Peer-reviewed article
Artificial Intelligence Applications in Pediatric Primary Care for Weight Management: A Systematic Review.
Nicholas Cook, Anindita Roy, Margret J Foster, Kyle M Holland, Alfiya Shaikh Mohd Rafiq et al. ·
Summary
Written by AI, not yet AICS's view · claude-haiku-4-5-20251001This systematic review examined how artificial intelligence tools have been used in primary care settings to help manage overweight and obesity in children and adolescents aged 5-19 years. The researchers found 20 studies where AI was applied, mostly targeting families rather than clinicians, with machine learning being the most common approach. Most AI tools focused on identifying children at risk of obesity early, but there is limited evidence on whether these tools actually change clinical practice or improve long-term health outcomes. The review highlights that future research needs to test AI tools that are actually integrated into clinical workflows and measure meaningful health improvements.
Key points
- 80% of studies targeted patient families, while 20% included both families and clinicians
- Machine learning was the most common AI approach, used in 60% of studies, primarily for obesity risk prediction or stratification
- Fewer studies examined family engagement, counseling, documentation review, or generative AI tools
- Methodological quality of included studies ranged from moderate to high
- Evidence on workflow integration, sustained engagement, and long-term clinical impact remains limited
Limitations
This review identified existing research but does not evaluate the actual effectiveness of these AI tools in improving children's health outcomes or reducing obesity rates. The review cannot tell us whether these AI approaches work in real-world clinical practice or lead to sustained behaviour change in families.
Abstract
BACKGROUND: Pediatric overweight and obesity (OW/OB) is a major public health concern associated with significant short- and long-term health risks. Primary care is a key setting for early identification and management, yet current approaches face barriers related to implementation, workflow, and family engagement. This systematic review examined how artificial intelligence (AI) has been used in pediatric primary care and closely connected clinical settings to support OW/OB management. METHODS: We conducted a systematic review registered in PROSPERO (CRD420251103662) and reported according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 guidelines. MEDLINE, Embase, CINAHL Ultimate, Compendex, and CENTRAL were searched for English-language studies published from 2007 to July 9, 2025. Eligible studies included children and adolescents ages 5-19 years with OW/OB, were conducted in or linked to primary care, and included at least one AI-based approach. Dual-reviewer screening, structured data extraction, and methodological appraisal using the Mixed Methods Appraisal Tool were completed. RESULTS: Of 3854 records identified, 20 studies met inclusion criteria. Most studies (80%) targeted patient families, while 20% included both families and clinicians. Machine learning was the most common AI approach (60%) and was primarily used for obesity risk prediction or stratification. Fewer studies examined family engagement, counseling, documentation review, or generative AI tools. Methodological quality ranged from moderate to high. DISCUSSION: AI tools in pediatric primary care primarily support early identification of OW/OB risk, but evidence on workflow integration, sustained engagement, and long-term clinical impact remains limited. Future research should prioritize prospective, workflow-embedded studies, meaningful clinical outcomes, equity, privacy, governance, clinician training, and policy oversight.