Gestational Diabetes Mellitus: What is Next on This Front with Artificial Intelligence?

Dear Sir, Gestational diabetes mellitus (GDM), marked by pregnancy-induced glucose intolerance, affects 3.8%–21% globally, impacting around 200,000 Indian women yearly. While often resolving post-delivery, 50% may develop type 2 diabetes mellitus within 5 years. Early onset GDM (<20 weeks) may suggest underlying conditions like insulin resistance or maturity-onset diabetes of the young, leading to potential misdiagnosis due to rising incidence and prevalent GDM in India.[1] Risk factors for GDM include pre-pregnancy body mass index, maternal age, ethnicity, family history of diabetes, smoking, and exposure to perfluorochemicals. It leads to maternal complications like pre-eclampsia and neonatal issues. GDM increases long-term risks of diabetes, obesity, and cardiovascular diseases for both mothers and offspring. Perpetuating a transgenerational cycle, research on pathogenesis, early diagnosis, and clinical intervention is vital, and artificial intelligence (AI) can ease this.[2] Limited post-diagnosis self-care/knowledge in GDM underscores its significance. Genetic testing aids early detection and management. Disparities exist between clinical guidelines and real-world implementation.[3] This gap can be reduced with AI-based interventions. Literature reports the use of support vector machine (SVM)-powered mobile app-based fasting blood glucose, sex, and age-based predictions for maternal healthcare and GDM in rural India, with ongoing efforts to enhance its effectiveness.[4] SVM stood out among nine algorithms for GDM diagnosis. AI brings hope to the rural community for pre- and post-natal care for prediction/diagnosis of GDM with an SVM-powered mobile app aiming to revolutionize precision medicine, and remote management in long-distance medical care and decision-making of GDM. Nevertheless, the use of AI in GDM needs to be prioritized based on the underlying purpose of employing AI and its pre-conceived use.[4] Emerging diagnostic molecular technologies enhance GDM management, surpassing glycated haemoglobin or haemoglobin A1c. Soluble factors from adipose tissue and placenta serve as promising markers. Early predictors include various blood markers: viz. mid-pregnancy biomarkers include plasma glycated CD59 and 1,5 anhydroglucitol, visfatin, omentin, leptin, ficolin-3, and fetuin-A and third-trimester markers include fetuin-B, fibroblast growth factor-21, and plasminogen activator inhibitor 1. They aid risk assessment and intervention timing greatly.[5] Indian Council of Medical Research (ICMR) recommends non-fasting 75 g 2-h post-glucose ≥140 mg/dL for GDM screening, balancing practicality and cost-effectiveness. Variability and reliance on a single threshold screening methods serve as limitations, impacting research. Recent GDM strategies involve probiotics and tailored dietary adjustments. Debates persist on probiotic specifics and carbohydrate intervention efficacy. Individualized attention and adherence are essential and can be bridged through AI.[6] An AI tool can send reminders for regular yoga practice, stressing its significance in managing GDM and its benefits for the mother and baby. These reminders can be customized to fit the mother's schedule and preferences, aiding in the effective integration of yoga into her daily routine. Moreover, the AI tool can offer tailored guidance on yoga poses to mitigate GDM risks and reduce stress, ultimately supporting the mother's health during pregnancy. A Mexican study used a piloted AI predictive model based on "The artificial neuronal network" viz. MIDO GDM for fostering the GDM prediction, aiding >1709 pregnant women. Multiple algorithms enable early risk identification in low-resource areas without oral glucose tolerance tests, potentially revolutionizing healthcare. Similar Indian studies could enhance patient care.[7] AI improves GDM pregnancy outcomes with education, support, and personalized interventions. It aids in postpartum glucose monitoring, assesses diabetes risk, supports family planning, and breastfeeding, and monitors maternal and child health. Early newborn screenings detect complications, whereas counseling and genetic evaluation before prenatal care are crucial, especially for those undergoing infertility treatments. Molecular testing integration is vital. AI customizes interventions based on family genetics and lifestyle choices, reducing risks. We propose that AI-enhanced personalized treatments improve GDM management, including data processing, glucose monitoring, and decision-making, benefiting fetal monitoring and complication prediction. Among the many healthcare apps that employ AI for GDM and decision support, focus only on data recording or generic advice only a few comprehensive mHealth apps like NICE pathway, MobiGuide, SineDie tool, and GDm-Health tool provide decision support. Barriers like mass-scale adoption of AI GDM apps by end users and clinical settings and ethical data concerns are often ignored, and further trials and research are imperative.[8] AI clinical support is a crucial aspect, particularly post-coronavirus disease 2019 era wherein remerging infections create barriers to impede clinical care.[9] Changes in clinical management and decision support powered by novel AI approaches will be major drivers for the future of healthcare. Further research is needed for intelligent holistic AI-driven apps to improve community-based self-care outcomes.[8] Ethical use, professional involvement, and patient privacy are crucial and remain the mainstay for employing AI interventions. Enhancing GDM awareness at fertility and antenatal centers requires diverse strategies: education, staff training, preconception counseling, online resources, collaboration with obstetricians, community outreach, support groups, and screening programs. Amid times of delayed clinical visits, embracing telemedicine is necessary, and likewise AI. Utilizing GDM tools like the SineDie app can mitigate pandemic impacts, reducing hospital visits and easing clinician workload and remote patient management and clinical decision-making while ensuring high-quality care.[9] These improvements would revolutionize healthcare. However, government involvement is essential for effective implementation. Authors contributions RU: design, literature search, and manuscript preparation; JB and ST: manuscript review and guidance; PKK: concept design, literature search, manuscript editing/review, guidance, and correspondence. Consent to participate Not applicable. Patients consent form Not applicable. Ethical approval or IRB approval Not applicable. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest.