Clinical Prediction Models in Sports Medicine: A Guide for Clinicians and Researchers

Abstract

Synopsis Participating in sport carries inherent risk of injury. Clinicians execute high-level clinical reasoning and decision making to support athletes to achieve the best outcomes. Accurately diagnosing a problem, estimating prognosis, or selecting the most suitable intervention for each athlete is challenging. Clinical prediction models are tools to assist clinicians in estimating the risk or probability of a health outcome for an individual by using data from multiple predictors. Although common in general medical literature, clinical prediction models are rare in sports medicine. The purpose of this article was to (1) describe the steps required to develop and validate (ie, evaluate) a clinical prediction model for clinical researchers, and (2) help sports medicine clinicians understand and interpret clinical prediction model studies. Using a case study to illustrate how to implement clinical prediction models in practice, we address the following issues in developing and validating a clinical prediction model: study design and data, sample size, missing data, selecting predictors, handling continuous predictors, model fitting, internal and external validation, performance measures, reporting, and model presentation. Our work builds on initiatives to improve diagnostic and prognostic clinical research, including the PROGnosis RESearch Strategy (PROGRESS) series of papers and textbook and the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) statement. J Orthop Sports Phys Ther 2021;51(10):517–525. doi:10.2519/jospt.2021.10697

Comparing “people-like-me” and linear mixed model predictions of functional recovery following knee arthroplasty

“…Future work investigating PLM or LMM prediction approaches should evaluate (1) strategies for implementing these predictions into practice and (2) the impact of using these predictions to inform clinical decisions. 32,35 Clinical implementation for both PLM and LMM predictions would require the development of user interfaces.…”

“…They can be used to inform patients about the future course of their condition, estimate prognosis, and select and evaluate treatments. 1,2 Prediction models can account for prognostic differences across individuals and subgroups in a population by adjusting for relevant factors. 2,3 In doing so, prediction models can facilitate a personalized medicine approach, where treatments and decisions are anchored to an individual’s prognosis instead of the average outcomes of the population.…”

“…2,3 In doing so, prediction models can facilitate a personalized medicine approach, where treatments and decisions are anchored to an individual's prognosis instead of the average outcomes of the population. 4 Prediction models may be particularly useful for personalizing care in postoperative settings where (1) clinical status changes rapidly, (2) early care decisions may have long-term impacts on recovery, and (3) personal factors influence the expected course of recovery. The total knee arthroplasty (TKA) population is a good example to illustrate the usefulness of prediction models in postoperative care because (1) patients experience dramatic functional loss acutely after TKA followed by a period of rapid recovery, 5 (2) targeted interventions during the acute phase can improve long-term function, [6][7][8][9] and (3) a combination of personal factors (e.g., BMI, sex) can affect an individual's functional recovery trajectory.…”

“…4 Prediction models may be particularly useful for personalizing care in postoperative settings where (1) clinical status changes rapidly, (2) early care decisions may have long-term impacts on recovery, and (3) personal factors influence the expected course of recovery. The total knee arthroplasty (TKA) population is a good example to illustrate the usefulness of prediction models in postoperative care because (1) patients experience dramatic functional loss acutely after TKA followed by a period of rapid recovery, 5 (2) targeted interventions during the acute phase can improve long-term function, [6][7][8][9] and (3) a combination of personal factors (e.g., BMI, sex) can affect an individual's functional recovery trajectory. 10 Therefore, prediction models can be used to augment clinical decision making and facilitate personalized treatments to improve functional recovery following TKA.…”

Improving Clinical Prediction Model Methods

“…Although there is a propensity to dichotomize continuous predictors for their perceived simplicity of use in clinical practice, this is ill-advised (7). Dichotomizing continuous predictors discards important information, is biologically implausible, and considerably reduces performance (3,5,7). Continuous predictors should remain continuous, and potential nonlinear relationships between predictors and the outcome should be investigated (7) Our motivation is not to critique the time and effort undertaken by the authors (1), but rather help improve the model, as prediction models can have direct effect on patient health.…”