Mental health is a global issue, and mobile applications, such as chatbots, offer a partial solution by providing improved services through various communication forms. This study aimed to identify chatbots and their technical features in mental health services. This study conducted a systematic review of mental health chatbots and their technical features from 2000 to 2025. A search was performed across databases such as PubMed, Scopus, ProQuest and the Cochrane database. The CASP (Critical Appraisal Skills Programme) appraisal checklist was used to assess the quality of the studies. In the next step, the Braun and Clarke's approach was utilized for conducting thematic analysis on the data. The search yielded 2,921 records, of which 10 were duplicates and removed. After screening for relevance and eligibility, 33 papers met all the requirements. The mean quality score of the included studies was 13.36 (standard deviation = 1.36). The studies had a moderate risk of bias, as they mostly had a clear question, searched for the right type of papers, included all relevant papers and reported the results precisely. The research conducted an analysis of 138 mental health chatbots, categorizing them based on five distinct attributes: the disorder they target, their input and output modalities, the platform they operate on and their method of generating responses. The research emphasized the need for designing chatbots that suit patients' preferences and needs, and also indicated that the digital divide within societies should be taken into account when designing and producing chatbots for mental health services. Although mental health chatbots can assist underserved communities, ethical concerns must be addressed before their deployment.
Khosravi et al. (Thu,) studied this question.