Sample size
In this study a convenient sample of 1504 patients was included as
described in the methods section (). Based on a conservative estimation
of one-third non adherent patients in this population, 501 non-adherent
patients were expected. As eight independent variables were planned to
be included in these multilevel regression analyses, 62 cases per
independent variable were available, which means enough power is
achieved, even taking into account the variance attributable to the
group level (based on an alpha of 0.05, a beta of 0.8).