Changes in depression symptom profile with gender-affirming hormone use in transgender persons 99 Table S3.3. (continued) Study aim Parameter Specification Results section Longitudinal analyses in the TF group, stratified by complete or incomplete GAHT use Dataset All measurements of TF participants who contributed both a baseline and at least one follow-up measurement. Supplementary materials table 1. Model Linear Mixed model: Total IDS-SR score. Baseline score (continuous) + Measurement time point (categorical; baseline, 3 months or 12 months) * GAHT use (categorical; complete GAHT use or incomplete GAHT use) + (1 | Participant number) Longitudinal analyses per group (i.e. TM and TF), adjusted for age, cohort and cycle regulation use Dataset All measurements (of participants who contributed both a baseline and at least one follow-up measurement. Only TM participants were included in the model with adjustment for cycle regulation and testosterone form, and only TF participants were included in the model with adjustment for estrogen form. Supplementary materials table 2. Model Linear Mixed model: Total IDS-SR score. Baseline score (continuous) + Measurement timepoint (categorical; baseline, 3 months or 12 months) * Group (categorical; TM or TF) + Covariates 2. + (1 | Participant number) 1. IDS-SR total score, and IDS-SR subscores (i.e. mood, anxiety, lethargy and somatic symptoms). All outcome scores were log-transformed due to skewing of the absolute scores. 2. Age (continuous variable), cohort (categorical variable; RESTED or ENIGI), cycle regulation (categorical variable; use or no use of cycle regulation), testosterone form (categorical; transdermal or intramuscular) and estrogen form (categorical; transdermal or oral) were added as covariates. 3.3. EFA results Bartlett´s test of Sphericity was significant (p-value<0.05) and the KaiserMeyer-Olkin factor adequacy index indicated adequate sampling (KMO=0.87). Items with a loading greater than 0.4 were included in the clusters. The conducted four-factor EFA showed a Tucker Lewis fit index (TLI) of 0.92 and a Root mean square error of approximation (RMSEA) of 0.043 (90% confidence interval from 0.03 to 0.056). The resulting factor loadings and item statistics of the four-factor EFA model are reported in Supplementary Table S3.4, as displayed below.
RkJQdWJsaXNoZXIy MTk4NDMw