Psychometric properties of loaded and unloaded timed stair tests in assessing people with knee osteoarthritis.

Single-site study in Hong Kong; results may not generalise to populations with different staircase dimensions or cultural contexts. Only unilateral knee osteoarthritis was included and severity was not stratified, so the findings may not apply to patients with bilateral involvement or more advanced disease. The affected knee was not necessarily the dominant limb, which may have added variability. Handrail use was allowed but not standardised, so completion times may reflect different movement strategies. Only completion time was measured; movement quality, compensatory strategies, and safety risks during stair negotiation were not assessed. Inter-rater reliability was established with only two experienced raters who timed each trial simultaneously, so the near-perfect agreement may not hold across a broader pool of clinicians. The sample was predominantly female (42 of 60 in the KOA group), which may limit generalisability given known sex-related differences in knee osteoarthritis. Cross-sectional design: the study establishes test properties at a single time point but does not track how the test performs as a measure of change over a rehabilitation course.

No conflicts of interest declared. Funded by the General Research Grant from the Research Grants Council of the Hong Kong Special Administrative Region (Project No. 15101023), awarded to Prof. Shamay S.M. Ng and her team.

Do not read the inter-rater ICC of 1.000 as proof that any two therapists will score identically. The two raters timed the same trial simultaneously with stopwatches, so their readings were essentially the same measurement, not an independent check. And do not treat the cut-off scores as a definitive diagnostic test: the AUC of 0.771–0.782 is only moderate discriminative ability, meaning a meaningful proportion of classifications will be incorrect.