Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Consensus on the descriptors, definitions, and reporting methods for heading in football studies: A Delphi study

  • Kerry Peek ,

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Validation, Writing – original draft

    Kerry.Peek@FIFA.org

    Affiliations Sydney School of Health Sciences, Faculty of Medicine and Health, The University of Sydney, Sydney, Australia, FIFA Medical, Fédération Internationale de Football Association, Zurich, Switzerland

    ⨯
  • Andrew G. Ross,

    Roles Data curation, Investigation, Methodology, Validation, Writing – review & editing

    Affiliations Sydney School of Health Sciences, Faculty of Medicine and Health, The University of Sydney, Sydney, Australia, Physiotherapy Department, College of Sport, Health and Engineering, Victoria University, Melbourne, Australia

    ⨯
  • Paula R. Williamson,

    Roles Methodology, Writing – review & editing

    Affiliation Department of Health Data Science, University of Liverpool, Liverpool, United Kingdom

    ⨯
  • Julia Georgieva,

    Roles Investigation, Methodology, Writing – review & editing

    Affiliations Sydney School of Health Sciences, Faculty of Medicine and Health, The University of Sydney, Sydney, Australia, Curtin School of Allied Health, Curtin University, Perth, Australia

    ⨯
  • Thor Einar Andersen,

    Roles Investigation, Methodology, Writing – review & editing

    Affiliations Oslo Sports Trauma Research Center, the Norwegian School of Sport Sciences, Oslo, Norway, The Norwegian Football Association’s Sports Medical Centre, Oslo, Norway

    ⨯
  • Tim Meyer,

    Roles Investigation, Methodology, Writing – review & editing

    Affiliation Institute of Sports and Preventive Medicine, Saarland University, Saarbrucken, Germany

    ⨯
  • Vincent Gouttebarge,

    Roles Investigation, Methodology, Writing – review & editing

    Affiliations Department of Orthopedic Surgery and Sports Medicine, Amsterdam UMC location University of Amsterdam, Amsterdam, The Netherlands, Section Sports Medicine, Faculty of Health Sciences, University of Pretoria, Pretoria, South Africa, Football Players Worldwide (FIFPRO), Hoofddorp, The Netherlands

    ⨯
  • Sara Dahlen,

    Roles Investigation, Methodology, Writing – review & editing

    Affiliation Oslo Sports Trauma Research Center, the Norwegian School of Sport Sciences, Oslo, Norway

    ⨯
  • Mike Clarke,

    Roles Methodology, Writing – review & editing

    Affiliation Centre for Public Health, Queen’s University Belfast, Belfast, United Kingdom

    ⨯
  • Andreas Serner

    Roles Conceptualization, Investigation, Methodology, Writing – review & editing

    Affiliation FIFA Medical, Fédération Internationale de Football Association, Zurich, Switzerland

    ⨯

Abstract

Heading in football (soccer) is a complex skill involving deliberate head-to-ball contact, which may pose short-, medium-, and long-term risk to player brain health. However, understanding header exposure during matches and training sessions, as well as comparing header incidence between studies is currently challenging given the lack of standardisation in descriptors, definitions, and reporting methods. This Delphi study aimed to establish a consensus on the descriptors, definitions, and reporting methods for heading in football research to improve consistency and quality. The study involved 167 participants from diverse football-related backgrounds including coaches, players, medical personnel, and researchers, with consensus achieved to include 27 descriptors in minimum reporting criteria for heading in football research. An additional 27 descriptors were also defined for inclusion in an expanded framework. The operational definition of a header was standardised as “a head-to-ball contact where the player makes a deliberate movement to redirect the trajectory of the ball using their head.” The consensus framework provides a standardised approach to heading in football research to enhance data quality and comparability across studies. Improved header incidence data quality has the potential to contribute significantly to our understanding of the risks associated with heading in football to inform future research and practice guidelines.

Citation: Peek K, Ross AG, Williamson PR, Georgieva J, Andersen TE, Meyer T, et al. (2025) Consensus on the descriptors, definitions, and reporting methods for heading in football studies: A Delphi study. PLoS One 20(7): e0327189. https://doi.org/10.1371/journal.pone.0327189

Editor: Mario Lopes, Universidade de Aveiro Escola Superior de Saude de Aveiro, PORTUGAL

Received: March 5, 2025; Accepted: June 11, 2025; Published: July 17, 2025

Copyright: © 2025 Peek et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the paper and its Supporting information files.

Funding: This project was part funded by a Thompson Equity Prize (The University of Sydney) awarded to KP.

Competing interests: KP, AS, TM, TEA, VG are all currently employed by non-profit football organisations. AS and KP declare employment with FIFA. TM is chairman of UEFA´s and the German FA´s (DFB) Medical Committee as well as of the German League Organisation´s (DFL) working group “Medicine in Professional Football”. TEA is Chief Medical Officer of the Football Association of Norway (NFF). VG is the Chief Medical Officer at FIFPRO. JG has been a paid contracted Injury Spotter for FIFA organised tournaments (2023). TM and VG are non-paid members of FIFA’s Heading Expert Group. This does not alter our adherence to PLOS ONE policies on sharing data and materials. All authors declare no other relevant financial or non-financial competing interests.

Introduction

Research on heading in association football, which includes the incidence and characteristics of headers, could be explored as one approach of assessing the potential causal inference between header exposure and long-term issues with brain health in retired players [1]. A recent systematic review and meta-analysis exploring the risk to football players in developing any neurodegenerative disease (including Alzheimer’s disease and other dementias, motor neurone disease, amyotrophic lateral sclerosis, and Parkinson’s disease) compared to general populations from nine included studies reported an odds ratio of 1.69 (95% CI 1.11–2.58) with significant heterogeneity between studies [2]. However, header, head impact, and/or head injury exposure was either not objectively collected or completely absent as an outcome measure in these studies, with causal inferences often based on career length [3], and playing position (out-field versus goal keeper) [1]. Causal inference frameworks assist in understanding causal relationships between factors particularly in circumstances where randomisation of exposure that ensures exchangeable comparison groups is considered infeasible [4]. The longitudinal collection of objective header data (for instance, using video analysis), based on a standardised framework of descriptors and definitions, could provide a more accurate report of header exposure, when compared with previous measures relying on player estimation [5]. Furthermore, objectively collected header exposure data can assist when exploring the potential short-term (defined as a single session of headers) [6], or medium-term (which includes header exposure over multiple practice sessions or across a season/s in active players) risk to player’s brain health. Exploring the potential short- and medium-term risk associated with repeated headers offer an alternative study design given the infeasibility of completing a randomised clinical trial in which young players are assigned to play football with or without heading with their cognitive function monitored over many decades [7]. However, if heading in football research is to contribute to our understanding of the potential relationship between header exposure and short-, medium, or long-term brain health of football players then the collection of header exposure data needs to be of the highest quality possible.

A recent systematic review highlighted several shortcomings related to current heading in football research, most importantly the lack of a standardised operational definition of a header, with some studies making no distinctions between a deliberate head-to-ball contact (i.e., a header), and an accidental or unintentional ball-to-head impact (such as a player being hit in the face by a ball delivered at close range) [5]. This distinction is important given that higher head impact forces have been measured when the ball contact is unexpected [8]. Furthermore, there was an absence of minimum reporting criteria in these studies related to player demographics, heading descriptors and their definitions, as well as reporting methods [5]. This Delphi study aims to bridge this gap by establishing a framework of descriptors and definitions as well as reporting standards to improve the consistency and quality of heading in football research. Therefore, the objective of this study is to achieve consensus in the descriptors, definitions, and reporting options that should be included in minimum reporting criteria for heading in football research.

Methods and analysis

Project steering committee

A Project Steering Committee was convened that included prominent experts in heading in football research, and research methods, while also considering representation across professional backgrounds, research career stage, and country of origin. The final Project Steering Committee included 10 members, 4 women, including the project lead, and 6 men, who originated from 9 different countries worldwide. Professionally, 2 members were Delphi and consensus study experts, and 6 members had health, and health related research backgrounds. Two members were also former professional football players, and 1 has experience in coaching. Seven of the members were mid- to late-stage researchers and 3 were early-stage researchers, including 2 members completing PhDs which included heading incidence research. Collectively the committee have published more than 50 research papers on heading, and/or head injuries in football. See S1 Appendix for more details.

Protocol, and ethics

The protocols for this study were registered on Open Science Framework (https://osf.io/qh6un and https://osf.io/qarxg) and was conducted using a phased approach, as described in Fig 1, following Guidance on Conducting and Reporting Delphi Studies (CREDES) [9], and ACcurate COnsensus Reporting Document (ACCORD) [10]. As this study did not involve the collection of personal health-related data, or other sensitive data, it was granted an exemption from ethics provided by Swiss Association of Research Ethics Committee, Switzerland (BASEC-Nr: Req-2O2 4-OO3.23).

thumbnail
Download:
Fig 1. Overview of the Delphi Method (adapted from Prinsen at al. [11]) for each phase 1-3 used in this study.

https://doi.org/10.1371/journal.pone.0327189.g001

Systematic review

A systematic review was completed in December 2023 to collate the number and type of heading descriptors used in all published studies which report on heading incidence in football, as well as documenting the data collection and reporting methods used in the included studies to present heading incidence data. The review is published separately [5]. This review identified 71 studies which included header incidence data as an outcome measure for inclusion with the following key findings: 1) only 61% of studies defined a header with even fewer (23%) providing an operational definition of a header within the methods; 2) important study and player demographic data including year and country were often not reported; 3) reported heading descriptors and their reporting options varied greatly between studies; 4) visual identification of headers was essential when inertial measurement units were used to collect header incidence data; and 5) there was a lack of standardisation in the reporting methods used in header incidence studies making comparison between studies challenging, if not impossible [5]. All identified heading descriptors from the systematic review were included in the Phase 1 questionnaire.

For this study, heading was defined as a player performing a deliberate movement to redirect the trajectory of a ball using their head. Therefore, heading could result in either a header (with head-to-ball contact) or an attempted header (without head-to-ball contact) and could be performed in either a contested (duel) or uncontested (no duel) situation.

Questionnaire development

Questionnaires were developed for each objective and phase of this project (Phase 1- descriptors, Phase 2- definitions, Phase 3- reporting methods). These questionnaires were distributed sequentially. In other words, once consensus was reached on the descriptors (Phase 1), a questionnaire to gain consensus on the definitions (Phase 2) was developed and distributed, followed by a questionnaire to gain consensus on the methods of data reporting (Phase 3), Fig 1. All questionnaires were developed and distributed using REDCap electronic data capture tools hosted by The University of Sydney. The Phase 1 first-round questionnaire also included a number of demographic questions to gain more information about each participant including: self-reported role in football (coach, medical personnel, player, researcher), age, gender, country where they had spent most of their football or working career, highest playing level (players only), gender and age group of players they mainly worked with (coaches, medical personnel and researchers only).

The following methods were applied to the development of each questionnaire.

Phase 1 questionnaire.

Phase 1 included two questionnaires, first-round and second-round. These questionnaires focused on gaining consensus on the descriptors that should be included in minimum reporting criteria for heading in football research.

  1. a). All descriptors included in the Phase 1 questionnaires were scored using options 1–9 where 1–3 was considered not important/do not include in minimum reporting criteria, 4–6 important but not critical to include, and 7–9 critical/definitely include in minimum reporting criteria.
  2. b). Any descriptor scoring >80% agreement for options 7–9 (definitely include) in the first-round questionnaire was removed from the second-round questionnaire (as consensus to include had been reached) but were included in the Phase 2 (definitions), and Phase 3 (methods of reporting) questionnaires. These items are included in the final recommended list of descriptors to be included in minimum reporting criteria for heading in football research.
  3. c). Any item scoring >80% agreement for options 1–3 (do not include) in the first-round questionnaire was removed from the second-round questionnaire and but were included in the questionnaires of Phase 2 (definitions) and Phase 3 (methods of reporting) to gain consensus on the definition and methods of reporting even though they were not recommended to be included in minimum reporting criteria for heading in football research.
  4. d). Any items that were removed between rounds were clearly communicated to participants (including what and why). Participants had access to a ‘free text’ box to voice any concerns they had about any removed item in the second-round questionnaire.
  5. e). The remaining items that did not reach consensus in the first-round questionnaire to be included, or excluded, from minimum reporting criteria were added to the second-round questionnaire,
  6. f). Any descriptors suggested by any participant in the first-round questionnaire were also included in the second-round questionnaire. These additional descriptors were highlighted to participants as being suggested by a peer-participant.
  7. g). To assist participants in completing the second-round questionnaire, participants were provided with the mean score for each descriptor (across all participants) as well as their previous score for that descriptor.
  8. h). Descriptors that achieved >80% consensus at that point were automatically included in (or excluded from) minimum reporting criteria. Descriptors that did not achieve consensus (but scored >70%) were discussed and anonymously voted on by the Project Steering Committee as to whether to include, or exclude, and reported accordingly in subsequent questionnaires.

Phase 2 questionnaire.

Phase 2 included two questionnaires, one in a first-round and the other in a second-round. These questionnaires focused on achieving consensus on the definition of each descriptor included in Phase 1.

  1. a. In the Phase 2 questionnaires, participants were provided with three lists of descriptors:
    1. i. The first list contained the descriptors that reached >80% consensus in Phase 1 to be included in minimum reporting criteria for heading in football research as scored by the participants in Phase 1.
    2. ii. The second list contained the descriptors that did not reach consensus in Phase 1 but were voted for inclusion in minimum reporting criteria by the Project Steering Committee.
    3. iii. The third list contained the remaining descriptors that did not reach consensus by participants in Phase 1 or the Project Steering Committee but were included in this questionnaire to gain a standardised definition for these descriptors (for potential use in future research).
  2. b. Below each descriptor, participants were provided with one or more definitions as reported in the published literature (extracted from the published systematic review) [5] to define or describe that descriptor. In cases where no published definition was available, a suggested definition was developed by the Project Steering Committee.
  3. c. Participants were instructed to either select their preferred definition, or to propose a definition of their own.
  4. d. Any preferred definition scoring >80% agreement in the first-round questionnaire (and without a suggested alternative definition) was removed from the second-round questionnaire (as consensus to include had been reached).
  5. e. The remaining descriptors without >80% agreement on the preferred definition were included in the second-round questionnaire, along with any alterative definition/s suggested by any participant in the first-round questionnaire where available. These alternative definitions were highlighted to participants in the second-round questionnaire as being suggested by their peers.
  6. f. To assist participants in completing the second-round questionnaire, participants were provided with the mean results (percentage) on the preferred definition of each descriptor as well as their previously selected preferred definition.
  7. g. Descriptors that reached >80% agreement on the preferred definition were copied over to the Phase 3 questionnaire.

Phase 3 questionnaire.

Phase 3 only included one questionnaire (although a second-round questionnaire was possible, if needed). This questionnaire focused on the methods and reporting options for each descriptor from Phase 1.

  1. a. In Phase 3 participants again were provided with the three lists of descriptors as per the Phase 2 questionnaire.
  2. b. Below each descriptor, participants were provided with the preferred definition as per the results of the Phase 2, as well as the reporting options and methods of reporting for each descriptor. Methods and reporting options were identified from the included studies as part of the systematic review [5]. In cases where no published definition was identified, methods and reporting options were developed by the Project Steering Committee.
  3. c. Participants were asked if they agreed with the methods and reporting options for each list of descriptors with an open text box for participants to add their comments or feedback where they felt edits or amendments should be made.
  4. d. Where >80% of participants agreed with the methods and reporting options in Phase 3, these were included in the final framework of recommended descriptors, definitions, and reporting methods to be used in heading in football research.

Selection of stakeholder groups and participants.

The selection of stakeholder groups from which participants were derived, were discussed, and selected by the Project Steering Committee to reflect the wider football population that might conduct heading in football research or review the results of such studies. The final selected stakeholder groups included:

  • Researchers
  • Medical personnel (from varying professional disciplines including team physicians and physiotherapists),
  • Technical and performance coaches and analysts, and
  • Players (including past and present players)

All participants needed to be 18 years or older to participate. Researchers were selected based on the corresponding author of the included studies in the published systematic review [5]. Medical personnel, coaches and players were recommended by members of the Project Steering Committee based on their networks and knowledge of who could be considered as having expertise on heading in football. Diversity in terms of confederation, country of origin, sex, and ethnicity of participants was also considered. Prior to the distribution of the first questionnaire all potential participants (except researchers as their email addresses were publicly available) were emailed information about this project by the member of the Project Steering Committee who recommended them, with only those agreeing to participate receiving the first questionnaire in Phase 1.

Questionnaire distribution timeframe.

The first questionnaire was distributed on 1st of May 2024, with the fifth and final questionnaire distributed on 29th November 2024, closing on 23rd December 2024. Each questionnaire was open for four weeks with two reminders sent to non-responders.

Completion rate and missing data.

The final list of participants as confirmed by the Project Steering Committee were emailed the first-round questionnaire in Phase 1. Any participant who did not complete this first questionnaire (or declined to participate) were removed from further involvement in the project. Although participants were encouraged to complete every questionnaire, only completion of the Phase 1 first round questionnaire was required to remain involved in this study. Records of completion rate and any missing data were recorded and reported below (see also S1 Appendix for the complete dataset for each questionnaire).

Results

Demographic and completion data

Initially 244 people were invited to participate including researchers (n = 66), medical personnel (n = 79), coaches (n = 58), and players (n = 41). Seventeen emails were returned as undeliverable (all researchers), attempts were made to locate alternative email addresses where able. Three people declined to participate (one coach, one medical personnel, and one researcher) citing time-related factors as a reason to decline. No response was received from a further 57 people. Non-responders included 31 coaches, 11 medical personnel, 8 players and 7 researchers. Of the non-responders, 46 were men and 14 were women, with 15 from Confederation of African Football (CAF), 6 from Confederation of North, Central America and Caribbean Association Football (CONCACAF), 1 from Oceania Football Confederation (OFC), and 38 from Union of European Football Associations (UEFA).

This resulted in 167 participants completing the first questionnaire (68% of invitees). In total, 18 participants (11%) only completed the first questionnaire. All other participants completed a minimum of two questionnaires, with a total of 88 participants (53%) completing all five questionnaires. The highest retention rate of participants were coaches (69%) and medical personnel (64%). Participant demographic and completion data are reported in Table 1.

thumbnail
Download:
Table 1. Participant demographic and completion data.

https://doi.org/10.1371/journal.pone.0327189.t001

Two-thirds (n = 111, 66%) of participants were men with a mean age of 41.7 years, with approximately half (n = 66, 49%) of coaches, medical personnel, and researchers (n = 134) reporting that they mainly work with boys/men players (n = 26, 19% reported that they mainly work with girls/women players, with the remaining n = 42, 32% working with both boys/men and girls/women). Only 4 (3%) coaches, medical personnel and researchers worked mainly with children (12 years or younger, these participants were all coaches), 37 (28%) worked with adolescents (13–17 years), and 93 (69%) worked with adult players. Two medical personnel worked in para-football (one physiotherapist and one doctor).

Participants also played or worked in a range of geographical locations and confederations; Europe (UEFA, n = 82), Africa (CAF n = 35), North America, Central America and the Caribbean (CONCACAF n = 31), Asia (Asian Football Confederation (AFC) n = 9), Oceania (OFC n = 8), and South America (South American Football Confederation (CONMEBOL) n = 2). Of the 33 players who completed the first questionnaire, 32 (97%) were professional players.

Phase 1- Descriptors

The Phase 1 first-round questionnaire included 46 descriptors, 10 of which reached >80% consensus as essential to be included in minimum reporting criteria for heading in football research, particularly where it relates to header incidence and characteristics during match play or training sessions. These were:

  • Activity type
  • Playing level
  • Number of players (overall)
  • Number of teams
  • Sex
  • Age
  • Age group
  • Type of header
  • Head injury
  • Non-header related head impact

In addition, three descriptors reached >80% consensus to not be included in minimum reporting criteria for heading in football research. These were:

  • Ball deliverer
  • Match score at the time the header occurred
  • Final score

Participants also suggested the following additional descriptors, which were then added to the second-round questionnaire:

  • Head covering or head personal protective equipment: The player performing the header was observed to be wearing something that covered all or part of their head or face at the time of the header (such as a hijab) or a form of personal protective equipment (such as head gear, or face mask/shield)
  • Head movement: The movement direction of the player’s head immediately before and/or after they completed the header
  • Head impact magnitude: The linear and/or rotational acceleration forces associated with the header
  • Weather: The weather at the time of the header (i.e., raining, snowing, sunny)
  • Ball: Information about the ball being used
  • Ball speed: The speed of the ball immediately before and/or immediately after the header
  • Referee: The action of the referee in response to the heading event (such as foul called, signal for medical team, play on, VAR)
  • Player size: The relative size of the player performing the header (such as smaller, larger or similar in size to the other players on the pitch)
  • Player stability: How balanced the player heading the ball was at the time of the header
  • Ball tracking: Whether the player was tracking the in-coming ball immediately prior to the header.
  • Playing surface: The specific playing surface of the pitch (i.e., grass, artificial turf)

After completion of the Phase 1 second-round questionnaire, a further 14 descriptors were included as minimum reporting criteria for heading in football research. Five descriptors scored over 70%, which were then discussed by the Project Steering Committee, who then voted for their inclusion in minimum reporting criteria via completion of an anonymous questionnaire (three of which were then added to the minimum reporting criteria). Thus, a total of 27 descriptors were determined to form the minimum reporting criteria (Table 2, Column 1).

thumbnail
Download:
Table 2. Descriptors with >80% consensus to be included in minimum reporting criteria for heading in football research with their definitions and reporting options.

https://doi.org/10.1371/journal.pone.0327189.t002

Phase 2 Definitions

For the Phase 2 questionnaire, 3 lists were created. Forty-two out of 54 descriptors attained >80% consensus on their preferred definition after the first-round questionnaire, with 8 of the 12 remaining descriptors having at least 1 alternative definition suggested by a participant in the first-round questionnaire. These 12 descriptors were then included in the second-round questionnaire along with the 8 alternative definitions. Following completion of the second-round questionnaire, all 54 descriptors had achieved >80% consensus on the preferred definition (including three descriptors which were suggested by a participant in the first-round: Time between headers, Head impact magnitude, and Player size).

Phase 3 Methods and reporting options

All descriptors in the Phase 3 questionnaire reached >80% agreement for the final definition, methods and reporting options, except for 2: ‘head injury’, and ‘body contact during a duel/header’. For the ‘head injury’ descriptor, 78% of participants selected the definition “the header resulted in a potential head injury or diagnosed head injury,” the other 4 definitions ranged from 2–13%. For the ‘body contact during a duel/header’ descriptor 78% of participants selected “a situation where two or more players are competing for the header and the nature of the body contact observed between players” definition, the other 2 definitions ranged from 6–16%.

There were 5 comments from participants that suggested minor wording clarification of 2 reporting options. These were for the descriptor of ‘Playing level’ in list one where the addition of “within a national league system” was recommended to the ‘amateur’ reporting option, as well as the descriptor ‘Referee action’ in list 3 where the option “no referee action” was suggested by 3 participants. As these were minor suggestions, these were discussed by the Project Steering Committee only. Five additional comments were received from participants to outline that they did not agree with the descriptor being included in either that list or being collected at all, no changes were made based on these comments (as consensus had already been achieved across the 3 Phases) although these comments were noted. Two further comments were made regarding being able to reliably collect data for a number of descriptors, again no changes were made as this was beyond the scope of this project.

Operational definition of a header and heading incidence reporting.

In Phase 3, the operational definition of a header was determined to be:

  • A head-to-ball contact where the player makes a deliberate movement to redirect the trajectory of the ball using their head.

It was noted that minimum reporting for heading in football research should include raw number of headers per activity as well as incidence rate per 1000 match (or training) hours. Where data are collected from both matches and training sessions, these data should be reported separately with the exposure hours used to calculate the incidence rate also being reported.

Descriptors, definitions and reporting options

The final descriptors, definitions and reporting options that reached >80% consensus to be included in minimum reporting criteria for heading in football research can be seen in Table 2. Descriptors that did not reach >80% consensus to be included in minimum reporting criteria but were included to gain a standardised definition and reporting options for these descriptors are reported in Table 3. See S1 Appendix for a full list of results.

thumbnail
Download:
Table 3. Descriptors that did not reach >80% consensus to be included in minimum reporting criteria for heading in football research but were included in Phase 2 and 3 to gain consensus on their definitions and reporting options.

https://doi.org/10.1371/journal.pone.0327189.t003

Discussion

The aim of this Delphi project was to gain consensus on a list of descriptors, their definitions, and reporting options that should be included in minimum reporting criteria for heading in football research, particularly research related to header incidence and characteristics during match play and/or training sessions. This resulted in 27 descriptors being recommended, including 24 descriptors which achieved consensus from participants across a range of relevant footballing backgrounds including coaches, medical personnel, players, and researchers. An additional 3 descriptors that reached >70% and were added to the minimum reporting criteria by the Project Steering Committee. A further 27 descriptors also achieved agreed definitions, and reporting options by participants, thereby expanding the framework of standardised descriptors to 54. Having an expanded framework provides increased flexibility when collecting heading data, where additional descriptors beyond the minimum recommended can be collected in a standardised format to address the varied and specific research aims and objectives of future studies.

While having 27 recommended descriptors as minimum reporting criteria for heading in football research may seem a lot, 12 of these descriptors are categorised as demographic descriptors which should only require collection as summary data (such as data collection dates, country (or countries) where data were collected, and number, playing level, sex, and age of players included in the research). Of the remaining 15 descriptors, 7 relate to the player heading the ball, 5 relate to the header itself, 2 relate to the activity type, and the last 1 relates to the ball. Many of the recommended descriptors that relate to the player heading the ball, as well as the header itself, are important to capture as they can assist in understanding the potential risk posed by headers, and the action of heading, as opposed to non-header related head impacts.

One of the most fundamental findings from the systematic review completed to inform this Delphi study was that there were inconsistencies in the operational definition of a header (where one was provided) [5]. We now recommend that a header is defined as a “head-to-ball contact where the player makes a deliberate movement to redirect the trajectory of the ball using their head”. This is considered important to assist in differentiating between headers, which are a deliberate sport specific action, and head impacts which are usually unintentional, and may pose a different risk to brain health.

While a number of studies have reported on the potential increased risk posed to former football players in developing a neurodegenerative disease [1,2,12], incidence data related to headers and head impacts have not been reported. One recent study of 199 former players (aged 50 years and older) recruited through the English Professional Footballers’ Association, which included an estimation of incidence data, reported no overall association between headers, or other non-header related head impacts, and cognitive function, except possibly in forward players who were exposed to higher numbers of non-heading related head impacts [13]. Given that forward players have consistently been shown to head the ball less than other playing positions, particularly central defenders [14], underscores the importance of categorising headers separately from head impacts. In our study, participants recommended that non-header-related head impacts are defined as “a head impact that is not classified as a header”, with reporting options including unintentional ball-to-head impacts, head impacts with another player, and head impacts with an object.

Further, capturing injury data that separate the mechanism of injury between those caused by headers, and those caused by head impacts is vital, given that headers are a less common cause of a potential head injury than head impacts [15]. While players can be injured by the ball, this is more commonly as a result of unintentional ball-to-head impacts [16], with reportedly higher rates of ball-related head injuries in women and girls [17]. An additional finding from the earlier study was an association between self-reported concussion and lower cognitive function using the Preclinical Alzheimer Cognitive Composite [13]. If the data reported from this study [13] are supported and replicated in other cohorts (including women), then it would appear that history of concussion is more deleterious to long-term cognitive function than headers, which has implications for the focus of both head injury prevention initiatives and protection of long-term brain health in football players.

While we recommend that all future heading in football research includes the 27 recommended descriptors, there may be other descriptors from the extended framework that are useful to capture for other purposes. For instance, documenting the time between headers performed by the same player might be considered to inform the design of studies investigating the acute or short-term effects of a single bout of headers on brain health [18]. Specific recommendations on the quality criteria for studies assessing the acute effects of headers have recently been published [6]. Accurately capturing how many headers individual players perform (rather than calculating mean header incidence per player per match based on the numbers of players on the pitch), as well as the time between headers would ensure that studies exploring the short-term effects of headers not only replicate match, or training, conditions, but also consider the ethical implications of requiring players to complete a high number of headers (particularly from high velocity balls) over a short period of time [6].

Additionally, many descriptors recommended for inclusion in minimum reporting criteria overlap with heading performance or technique. For instance, the 3 descriptors: controlled or uncontrolled header, ball-to-head point of contact, and the consequence/outcome of the header. There is limited but emerging research that suggests that head impact magnitude during headers can be reduced by improved technique [14,19]. Therefore, capturing aspects of heading performance or technique alongside incidence data can potentially be used to inform heading coaching frameworks, particularly in young and/or inexperienced players when starting to head the ball. Heading performance can also be used as a metric to distinguish between ‘high and low value headers’ as a means to reduce heading burden, particularly when focusing on headers more or less likely to lead to a goal, or a turnover of possession [20].

We also recommend that minimum reporting on heading incidence should include raw number of headers per activity (matches, and/or training/ practice sessions) as well as incidence rate per 1000 match (or training) hours, with relevant reference to confidence levels. Where data are collected from matches and training sessions, these data should be reported separately with the exposure hours used to calculate the incidence rate (with 95% confidence intervals) also being reported. While it was outside of the scope of this study to report on the best approach to calculate incidence rates, incidence rates of headers have been presented using match exposure time based on earlier published formulas (number of headers/match exposure time) × 1000), with similar calculations recommended for headers during training [21,22].

Limitations

While agreed descriptors, definitions, and reporting options have been provided through this project, assessing the feasibility and reliability of recording each of these descriptors was beyond the scope of this study. However, it is recognised that some of the descriptors may be challenging to collect, both when collecting in-person, and particularly when using lower quality footage for review. Furthermore, it is recognised that individual researchers may interpret some options differently to others. Accordingly, it is expected that intra-rater and/or inter-rater reliability of recording for each included descriptor is completed and reported in any future studies. For minimum reporting descriptors which are not found to have acceptable levels of inter-/intra-rater agreement of >0.80 with an appropriate agreement statistic (e.g., Cohen’s Kappa, Fleiss Kappa, Intraclass Correlation Coefficient, percentage agreement) analysts may wish to combine certain reporting options to enhance reliability. For example, ball-to-head point of contact, which can be challenging to distinguish due to factors such as low footage quality and varying hair lines, may have greater reliability by combining the Forehead and Top of head reporting options.

While the strength of this project was the large number of participants, from a wide geographical area across 4 stakeholder groups, it should be acknowledged that we observed a reduction in completion rate as the project progressed particularly from researchers and players. While the completion rate across all stakeholder groups was > 60% for each of the 5 questionnaires, only 53% completed all 5 questionnaires. Further, it is possible that the strategy used to recruit participants missed some people (particularly players, coaches and medical personnel) with expertise in assessing heading incidence. This may be particularly true for the confederation, CONMEBOL, which had the lowest number of invitees across all 6 confederations. This is important to note for future research to maximise representation across all confederations where this is relevant to the research purpose.

Considerations for future research

While our Delphi study has provided a comprehensive list of descriptors, definitions and reporting options for future heading in football research, adherence to this framework could be further improved by developing training resources using match video examples to highlight how this coding should be applied, which in turn should improve intra- and inter-rater reliability of reporting, further improving research quality. While our descriptors did include whether a header resulted in a potential or diagnosable head injury, further details regarding potential injury events are not fully captured with the included descriptors, particularly if the injury related to a non-header related head impact. For these events, it is suggested that researchers consider the FIFA Football Language with a Medical Extension for coding injury inciting circumstances [23–25]. Finally, as heading technique and performance is an emerging area of scientific enquiry, further exploring additional descriptors related to these aspects of heading might be a useful addition to the literature.

Conclusion

This Delphi project gained consensus from participants across a range of relevant football backgrounds including coaches, medical personnel, players, and researchers, on a list of 27 descriptors, their definitions, and reporting options, that should be included in minimum reporting criteria for heading in football research. A further list of 27 descriptors also achieved agreed definitions, and reporting options. Importantly, a standardised operational definition of a header was agreed, with recommendations that minimum reporting on heading in football research should include raw numbers of headers per activity (matches, and/or training/ practice sessions) as well as incidence rate per 1000 match (or training) hours. Collectively, these recommendations should improve the quality in reporting of heading in football data and enable comparison of header exposure data between studies regardless of when or where the data were collected.

Key points:
  • This Delphi study aimed to establish a consensus on descriptors, definitions, and reporting methods for heading in football research to improve consistency and quality when exploring the incidence and characteristics of headers.
  • We involved 167 participants from diverse football-related backgrounds, including coaches, players, medical personnel, and researchers.
  • Consensus was achieved on 27 descriptors for minimum reporting criteria and an additional 27 descriptors for an expanded framework. These descriptors cover various aspects such as player demographics, type of header, head injury, and non-header related head impacts.
  • We defined a header as “a head-to-ball contact where the player makes a deliberate movement to redirect the trajectory of the ball using their head.” This definition helps differentiate between intentional headers and unintentional ball-to-head impacts.
  • We recommend that future heading in football research includes raw number of headers per activity as well as incidence rate per 1000 match (or training) hours. Where data are collected from both matches and training sessions, these data should be reported separately with the exposure hours used to calculate the incidence rate also being reported.
  • Furthermore, researchers should document the reliability (intra- and/or inter-rater) for each descriptor reported.

Supporting information

S1 Appendix. Additional project data.

https://doi.org/10.1371/journal.pone.0327189.s001

(DOCX)

References

  1. 1. Ueda P, Pasternak B, Lim C-E, Neovius M, Kader M, Forssblad M, et al. Neurodegenerative disease among male elite football (soccer) players in Sweden: a cohort study. Lancet Public Health. 2023;8(4):e256–65. pmid:36934741
  2. 2. Howarth NE, Ji C, Batten J, Pearce AJ, Dawes H, White AJ, et al. Neurodegenerative Disease and Association Football (NDAF): systematic review and meta-analysis. Int J Environ Res Public Health. 2025;22(5):806. pmid:40427919
  3. 3. Orhant E, Carling C, Chapellier J-F, Marchand J-L, Pradat P-F, Elbaz A, et al. A retrospective analysis of all-cause and cause-specific mortality rates in French male professional footballers. Scand J Med Sci Sports. 2022;32(9):1389–99. pmid:35611613
  4. 4. Glass TA, Goodman SN, Hernán MA, Samet JM. Causal inference in public health. Annu Rev Public Health. 2013;34:61–75. pmid:23297653
  5. 5. Peek K, Ross A, Andersen TE, Meyer T, Dahlen S, Georgieva J. Heading in football: a systematic review of descriptors, definitions, and reporting methods used in heading incidence studies. Sci Med Footb. 2024:1–18.
  6. 6. Peek K, Franchi M, Lemmink K, Balsom P, Meyer T. Quality criteria for studies assessing the acute effects of heading: results from a UEFA expert panel. Sports Med. 2024;54(5):1089–95. pmid:38150166
  7. 7. Ueda P. Cognitive function and heading frequency among retired professional soccer players-heading for clarity? JAMA Netw Open. 2023;6(7):e2324368. pmid:37459102
  8. 8. Perkins RA, Bakhtiarydavijani A, Ivanoff AE, Jones M, Hammi Y, Prabhu RK. Assessment of brain injury biomechanics in soccer heading using finite element analysis. Brain Multiphys. 2022;3:100052.
  9. 9. Jünger S, Payne SA, Brine J, Radbruch L, Brearley SG. Guidance on Conducting and REporting DElphi Studies (CREDES) in palliative care: recommendations based on a methodological systematic review. Palliat Med. 2017;31(8):684–706. pmid:28190381
  10. 10. Gattrell WT, Logullo P, van Zuuren EJ, Price A, Hughes EL, Blazey P, et al. ACCORD (ACcurate COnsensus Reporting Document): a reporting guideline for consensus methods in biomedicine developed via a modified Delphi. PLoS Med. 2024;21(1):e1004326. pmid:38261576
  11. 11. Prinsen CAC, Vohra S, Rose MR, King-Jones S, Ishaque S, Bhaloo Z, et al. Core Outcome Measures in Effectiveness Trials (COMET) initiative: protocol for an international Delphi study to achieve consensus on how to select outcome measurement instruments for outcomes included in a “core outcome set”. Trials. 2014;15:247. pmid:24962012
  12. 12. Mackay DF, Russell ER, Stewart K, MacLean JA, Pell JP, Stewart W. Neurodegenerative disease mortality among former professional soccer players. N Engl J Med. 2019;381(19):1801–8. pmid:31633894
  13. 13. Gallo V, Seghezzo G, Basinas I, Williamson E, van Hoecke Y, Davoren D. Cognitive function among former professional male soccer players–the HEADING study. medRxiv. 2024:2024–03.
  14. 14. Peek K, Duffield R, Cairns R, Jones M, Meyer T, McCall A, et al. Where are we headed? Evidence to inform future football heading guidelines. Sports Med. 2023;53(7):1335–58. pmid:37285067
  15. 15. Mooney J, Self M, ReFaey K, Elsayed G, Chagoya G, Bernstock JD, et al. Concussion in soccer: a comprehensive review of the literature. Concussion. 2020;5(3):CNC76. pmid:33005435
  16. 16. Boden BP, Kirkendall DT, Garrett WE Jr. Concussion incidence in elite college soccer players. Am J Sports Med. 1998;26(2):238–41. pmid:9548117
  17. 17. Blyth RJ, Alcock M, Tumilty DS. Why are female soccer players experiencing a concussion more often than their male counterparts? A scoping review. Phys Ther Sport. 2021;52:54–68. pmid:34419893
  18. 18. Snowden T, Reid H, Kennedy S, Kenny R, McQuarrie A, Stuart-Hill L, et al. Heading in the right direction: a critical review of studies examining the effects of heading in soccer players. J Neurotrauma. 2021;38(2):169–88. pmid:32883162
  19. 19. Quintero LM, Moore JW, Yeager MG, Rowsey K, Olmi DJ, Britton-Slater J, et al. Reducing risk of head injury in youth soccer: an extension of behavioral skills training for heading. J Appl Behav Anal. 2020;53(1):237–48. pmid:30924148
  20. 20. Georgieva J, Peek K, Smith A, Lavender AP, Serner A, Arnold EJ. Heading performance outcomes at the FIFA Women’s World Cup France 2019TM. Int J Sports Sci Coach. 2024:17479541241298757.
  21. 21. Peek K, Georgieva J, Serner A, Orest F. Differences in the technical performance of heading between men and women football players during FIFA World Cup 2022 and FIFA Women’s World Cup 2023 matches. BMJ Open Sport Exerc Med. 2024;10:e002066.
  22. 22. Beaudouin F, Gioftsidou A, Larsen MN, Lemmink K, Drust B, Modena R, et al. The UEFA Heading Study: Heading incidence in children’s and youth’ football (soccer) in eight European countries. Scand J Med Sci Sports. 2020;30(8):1506–17. pmid:32335945
  23. 23. Peek K, Aiello F, Avery L, Gardner T, Rutherford H, Massey A, et al. Characteristics of potential head injury situations at the FIFA World Cup Qatar 2022TM. Sci Med Footb. 2024:1–9.
  24. 24. Aiello F, McCall A, Brown SJ, Serner A, Fortington LV, Huurman SAE, et al. Development of a standardised system to classify injury-inciting circumstances in football: the Football Injury Inciting Circumstances Classification System (FIICCS). Sports Med. 2023;53(9):1805–18. pmid:37233947
  25. 25. Aiello F, Avery L, Gardner T, Rutherford H, McCall A, Impellizzeri FM. Broadening our understanding of injury mechanisms to include at-risk situations: an overview of potential injuries at the FIFA men’s world cup Qatar 2022TM. Sci Med Footb. 2024:1–10.