Sonja Mensch

100 Chapter 5 RESULTS Participants Sixty children with SMD from 15 different care settings were included in the study (Table 2). Mean age of the children was 7.7 years (2-16), 45% had a cognitive development level of less than 6 months of age and 51% of the children were classified as GMFCS level V. Only baseline data were obtained for six children who were not included in the longitudinal analyses; one child died, two children were severely ill, two children had been transferred to another day-care centre and one moved home. Events Physiotherapists noted a large number of meaningful events, however many events occurred without a clear influence on motor abilities e.g., change in type of medication. Because of this lack of information, we only used the event categories ‘surgery’, ‘increase of contracture’ and ‘pain’ for the analyses. Using these categories, a total of 68 events were recorded during the course of the study for 49 out of 54 children. Construct validity Scores were obtained for 60 children at baseline, however only 27 children remained in the study after 15 months. Children were lost to follow up mainly due to severe illness, transfer to other day-care centres, or circumstances of the participating therapists. Correlations of Movakic and expert judgment scores are presented in Table 3. Responsiveness Classification in the event or no event group was not correlated with the number of measurements per participant ( r =.06; p =.27). Mean absolute Movakic score-changes were significantly larger in the event group (Table 4) ( t =2.21; p =.03), with a moderate effect size (Cohen’s d =.62, 95%-CI [.23, 1.00]). Intraclass correlation coefficients (ICC) of T0 and T1 scores in the event group (ICC = .78, 95%CI .64-.86) was lower than in the no-event group (ICC = .95, 95%CI .94-.96), as presented in Table 4. The confidence intervals of these ICCs do not overlap, indicating that the agreement between the two measurements was stronger in the no-event group than in the event group.