Abstract
BACKGROUND: Accessible early intervention is crucial given the high prevalence of youth mental health problems. In addition to walk-in peer counselling for 12-25-year-olds, the @ease centres in the Netherlands offer an online chat service. This study aimed to contribute to the limited literature on online peer support by analysing links between topics, counselling techniques, and young people's responses, informing larger future analyses and practice. METHODS: In 2023, a total of 2145 online chat sessions were held with 1160 individuals. A randomized subset was drawn for feasibility of this exploratory retrospective mixed methods study (n = 26). For the qualitative part, hybrid latent content analysis was performed using Atlas.ti 24 to infer topics, techniques, and response types. User responses were categorized as indicators of either good (e.g. elaborative, reflective) or poor client collaboration (e.g. closed, dismissive). A logistic regression analysis in SPSS 27 was conducted to explore associations between counselling techniques and response types. RESULTS: Most-discussed topics were social and occupational problems, problems related to parents, professional help and diagnoses, intimacy and sexuality, and suicidality. Of the 623 user responses within the 26 online chats, 423 (67.9%) reflected good client collaboration. Collaborative responses were positively associated with peer counsellors' empathising/affirming (OR = 4.13, 95%CI 1.53-11.16, p = 0.005) and negatively linked with their asking closed-ended questions (OR = 0.23, 95%CI 0.11-0.47, p < 0.001) (also when combined with empathising/affirming: OR = 0.25, 95%CI 0.12-0.54, p < 0.001), the combination of empathising/affirming with giving advice (OR = 0.17, 95%CI 0.07-0.46, p < 0.001), and giving advice only (OR = 0.27, 95%CI 0.12-0.61 p = 0.002). CONCLUSIONS: Actively listening through empathising/affirming, without directly asking a question or giving advice, helped young people in opening up when accessing online chat peer counselling. Future research could leverage large language models to enable extensive, topic-specific analyses, as the present sample size warrants carefulness in drawing conclusions, though complying with standards for qualitative research and sufficient counts for quantitative analysing. Gained insights can help optimize online peer support for youth such as that offered by @ease.