Chapter 3 78 Longitudinal course of symptoms Since all models violated the assumption of normally distributed residuals, the IDS-SR score and subscales were log-transformed (i.e. log(Y + 1)). All estimated changes have therefore been reported as estimated percentages of change. To analyze changes in the IDS-SR total score and the subscales derived from the clusters as found in the EFA analyses, we used linear mixed models with the lme4 package (Bates et al., 2015) and lmerTest package (Kuznetsova et al., 2017) in Rstudio. For all analyses addressing baseline measurements (i.e. EFA analyses and comparison of IDS-SR scores and subscores between the TM and TF groups at baseline) we used the full baseline sample. Linear regression was used to assess differences in baseline scores between TM and TF groups. Longitudinal analyses comparing baseline to follow-ups (i.e. examination of changes in IDS-SR scores and subscores after 3 or 12 months of GAHT) were conducted in participants who contributed both a baseline and at least one follow-up measurement. Linear mixed models were used to assess changes in the IDS-SR scores and cluster scores after 3 and 12 months of GAHT. The linear mixed models assessing changes over time included time point as categorical fixed predictor, participants' baseline score as fixed predictor, and a random intercept per participant. Differences between the TM and TF group over time were studied by adding an interaction term to the aforementioned model. Analysis outcomes were reported using medians of the scores as well as the estimated change resulting from the linear mixed models. The full model specifications are also reported in Supplementary Table S3.3. 3. Results 3.1. Participant demographics The total participant sample included 199 participants, of whom 164 completed the IDS-SR at baseline, 148 at 3 months of GAHT and 105 at 12 months of GAHT. In the full cohort (i.e. n=199), participants’ median age was 24 years (IQR: 21 to 29) old, and 13.1% used psychotropic medication.
RkJQdWJsaXNoZXIy MTk4NDMw