Can a computer system spot growth problems in children faster and more fairly than doctors alone? Researchers from multiple institutions tested this question by creating an artificial intelligence tool called the Intelligent Nutritional Monitoring Model (INMM). The system was trained on health data from two large surveys—one from the United States and one from India—to predict whether a child has stunting, which is abnormally slow growth caused by malnutrition.
Stunting remains a significant concern across Asia, affecting children's development and school performance. In Singapore, while overall nutrition has improved, pockets of vulnerability remain in low-income communities and among migrant families.
The AI system learns patterns humans might miss
The INMM works by analyzing basic measurements collected at routine check-ups: age, gender, height, weight, and body mass index. Rather than applying simple rules, the AI identifies hidden patterns in how these measurements relate to each other. For instance, it automatically understands that what counts as normal height and weight varies by age—something a human screener might overlook under time pressure.
When tested against traditional computer programs and other AI systems, the INMM came out ahead. It was more accurate, faster, and required less computing power. Importantly, the researchers tested it on data from two different countries and populations, and it performed reliably on both.
Fast screening without bias, especially valuable in stretched clinics
The system's main strength for busy healthcare settings is efficiency. Community health workers and clinic staff could use it to quickly flag children needing closer attention, without expensive equipment or specialist knowledge. Unlike human judgment, the AI applies identical standards to every child, reducing the chance that some children are overlooked due to unconscious bias.
For Singapore and other Asian contexts with stretched healthcare resources, this could mean faster, more consistent screening across all neighbourhoods. However, it's important to note: this study tested the AI on existing data from two countries. Real-world use in different communities would require further testing to ensure it works equally well in new settings.
What you can do with this information
- Ask your child's doctor about their growth screening process at routine check-ups; systematic tracking catches problems earlier than sporadic checks.
- Keep records of your child's height and weight measurements from each visit so you can spot trends yourself and share them with your healthcare provider.
- If your child attends a clinic with limited resources, ask whether they use screening tools to identify children at nutritional risk, as AI-assisted systems may become more common in the coming years.