Predicting complications in childbirth: a novel model integrating hypertension risk factors

This study presents a novel predictive model for identifying pregnant women at elevated risk of hypertensive disorders during childbirth, integrating maternal characteristics and extrinsic protective factors. Conducted over two years (2023–2024) in South Kalimantan, Indonesia, the research enrolled 413 pregnant women using stratified sampling, adjusted for attrition. Data were collected via structured interviews and clinical records, focusing on demographics, socioeconomic status, health behaviors, and social support. Structural Equation Modeling (SEM) with Partial Least Squares (SmartPLS 3.3.3) was employed to evaluate relationships among constructs. The model demonstrated strong convergent and discriminant validity, with all indicators retaining loadings ≥ 0.70 and adequate reliability (CR > 0.7, AVE > 0.5). Hypothesized pathways were statistically significant: maternal characteristics (β = 0.209) and extrinsic factors (β = -0.349) both significantly predicted hypertension risk. Notably, higher levels of knowledge, self-care, and spousal support were inversely associated with risk suggesting these factors act as protective moderators. Model fit indices (SRMR = 0.098, NFI = 0.654) indicated acceptable explanatory power, particularly given the exploratory nature and modest sample size. The combined predictors explained 36.6% of variance in hypertension risk, highlighting the need for additional variables (e.g., access to care, genetic predisposition). All variables showed normal distribution, and no multicollinearity concerns emerged (VIF < 5.0). This model offers a practical, evidence-based tool for early risk stratification in antenatal care, supporting targeted interventions to reduce maternal morbidity in resource-constrained settings.