14781-koppen

232 Chapter 11 If participants reported the information for an item to be “about right” or “not needed”, they were classified as satisfied (scored 1). Participants reporting that the information was “too much”, “too little” or “not received”, were classified as dissatisfied (scored 0). For this study 2 items were excluded; “whether you can drink alcohol whilst taking this medicine” and “whether the medication will affect your sex life” since these were not considered relevant in our pediatric sample. Consequently, 15 items were included and the total SIMS score ranged between 0-15 points. The SIMS has been used in other pediatric and adult studies in combination with adherence assessment. 14,37 Medication diary In addition to the questionnaire, families were asked to keep a one-week medication diary where they reported the intake of their prescribed medication doses. This diary was returned via mail. Statistical analyses Normally distributed data are presented as means and standard deviations whereas non- normally distributed data are presented medians and interquartile ranges. Data distribution was evaluated using histograms and Q-Q plots. We used parental questionnaires, which were available for all children, for the analyses pertaining to treatment adherence and the correlation with beliefs about medication, illness perceptions, satisfaction about treatment and satisfaction with information about medication. Therefore, all our results are based on proxy-reported data. In accordance with previous studies applying the MARS-5, children with a MARS-5 score ≥23 were defined as adherent and children with a score <23 were defined as nonadherent. 11,17,26–30 For comparisons between the adherent and nonadherent groups, the Mann-Whitney U test was used for non-normally distributed data and the Chi-squared test was used in case of categorical variables. To compare scores for the MARS-5, BMQ subscales, B-IPQ questions, TSQM domain score and SIMS total score between parents and children, the Wilcoxon- signed rank test was used. Spearman’s rank correlation test was performed to analyze the correlation between the BMQ-subscale scores, the B-IPQ-question scores, the TSQM domain scores, the SIMS total score and the MARS-5 score as a continuous variable. Separate analyses were performed for intentional and unintentional nonadherence items of the MARS-5. To test the effects of predictors on treatment adherence (MARS-5 score as a continuous variable), a hierarchical multivariate regression model with two blocks was created. The first block contained variables correlating both to MARS-5 according to Spearman’s