Overcoming the novelty effect on YouTube: visibility patterns in evidence-based psychotherapy videos-Findings of a...

As the second most visited website globally, YouTube offers a major opportunity for disseminating mental health knowledge and reducing stigma around psychotherapy. However, visibility remains a key challenge, particularly in the long-term because YouTube algorithm favors novelty over relevance. This novelty effect may discourage scientifically trained psychotherapists from publishing high-quality content due to its limited lifespan. Yet, videos created by clinicians can be directly recommended during in-person therapy or professional exchanges, potentially bypassing algorithmic limitations and maintaining long-term visibility. This study investigated whether such videos maintain long-term viewership and explored the factors influencing their engagement.

We analyzed viewership data automatically collected by YouTube for 51 videos posted between 2018 and 2022 on the GREMO HUS channel. All videos were developed by academic psychotherapists, grounded in evidence-based practices and organized into three playlists: "Meditation & Relaxation," "DBT Skills," "Compassion Focused Therapy."

Average monthly channel views increased significantly from 365.3 in 2019-2020 to 1,669.5 in 2021-2022 (p < 0.001), alongside a rise in average viewer retention from 24.7 to 33.9% (p < 0.001). At the video level, monthly views increased both with time since publication (p < 0.001) and by calendar year (from 3.8 views per month in 2019 to 45.9 per month in 2022; p < 0.001). Video length was positively associated with views for meditative videos, but negatively for "DBT skills" videos. Viewer retention improved with the inclusion of role-play scenes but not by speech rate, except for "Meditation and Relaxation." User comments frequently mentioned the videos' utility in face-to-face psychotherapy and professional training.

Our findings demonstrate that evidence-based videos can overcome the novelty effect commonly associated with YouTube when embedded in real-life therapeutic practices. This suggests a hybrid dissemination model-leveraging both digital platforms and offline clinical networks-may effectively bypass algorithmic limitations. Future research should investigate the generalizability of these findings to other healthcare domains and platforms, and further explore the mechanisms through which professional recommendation influences digital engagement.