Answers are grounded in this paper and cite their sources.
Prospective analysis protocol with executable code, deposited before any analysis of Osteoarthritis Initiative participant data. The author has not accessed those data. Background. A single gait speed summarizes current functional capacity but cannot distinguish people who walk at the same speed today and have been losing speed at different rates. Against a requirement common to everyone, baseline reserve carries no information beyond the gait speed itself; the rate of change is the candidate that could add information. This protocol specifies the empirical test of that claim. Design. Landmark analysis in the Osteoarthritis Initiative. In the primary design, each participant's exponential loss rate is estimated from gait speed at years 0, 1, 2 and 3; current level is the separate year-4 gait speed; first crossing below the requirement is ascertained at years 6 and 8. A secondary design estimates rate from years 0 to 2, takes level at year 3, and ascertains crossing at years 4, 6 and 8. Primary comparison. Level-only versus level-plus-rate discrete-time hazard models for first observed crossing below 0.91 m/s, with a likelihood-ratio test and bootstrap out-of-bag change in discrimination and prediction error. Interpretation. Three outcomes are defined in advance: a positive result, an informative null in which rate is measured reliably and adds nothing, and a measurement-limited null in which rate is measured too imprecisely to test the question. Only an informative null counts against the hypothesis. Files: protocol (PDF) and a package with the analysis code, a synthetic-data generator for testing, and a README. The code refuses to run until the visit-code and variable mapping has been confirmed against the OAI data dictionaries.
No takes yet. Share an insight, caveat, or question.
Robert T. O'Leary (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: