156 CHAPTER 7 Bias management In case of subsequent invasive or ongoing medical treatment for drooling, drooling was considered (partially) refractory to 2-DL. When new measurements were made before reintervention these were identified as the long-term data, since reintervention took place in the period after 32 weeks. Statistical methods Paired samples t-tests were used for the continuous variables. DS and DF at baseline were compared to DS and DF at subsequent measurements using a Wilcoxon rank test. Descriptive statistics were used for the percentage of patients undergoing additional surgery and/or pharmacological treatment after 2-DL and the analysis on adverse events and caregivers’ satisfaction. The level of significance was set at ≤0.05. We present 95% Confidence Intervals (CIs) when applicable. RESULTS Fifty-four patients were analyzed for eligibility (figure 1). Eleven patients were not available for long term follow-up: eight had deceased from causes unrelated to the 2-DL and three were excluded because of unknown contact details. One patient was excluded from the effect analysis of 2-DL because of additional treatment within the 32 weeks follow-up. This patient and the eight deceased patients were not excluded from the AE analyses. The mean time between the 2-DL and the survey was 5.2 years (SD 3.1, minimum 1.2, maximum 11.9 years). See table 1 for patient characteristics. Primary outcome VAS for severity of drooling (n=30) VAS at 32 weeks and in the long-term were significantly lower than baseline measurements (mean difference -36.5, CI -47.0 to -26.0, p ≤ 0.001; -26.1, CI -36.2 to -15.9, p ≤ 0.001). Moreover, long-term VAS significantly increased compared to VAS at 32 weeks (+10.4, CI 1.0 – 19.8, p = 0.031) (figure 2).
RkJQdWJsaXNoZXIy MjY0ODMw