<p>Patient-reported outcome measures (PROMs) are widely used in musculoskeletal rehabilitation to assess pain, physical function, disability, health-related quality of life, psychological status, and patient-perceived recovery. Although traditionally applied as outcome measures, PROMs may also provide prognostic information and contribute to prediction models of rehabilitation response. However, their dual role as both endpoints and predictors has not yet been sufficiently integrated within a unified methodological framework. This narrative review critically examines the role of PROMs in musculoskeletal rehabilitation, with particular emphasis on their use as endpoints of functional recovery, predictors of rehabilitation outcomes, longitudinal monitoring tools, and instruments supporting shared decision-making. A search of peer-reviewed literature was conducted across global databases, focusing on commonly used PROMs. The available evidence indicates that PROMs capture patient-centred domains that are not fully represented by objective clinical or performance-based measures. As endpoints, they permit evaluation of treatment effectiveness. As predictors, baseline and early-treatment PROMs may contribute to prognostic models, risk stratification, treatment matching, and identification of patients at risk of persistent disability or non-response. Their predictive value is influenced by measurement validity, construct overlap, baseline severity, regression to the mean, ceiling and floor effects, assessment timing, missing data, and validation strategy. The clinical utility of PROM-informed models depends on rigorous instrument selection, transparent modelling, calibration assessment, external validation, and adherence to recognised methodological and reporting standards. PROMs should be considered as potentially valuable components of personalised rehabilitation assessment and prognostic modelling.</p>