Measuring adherence to pulmonary rehabilitation 47 Table 5 Construct validity of the RAdMAT-NL; Spearman correlation coefficients between RAdMAT-NL subscales and adherence measures Measures RAdMAT-NL Subscale Participation RAdMAT-NL Subscale Communication SIRAS 0.76 0.55 Percentage change on FTSST 0.73 0.49 Self-reported home-based adherence 0.33 0.32 Overall adherence 0.50 0.30 Percentage attendance at scheduled rehabilitation appointments 0.67 0.24 Note: RAdMAT-NL: Rehabilitation Adherence Measure for Athletic Training the Dutch version; SIRAS: Sport Injury Rehabilitation Adherence Scale; FTSST: Five Times Sit to Stand Test All correlations are significant p < 0.01 Discussion This study explored the dimensionality and the construct validity of the Dutch version of the Rehabilitation Adherence Measure for Athletic Training (RAdMAT-NL) in patients with COPD undertaking pulmonary rehabilitation in a primary physiotherapy practice in the Netherlands and Belgium. The study also examined if the RAdMAT-NL can be used as a single measure of adherence. A two-factor model of exercise adherence (Participation and Communication) was found. The RAdMAT-NL demonstrated good internal consistency and evidence for the validity of the questionnaire was indicated. Without the communication items the RAdMATNL can be used as a single measure of adherence. Dimensionality (structural validity) The results of this study show a different number of dimensions present in the RAdMAT-NL compared to the original research [13]; two factors instead of three factors. Factor 2, Communication, is the same factor as in the original research, however Factor 1 and 3 of the study of Granquist et al. are reflected in our study as one factor, Participation [13]. The difference in these results may be due to the translation into a different language and different culture, a different patient group and a different setting. Although a forced one-factor solution showed that all items strongly loaded on a single factor which is consistent with the study of Granquist et al. [13], the fit of the two-factor solution was better.
RkJQdWJsaXNoZXIy MjY0ODMw