Clinimetric properties of the Activity Monitor for assessing gait parameters in chronic stroke patients
Alphen, J.B. van
MetadataShow full item record
Introduction: A substantial part of the community-dwelling stroke survivors becomes less physically active over time. Post-stroke physical inactivity may cause a loss of physical fitness and increases the risk of comorbidities and a second stroke. There is no ‘gold standard’ to measure walking activity in this population. Assessment of walking activity generally involves subjective or observer-rated instruments. These instruments have disadvantages such as the risk of recall-bias and social desirability of answers. Valid step counters are also used frequently. Nevertheless, these are restricted to the measurement of number of steps. The ‘gold standard’ for measuring walking activity should measure the number of steps, covered distance and walking time over a representative period of time in a reliable and valid way. Therefore, SUSTAIN developed an ActivityMonitor (AM) to measure walking activity in chronic stroke patients. Aim: To investigate the criterion validity and test-retest-reliability of the AM as measure of walking activity in chronic stroke patients. Methods: Community-dwelling chronic stroke patients will be tested twice, with an interval of two to three weeks. They will perform a six-minute walk test and a treadmill test at different speeds at both testing days. Walking activity will be expressed in different gait parameters (i.e. number of steps, distance, mean step length). Output data of the AM will be compared to video analysis as criterion measurement. Intraclass Correlations Coefficients (ICCs), limits of agreement by Bland-Altman, Mean Relative Root Squared Error (validity) and the Smallest Detectable Change (reliability) will be calculated to determine the clinimetric properties of the AM. Results: 29 patients were tested to determine the validity, 21 patients of this group were tested twice to determine the test-retest reliability. ICC values for validity and reliability are high, ranging from .826 to .967. Mean error parameters are all around or smaller than 10%. However, individual error parameters based on the 95% limits of agreement are higher, up to 38.5%, and exceed the minimal clinically important differences for gait parameters. Conclusion: Although mean differences between methods and measurement moments were 10% or lower, the individual measurement errors were larger. Therefore, the AM cannot detect a clinical important difference for the gait parameters. Further research is needed to improve the algorithms of the AM to avoid outliers and to decrease the measurement error. This will offer opportunities for the use of the AM in clinical practice and research of walking activity in this population.