Statistics from Altmetric.com
Sports medicine clinicians are expected to make accurate diagnoses, estimate prognoses and identify athletes at risk of sustaining an injury.1 These complex decisions are dependent on clinical reasoning, which is informed by, and often biased toward, a practitioner’s scientific knowledge and experience. Clinical prediction models are developed by researchers to help facilitate such decisions in practice2; data for multiple predictor variables are combined to estimate an individual’s risk of a health outcome either being present (diagnosis) or whether it will occur in future (prognosis).3 Despite being employed widely in clinical medicine, clinical prediction models are uncommon in sports medicine. Clinical prediction models can offer benefits to both practitioners and athletes, but only if they are developed and validated using rigorous methods and transparently reported so that potential users can judge their accuracy and usefulness.
Therefore, the purpose of this editorial is to describe the recommended steps for clinical prediction development and validation and to guide practitioners using and interpreting prediction models in sports medicine.
The first step in developing a prediction model is to identify its clinical need, the target population and how and when it would fit into the clinical workflow. Models should predict outcomes that are relevant to sport stakeholders, and be clearly defined, including how and when assessed.4
Next is to identify any existing models that could be evaluated or updated. If not, …
GSB and TH are joint first authors.
Twitter @DRGSBullock, @dt_hughes
Correction notice This article has been corrected since it published Online First. The author affiliations have been corrected.
Contributors GSB, TH, JCS, MJC, RR and GC conceived the study idea. GSB, TH, JCS, MJC, RR and GSC were involved in design and planning. GSB, TH, RR and GC wrote the first draft of the manuscript. GSB, TH, JCS, MJC, RR and GC critically revised the manuscript. GSB, TH, JCS, MJC, RR and GC approved the final version of the manuscript.
Funding GSC was supported by the NIHR Biomedical Research Centre, Oxford, and Cancer Research UK (programme grant: C49297/A27294).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
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.