A risk assessment model for gestational diabetes mellitus integrating maternal demographics and hematological indices in Thai women

Universal screening for Gestational Diabetes Mellitus (GDM) is the gold standard, but it can be difficult and expensive in resource-limited settings. Many current prediction models rely on specialized tests that are not available in primary care. This study aimed to create a cost-effective risk assessment model using common maternal demographic and blood test data to help identify GDM risk in Thai pregnant women. We conducted a cross-sectional study with 200 pregnant women (24–28 weeks of gestation) at a secondary care hospital in Southern Thailand. We analyzed routine antenatal data, including complete blood count and other lab results. A risk-scoring system was built using multivariate logistic regression. We then tested the model’s performance using Receiver Operating Characteristic (ROC) curve analysis to determine the optimal cut-off point. Six independent predictors were found maternal age over 35 years, systolic blood pressure above 130 mmHg, glycosuria, hemoglobin at least 11 g/dL, platelets at least 260.50 × 10 3 /µL, and neutrophil-to-lymphocyte ratio (NLR) of 2.89 or higher. The combined risk model performed better (AUC = 0.804; 95% CI: 0.74–0.87) than single-marker models. With a risk score cut-off of 2.5 (indicating three or more risk factors), the model had a sensitivity of 71.4% and a specificity of 78.1%. The new risk assessment model, which uses basic demographic and routine blood test data, offers a reliable and affordable way to screen for GDM. This method enables early risk identification and helps healthcare providers use resources more efficiently, potentially reducing unnecessary tests in resource-limited settings.