approaches for studying severe sport injuries have been proposed.20 47 Though not 100% event capture, this example and recent advances in injury surveillance48emphasises the advantage of linking multiple data sources to obtain a more accurate and comprehensive understanding of death in sport. However, linkage requires availability of the data and appropriate fields (ie, name, event and death dates, sport and location) and methodology (ie, capture–recapture modelling49 50 and probabilistic linkage51) to link the data.
Although we made every attempt to be inclusive and comprehensive, it is possible that some information that would be germane to this review was unintentionally omitted. Additionally, subjective integration of this information is inherent in a review of this nature, and the conclusions to some extent represent the views of the authors; a different group of researchers, looking at the same body of information, may come to different conclusions.
National data collection systems and international pooling of data have been successful strategies in informing prevention for other injury-related outcomes such as motor vehicle traffic deaths24 and occupational deaths.25 Development of national data systems for collection of sport-related deaths, through a surveillance system or registry, would ensure that the appropriate data elements are captured, including fields that define sport participant status and sport-related activity. In order to be useful for prevention, the system should be ongoing and timely with the ability to disseminate information for public health action that inform prevention action in specific sports. Such a system would be enhanced by collaborative efforts between organisations in different countries and resources to support international data sharing as well as resources to support data collection using standardised criteria that are reproducible with fidelity in multiple global settings.
This review identified several overarching challenges in data sources for sport-related deaths including: establishing sport participant status and whether the activity was sport related, cause of death coding (ICD-10), injury prevention measures and timeliness of data reporting. Active surveillance and sports-specific sources can provide timely information on sport-related death that is useful for assessing current safety initiatives and adequacy of medical care. This information is critical if preventative action is to be informed by evidence.
What is already known on the subject
National injury statistics rely on death register and certificate reporting, which generally lack specific enumeration of sports-related causes.
Data sources have been previously described for studying injury mortality in general9 and for specific types of injury including occupational, traffic-related and sport-related injuries.
Accurate enumeration of the number of sport-related deaths, and how they occur, is a crucial first step towards their prevention.
What this study adds
Data on sport-related death are currently collected across a wide variety of data sources.
The most detailed sources were those that were dedicated to sports surveillance. Sport relatedness and type of sport in particular may not be reliably captured by the systems not dedicated to sports injury surveillance.
This review highlights the need for robust, comprehensive approaches with standardised methodologies enabling linkage between sources and international comparisons.
Contributors All authors have contributed to the following: (1) substantial contributions to conception and design, or acquisition of data, or analysis and interpretation of data; (2) drafting the article or reviewing it and, if appropriate, revising it critically for important intellectual content; (3) and final approval of the version to be published.
Funding KLK is the director of The National Center for Catastrophic Sport Injury Research (NCCSIR), which is funded by the following organisations: the American Football Coaches Association, the National Athletic Trainers’ Association, the National Collegiate Athletic Association, the National Federation of State High School Associations, the National Operating Committee on Standards for Athletic Equipment and the American Medical Society for Sports Medicine. SWM and the University of North Carolina Injury Prevention Research Center are partially supported by an Injury Control Research Center award (R49/CE002479) from the National Center for Injury Prevention and Control, Centers for Disease Control and Prevention. The Australian Centre for Research into Injury in Sport and its Prevention (ACRISP) is one of the International Research Centres for Prevention of Injury and Protection of Athlete Health supported by the International Olympic Committee (IOC).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.