| Popis: |
There is increasing interest in improving psychotherapy training using evidence-based supervision. One approach is live supervision (LS), in which the supervisor offers immediate feedback to the trainee (e.g., via microphone, text messages) during the session. This review summarizes the research on LS and its main results. The databases Web of Science Core Collection, PsycArticles, PsycBooks, PsycInfo, PSYNDEX, Psychology and Behavioral Sciences Collection, and PubMed were searched from inception to January 23, 2020 (including a backward search) and updated November 15, 2020. The inclusion criteria (i.e., main focus on LS, immediate feedback from a present supervisor, psychological setting) were met by k = 138 publications, including k = 8 randomized controlled trials (RCTs; N = 339). Two reviewers independently evaluated the RCTs' risk of bias using the revised Cochrane Risk-of-Bias Tool. Most publications had a family therapy background (59%), were categorized as nonempirical (55%), aimed primarily at describing or comparing specific LS methods (35%), and displayed positive views on LS (87%). Based on the RCTs, LS was superior to no-supervision in 78% of all comparisons, but only in 13% of the cases compared to a delayed supervision (DS) condition (i.e., regarding trainee skills, patient outcomes, or other variables). These results somewhat contradict the overall favorable views in the literature. However, the generalizability is limited due to a lack of high-quality studies and substantial heterogeneity in terms of LS methods, concepts, outcomes, and measurements. Ideas for more systematic research on LS regarding objectives and methods are proposed. Public Significance Statement This review summarizes research on live supervision (LS). LS is a form of supervision in psychotherapy training in which the supervisor observes the trainee's therapy session and provides immediate feedback. The review concludes that LS is probably as effective as delayed supervision (DS), although more high-quality research is needed. |
