Since the definitions of clinically acceptable Kappa and ICC values are both arbitrary and the Kappa statistics is very sensitive to the number of positive findings, the interpretation of the reliability of the tests will be based on all parameters in the analyses, including percentage agreement, Kappa/ICC values, measurement error and limits of agreement.
The agreement (measurement error) as measured using the SEM was 0.4 corresponding to 13% of the mean FSS values.
Bicondylar length, epicondylar breadth, anterior-posterior (AP) diameter, medial-lateral (ML) diameter and cortical area at the midshaft were measured by four observers to test the measurement error variance and observer agreement of the protocol (n = 6).
Results from each of the three methods were in complete agreement: measurement error gives rise to a systematic bias where a stationary animal is most likely to be measured as turning 180° or moving towards a fixed point in space.
In conclusion, our research shows that the vertical NRS-FRS has good test-retest reliability and validity in measuring physical fatigue after stroke, with good agreement, low measurement error, and high sensitivity and specificity.
Bias and random measurement error were small, indicating good agreement of both readings.
For the interpretation of the measurement error, the largest limit of agreement (either upper or lower limit) is most relevant.
The relative and absolute reliability of the NRS-FRS showed good test-retest reliability, with high agreement, small measurement error, and no systematic bias for the assessment of poststroke fatigue.
Agreement involves the measurement error of a score.
Reliability is highly dependent on the heterogeneity of the study sample, while the agreement, based on measurement error, is more a pure characteristic of the measurement instrument [ 14].
