73 The enhancing effects of testosterone in exposure treatment for social anxiety disorder the enhancement (session 1) and transfer to unenhanced exposure (session 2) separately. Its effects on SAD symptoms (SPS scores) were tested in an additional model (see below). Moreover, we explored augmentation effects on HR in similar models as subjective fear (see Chapter 4 - Appendix 1 for details). We used the Lme4 package in R (Bates et al., 2013) and p-values were calculated using the likelihood ratio tests (Afex package) (Singmann, 2013). Independent continuous predictors were centered and sum-to-zero contrasts used. Consistent with the recommendations for mixed models (Pek & Flora, 2018), we report unstandardized effect sizes (estimates). More specifically, to determine whether the added testosterone had affected self-reported fear (SUDs) during session 1, group (T/P) and time (start, 2 min, 4 min, 6 min, 8 min, end) were included as predictors. Linear, quadratic, and cubic time terms were modeled since we expected that the SUD scores would not necessarily follow a linear pattern (e.g. they could increase first [fear activation] before they decrease). Participant was included as random intercept and time (linear, quadratic, cubic) as random slope. Initial SUDs (rated after psychoeducation) were included as a fixed factor to control for variance unrelated to time or group; see also (Rodebaugh et al., 2013). To examine whether enhanced exposure affected fear patterns during the second unenhanced session, we ran the same analysis used for the first session. As regards the effects of the enhancement on symptom severity, we modeled SPS scores, with group (T/P) and time (pre/post/FU) as predictors and participant as the random intercept. Finally, in light of recent insights into the role of endogenous testosterone on the effects of exposure treatment for SAD (Hutschemaekers et al., 2020), we conducted post-hoc tests, re-running all our analyses now including basal testosterone levels (mean samples 1 and 2). Since we detected some outliers (visual inspection of boxplots) in the baseline data, we repeated the analyses after winsorizing (i.e., setting extreme baseline testosterone values to the second and 98th percentile to thus reduce the effects of spurious outliers). The results remained unchanged. Also, given that age and hormonal birth control are known factors affecting endogenous testosterone, we checked if the observed effects would hold after correcting for these variables in all models. Results Atrition One participant receiving placebo dropped out before the first exposure due to illness (see participant section). Another participant in the same group dropped out during the first session (3.6%). All other participants completed both sessions and the follow-up. 4
RkJQdWJsaXNoZXIy MTk4NDMw