Remco Arensman

79 Effectiveness and cost-effectiveness of e-Exercise low back pain - study protocol The required sample size was calculated according to the recommendations of Campbell et al. (2010) for cluster randomized trials (62,63), taking into account repeated measures of the primary outcome measure physical functioning on the ODI during follow-up (64). An intracluster correlation coefficient of 0.05 was used (65,66). Additionally, to detect a clinical relevant difference between groups at 12 months following baseline, a difference of > 6 points in physical functioning on the ODI (67,68) and a standard deviation of 14.5 (69) were used in the sample size calculation. For the repeated measures of physical functioning on the ODI a correlation of 0.5 is estimated between baseline and follow-up measurements until 12 months follow-up (64). Based on these assumptions (using a power of 80% and α =0.05) and average cluster size of 5, in total 165 patients are needed. With an expected dropout rate of 20% a total of 207 participating patients (104 patients per arm) are needed. Statistical analysis Descriptive statistics (e.g., means and proportions) will be used to explore baseline comparability. To investigate selective attrition, general characteristics and primary baseline variables of dropouts and non-dropouts will be compared All analyses will be performed according to the ‘intention-to-treat’ principle. Missing data for all outcomes and cost measures will be imputed using ‘Multivariate Imputation by Chained Equations’ under the assumption that data are missing at random given baseline confounders. For all analysis, a two-tailed significance level of p < 0.05 is considered to be statistically significant. All analyses will be carried out using IBM SPSS Statistics version 24.0 (Amork, New York, USA). Effectiveness The primary purpose of this study is to estimate the effectiveness of e-Exercise LBP for improving physical functioning compared to usual primary care physiotherapy in patients with LBP. The primary analysis time point for the study will be 12 months following baseline, however 3- and 24-month changes will also be evaluated. To evaluate the overall effectiveness of e-Exercise LBP, differences in change scores per group and time period will be estimated on primary and secondary outcomes using linear mixed models (LMM) with random effects to control for correlation within patients and physiotherapists (70,71). The three- level hierarchy will exist of repeated measurements (level 1), nested within patients (level 2), nested within physiotherapists (level 3). Analysis will be controlled for baseline variables that have been shown to be related to physical functioning, e.g., age, gender, pain severity scores, duration of pain (72–74). In addition, a per-protocol analysis that only includes patients of the intervention group which were adherent to the smartphone application and the entire usual care group will 4

RkJQdWJsaXNoZXIy MTk4NDMw