Measurement properties of the 30-second sit-to-stand test in post COVID-19 condition: Results from the PYCNOVID randomised controlled trial

Abstract

Background The 30-second sit-to-stand test (30s-STS) is a frequently used measure of functional capacity in post-COVID-19 condition (PCC), but its measurement properties have not been comprehensively investigated. Methods We used data from the PYCNOVID randomised controlled trial to examine feasibility, safety, test-retest reliability, construct validity (correlations with key symptoms and physical activity), and responsiveness to change of the 30s-STS. Data were collected at screening, at baseline (approximately 2 weeks after screening), and at follow-up (12 weeks after baseline). Using screening and baseline data, agreement was evaluated using Bland-Altman analysis (mean bias and limits of agreement), alongside the intraclass correlation coefficient (ICC) and the minimal detectable change (MDC95). Correlation coefficients were used to assess relationships between the 30s-STS test, physical activity, and key PCC-related symptoms. The minimal important difference (MID) was estimated with data from baseline and at 12-weeks using anchor- and distribution-based methods. Results Data from 150 adults with PCC (74.7% female) were analysed. The 30s-STS was safe (no adverse event) and feasible. Overall, test performance was reduced, 46% of participants had z-score values below the 5th percentile. The test showed good test-retest reliability (ICC 0.82 [95% CI, 0.76 to 0.86]) with no indication for a systematic learning effect; however, the limits of agreement were wide. Mean bias between was 0.84 repetitions (95% CI, 0.26 to 1.42). The MDC95 was 4.25 repetitions. Correlations with physical activity and key PCC-related symptoms were weak. Distribution-based methods yielded MID values between 1.38 and 3.23 repetitions. Change correlations between 30s-STS repetitions and the anchors were insufficient for anchor-based MID calculation. Conclusions The 30s-STS test is feasible, safe, and demonstrates good test-retest reliability. Changes exceeding four repetitions surpass measurement error (MDC95), while distribution-based estimates suggest a provisional MID of two repetitions. Further studies are needed to establish a triangulated MID to better capture patient-relevant change.