Measurement properties of activity monitoring for a rehabilitation (AMoR) platform in post-stroke individuals in a simulated home environment

ABSTRACT

Aim

The aim of this study was to evaluate the measurement properties of activity monitoring for a rehabilitation (AMoR) platform for step counting, time spent in sedentary behavior, and postural changes during activities of daily living (ADLs) in a simulated home environment.

Methods

Twenty-one individuals in the post-stroke chronic phase used the AMoR platform during an ADL protocol and were monitored by a video camera. Spearman’s correlation coefficient, mean absolute percent error (MAPE), intraclass correlation coefficient (ICC), and Bland–Altman plot analyses were used to estimate the validity and reliability between the AMoR platform and the video for step counting, time spent sitting/lying, and postural changes from sit-to-stand (SI-ST) and sit-to-stand (ST-SI).

Results

Validity of the platform was observed with very high correlation values for step counting (rs = 0.998) and time spent sitting/lying (rs = 0.992) and high correlation for postural change of SI-ST (rs = 0.850) and ST-SI (rs = 0.851) when compared to the video. An error percentage above 5% was observed only for the SI-ST postural change (7.13%). The ICC values show excellent agreement for step counting (ICC3, k = 0.999) and time spent sitting/lying (ICC3, k = 0.992), and good agreement for SI-ST (ICC3, k = 0.859) and ST-SI (ICC3, k = 0.936) postural change. Values of the differences for step counting, sitting/lying time, and postural change were within the limits of agreement according to the analysis of the Bland–Altman graph.

Conclusion

The AMoR platform presented validity and reliability for step counting, time spent sitting/lying, and identification of SI-ST and ST-SI postural changes during tests in a simulated environment in post-stroke individuals.

KEYWORDS:

• Stroke
• activity monitor
• physical activity
• sedentary behavior
• accelerometry

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/10749357.2024.2377520

Additional information

Funding

This work was supported by the [CAPES #1] under Grant [number 001]; and [FAPESP #2] under Grant [number 2019/25569-2]; [FAPESP #3] under Grant [number 2017/13655-6]; and [CNPq #4] under Grant [number 141311/2023-1].