Methodological heterogeneity biases physical activity metrics derived from the Actigraph GT3X in multiple sclerosis: A rapid review and comparative study.

The literature summary was a rapid review with single screening, no registered protocol, and no risk-of-bias or sensitivity analyses. Most included records did not report Actigraph methods clearly enough to be reproduced, so the review could not fully assess how common or valid each method was. The comparison used 42 analysable datasets from 47 enrolled people, all ambulatory with reduced walking ability, so it does not cover wheelchair users or the full MS range. Activity was measured during inpatient rehabilitation with structured exercise and physiotherapy, so the physical activity levels are not representative of normal home life. The analysis tested many method combinations on the same raw data, which can create dependencies and overfitting, although the researchers used participant random effects, model selection and multiple-testing corrections to reduce this. The study did not compare the methods against a gold standard, so it shows that choices change the numbers, not which choice is most accurate.

Do not read the 92.9% meeting a 7,500-step target as evidence that most participants were sufficiently active. The same sample met that target in only 33.3% of cases with the standard filter, so the result reflects the data-processing choice rather than a settled physical activity level.