Interventional trials that evaluate treatment effects using surrogate endpoints have become increasingly common. This paper describes four linked empirical studies and the development of a framework for defining, interpreting and reporting surrogate endpoints in trials. As part of developing the CONSORT (Consolidated Standards of Reporting Trials) and SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials) extensions for randomised trials reporting surrogate endpoints, we undertook a scoping review, e-Delphi study, consensus meeting, and a web survey to examine current definitions and stakeholder (including clinicians, trial investigators, patients and public partners, journal editors, and health technology experts) interpretations of surrogate endpoints as primary outcome measures in trials. Current surrogate endpoint definitional frameworks are inconsistent and unclear. Surrogate endpoints are used in trials as a substitute of the treatment effects of an intervention on the target outcome(s) of ultimate interest, events measuring how patients feel, function, or survive. Traditionally the consideration of surrogate endpoints in trials has focused on biomarkers (e.g., HDL cholesterol, blood pressure, tumour response), especially in the medical product regulatory setting. Nevertheless, the concept of surrogacy in trials is potentially broader. Intermediate outcomes that include a measure of function or symptoms (e.g., angina frequency, exercise tolerance) can also be used as substitute for target outcomes (e.g., all-cause mortality)—thereby acting as surrogate endpoints. However, we found a lack of consensus among stakeholders on accepting and interpreting intermediate outcomes in trials as surrogate endpoints or target outcomes. In our assessment, patients and health technology assessment experts appeared more likely to consider intermediate outcomes to be surrogate endpoints than clinicians and regulators. There is an urgent need for better understanding and reporting on the use of surrogate endpoints, especially in the setting of interventional trials. We provide a framework for the definition of surrogate endpoints (biomarkers and intermediate outcomes) and target outcomes in trials to improve future reporting and aid stakeholders' interpretation and use of trial surrogate endpoint evidence. SPIRIT-SURROGATE/CONSORT-SURROGATE project is Medical Research Council Better Research Better Health (MR/V038400/1) funded.

Purpose: To compare the Timed Up & Go (TUG) test between subjects with and without hemiparesis, considering the direction toward which they turned, and to determine the potential clinical variables that could explain possible observed differences between the groups and/or the turning directions. Method: Twenty-two hemiparetic and 22 matched control subjects performed the TUG twice, with each one turning in both directions. Measures of the strength and tonus of the quadriceps, gait speed, balance, and fear of falling were also collected. Results: Stroke subjects were slower in the TUG than the control group, independent of the direction toward which they turned (F = 45.87; p < .001). For both groups, similar performances were observed when turning toward the paretic and nonparetic/matched sides (F = 0.50; p = .48). The absolute differences between the two TUG trials were greater for the stroke subjects (p = .001) and were significantly correlated with gait speed, balance, and fear of falling (−0.69 < r < − 0.52; p < .013), with fear of falling being the only variable retained in the regression model (R² = 0.44; p = .001). For the control subjects, no significant correlations were found. Conclusions: The larger differences between the two TUG trials for the stroke subjects illustrated the impact of the turning direction on test performance. These differences were not related to hemiparesis, but to the fear of falling.

Objectives: To assess the intra- and inter-rater reliabilities and measurement errors of seven widely applied performance-based tests for stroke subjects (comfortable/maximal gait speeds and both stair ascending/descending cadences, as well as the Timed ‘Up and Go’ test) and to verify whether the use of different types of outcome scores (one trial, the means of two and three trials, and the best and the worst values of the three trials) affected the score values, as well as their reliability and measurement errors.

Key words: rehabilitation; nervous system diseases; movement; reproducibility of results; clinical protocols; revision