Maternal Mental Health And The Role of Artificial Intelligence-Based Conversational Agents: A Scoping Review

Abstract

Background: The physical and psychological changes, as well as events associated with the perinatal period, can have a very distressing impact on maternal mental health. About 20% of women experience a mental health problem during the perinatal period, and this remains a global public health problem. Due to the limitations of traditional in-person care, attention has recently shifted to digital methods for more innovative approaches to addressing mental healthrelated issues. Growing research interest is now towards the use of AI-based conversational agents within the digital mental health field. This thesis thus aimed to highlight how AI-based conversational agents developed for the perinatal context support maternal mental health during the perinatal period. Methods: This scoping review was conducted, guided by the framework of Arksey and O’Malley, the JBI scoping review recommendations, and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR). A search was performed across seven electronic databases (MEDLINE, CINAHL, ACM Digital Library, Cochrane Central, IEEE Xplore, Scopus, and ScienceDirect), as well as an additional search in Google Scholar. Primary studies of different study designs published between 2014 and 2024 in English were included during the search. Results: A total of six studies were included in this review, and data was extracted. The studies were mostly conducted in HICs (n = 4). The mental health condition that was most addressed was depression and depressive symptoms (n = 5). Narrative synthesis presented the basic shared functionalities of AI-based conversational agents used in the perinatal mental health context, with other unique features being specific to their intervention target. Thematic analysis identified five themes as facilitators and four themes as barriers. Digital benefits stood as the dominant facilitator, while Technological challenges were the dominant barrier to engagement with these agents among perinatal women. Conclusion: Interaction with AI-based conversational agents during the perinatal period has a positive impact on maternal mental health. Continued engagement with these agents and with the features of their associated platforms also provides support in other aspects of maternal well-being. To increase the generalisability of findings, and enhance engagement among perinatal women, future studies should prioritise higher methodological designs and the use of more advanced-level NLP techniques in developing these conversational agents.