Dataset for "Prevalence, Burden and Risk Markers of Health Problems in HYROX Athletes: A 12-Week Prospective Cohort Study"

Data collection method:

Design. Data were gathered through a 12-week prospective cohort study running from 1 December 2025 to 1 March 2026, in which a single cohort of HYROX athletes was followed and surveyed repeatedly over time. Sampling and recruitment. A convenience sample was recruited through emails to HYROX-affiliated gyms, flyer distribution across the University of Bath campus, and social media. The sample was self-selected rather than randomly drawn. Eligibility required participants to be aged 16 or over, fluent in English, engaged in HYROX-specific training at least once weekly, and to have at least three weeks of HYROX training experience; individuals with a current medical condition likely to substantially alter their training were excluded. Eighty-nine participants were enrolled. To encourage sustained response, prize draws (Amazon vouchers) were held every four weeks, contingent on at least 80% questionnaire completion. Baseline measurement. Following electronic informed consent, baseline demographic and background data (age band, sex, HYROX training experience, previous time-loss injury history, and recent competition history) were collected once at entry. Weekly measurement. Each week for 12 weeks, participants completed the Oslo Sports Trauma Research Center Questionnaire on Health Problems (OSTRC-H2), a validated self-report surveillance instrument that records any health problem (injury or illness) and its consequences for training participation, training volume, performance, and symptoms, regardless of whether training time was lost. Questionnaires were distributed each Sunday with a reminder issued to non-responders after 48 hours. Alongside the questionnaire, participants self-reported their training and competition exposure for the week, disaggregated by activity type (running, resistance training, HYROX-specific conditioning, prehabilitation, and competition). Instrumentation and capture. All data were collected electronically via the QuestionPro survey platform. The four OSTRC-H2 item responses were combined into a weekly severity score (0–100) for each reported problem, following the standard scoring of the instrument. Derivation and structuring. Reported problems were classified as acute injuries, overuse injuries, or illnesses, and coded for attributed activity, injured body region, mechanism, and onset type. Problems were flagged as "substantial" (a moderate or severe reduction in training or performance, equivalent to a score of 17 or above on the relevant items) or "time-loss" (at least one missed training day). The raw responses were compiled into a participant-level sheet and a weekly-records sheet (663 participant-weeks), with a separate dictionary defining every variable and its codes. Records were anonymised, with participants identified only by an arbitrary code. Ethics. Ethical approval was granted by the University of Bath Research Ethics Committee (reference 12792-15539), and procedures followed the Declaration of Helsinki.

Data processing and preparation activities:

Third-party datasets. No third-party or pre-existing datasets were used. All data were collected first-hand from study participants via weekly questionnaires, so no external data were modified or merged. Preparation for publication. The deposited file is the study's raw collected data, retaining the original weekly questionnaire responses and self-reported exposure as captured in QuestionPro. The following preparation steps were applied before deposit: Anonymisation. Direct and indirect identifiers were removed. Participants are identified only by an arbitrary participant code that cannot be linked back to an individual. Exact age was replaced with age bands, and any free-text or contact information collected during consent and administration was excluded from the deposited file. Structuring. Responses were organised into a participant-level sheet (one row per participant) and a weekly-records sheet (one row per participant-week), and a variable dictionary was added defining every field and its numeric codes. Derived variables. A small number of fields are derived rather than raw and are retained in the file for transparency: the OSTRC-H2 severity score (computed from the four item responses), the substantial- and time-loss-problem flags, and total exposure hours (summed across the reported activity categories). All underlying raw item responses are retained, so these derivations can be reproduced or recomputed. No values were imputed, smoothed, or otherwise altered, and no records were removed during preparation. Missing responses (for example, weeks in which a participant did not complete the questionnaire) are left as missing rather than filled.

Technical details and requirements:

Data collection. Data were collected electronically using the QuestionPro online survey platform (web-based; no specific version dependency for the collected data). Participants completed the weekly Oslo Sports Trauma Research Center Questionnaire on Health Problems (OSTRC-H2) and exposure items through any standard web browser on their own devices; no specialised hardware, sensors, or laboratory equipment were used. The OSTRC-H2 is a self-report instrument administered as described in the accompanying methodology; the four item responses were combined into a 0–100 severity score using the instrument's standard scoring. File format and viewing. The dataset is supplied as a single Microsoft Excel workbook (.xlsx). It can be opened in Microsoft Excel (2010 or later) or any compatible spreadsheet application (for example LibreOffice Calc, Google Sheets, or Apple Numbers). No macros, add-ins, or external links are used, and the file requires no password. The workbook contains three worksheets: a participant-level sheet, a weekly-records sheet, and a variable dictionary. Reuse and replication. No proprietary software is required to read or analyse the data; the .xlsx can be imported into any statistical environment. The analyses reported in the associated publication were produced in R (version 4.5.2) using the MASS, tidyverse, readxl, sandwich, broom and ggpubr packages, with the analysis code deposited separately. To reproduce those analyses, the code reads the workbook directly (via the readxl package); the variable dictionary sheet provides the field names and numeric codes needed to interpret every column.