JMIR Publications

ABSTRACT

Background:

Pediatric medical chatbots are increasingly being developed; however, evidence of their clinical applicability remains limited, and most studies emphasize feasibility, usability, and acceptability rather than objective outcomes. Because pediatrics is not simply scaled-down adult care, these interventions carry multigenerational implications, shaping lifelong health trajectories.

Objective:

We aimed to synthesize evidence on pediatric- and caregiver-facing AI medical chatbots by mapping existing tools, targeted conditions and functions, and their application in improving clinical and patient-centered outcomes.

Methods:

PubMed, Embase, Scopus, PsycINFO, the Cochrane Library, and the Web of Science were systematically searched without publication year restrictions. Randomized controlled trials and observational studies evaluating medical chatbots for children (<19 years) or caregivers were included. Non-original articles, end-of-life or palliative care studies, and non-English publications were excluded. Study quality was assessed using the Mixed Methods Appraisal Tool and the Oxford Levels of Evidence 2.

Results:

In total, 7 studies (mixed-methods, randomized controlled trials, and nonrandomized) were included, with sample sizes ranging from 20 to 10,387 and participants comprising children, adolescents, caregivers, and parent–child dyads. Chatbots were used for preventive care, procedural preparation, behavior change, and psychological support, using both AI-enabled and rule-based systems and often involving caregivers. Reported outcomes were most consistent in preventive and behavioral domains, including increased human papillomavirus vaccination uptake and improved health literacy, while oral-health and dietary interventions showed improvements in reported behaviors with limited changes in clinical outcomes. Procedural chatbots reduced preprocedural anxiety and improved knowledge and satisfaction; however, mental health outcomes were inconclusive. Engagement was generally low to moderate over 14 days to approximately 4 months, with improved outcomes associated with higher interaction frequency rather than longer duration.

Conclusions:

This systematic review maps the current landscape and applicability of medical chatbots in pediatrics, highlighting their use for health education and behavior change across diverse clinical contexts, particularly in caregiver-mediated care. However, substantial methodological heterogeneity and unclear reporting of underlying technologies limit the interpretability of the current evidence, underscoring the need for larger rigorous studies, developmentally tailored designs, and more standardized reporting of technological architectures.