111 Surgery vs botulinum for the impact of drooling on daily life in children with neurodevelopmental disabilities: a randomized controlled trial 5 (communication aids, tablets, electronic wheelchair) due to drooling during the past two weeks (dichotomous yes or no). The questionnaire was formerly developed by our multidisciplinary saliva control team to evaluate the effect of BoNT-A on the impact of drooling and has shown to be sensitive to change over time.2,16-18 It contains multiple visual analogue scales, several multiple choice and some openended questions.16 Statistical analysis All data was analyzed using SPSS 22.0 for Windows (SPSS Inc, Chicago, IL, USA). Patients who crossed over to the other treatment arm were analyzed according to the intention to treat principle. Patients who were excluded or withdrew from the study were not included in the analyses. Descriptive statistics were used to summarize patient characteristics. Paired and unpaired sample t-tests were used for comparing normally distributed continuous variables such as bib or shawl replacements, while ordinal level variables, including drooling severity were compared using Mann-Whitney and Wilcoxon tests and nominal level variables via chi-square tests. Missing data were missing at random and automatically imputed 20 times by SPSS, stratified by treatment group, with all patient characteristics and clinical outcomes as variables in the model. Multiple imputation for missing data on the last three questions regarding emotional reactions (Q15-17) were performed separately, excluding childrenwith an estimated developmental age <4 years as they were considered unable to express their feelings on physical appearance and social acceptance.16,17 Drooling severity in 13 positions and daily activities was analyzed in children able to participate in the displayed positions and daily activities, excluding children to whom the position or daily activity was not applicable. Differences in change between the interventions from baseline to 8 and 32 weeks were evaluated using mixed-model analysis for the continuous variables. Binary logistic regression was used to test for differences between treatments within dichotomous variables with the intervention and baseline values as independent variables. A p-value ≤0.05 determined statistical significance and 95% Confidence Intervals (95% CI) were given for interpretation of the point estimates and significance levels.
RkJQdWJsaXNoZXIy MjY0ODMw