<![CDATA[Timely referral of pregnant women with high-risk conditions is an important component of maternal healthcare. Delayed referral may increase the risk of maternal and fetal complications, particularly when clinical deterioration occurs rapidly. Conventional risk identification in primary midwifery care may depend on routine screening and clinical judgment, which can result in variation in referral decision-making. Artificial intelligence (AI)-based predictive risk assessment may provide additional decision support by identifying high-risk pregnancies and facilitating earlier referral. This study aimed to determine the effect of AI-based predictive risk assessment on the referral time of high-risk pregnant women in primary midwifery care. A quantitative quasi-experimental study with a pretest-posttest control group design was conducted among 80 pregnant women identified as having pregnancy risk factors. Participants were allocated into an intervention group receiving AI-based predictive risk assessment in addition to standard care (n = 40) and a control group receiving standard risk assessment (n = 40). The primary outcome was referral response time, defined as the interval between identification of referral indications and initiation of the referral process. Data were analyzed using the independent t-test or Mann–Whitney U test and multivariable regression analysis. Statistical significance was set at p < 0.05. The mean referral response time was 42.6 ± 16.8 minutes in the AI intervention group compared with 67.8 ± 24.5 minutes in the control group. The intervention group demonstrated a significantly shorter referral response time (mean difference = −25.2 minutes; 95% CI: −34.4 to −16.0; p < 0.001). After adjustment for maternal age, gestational age, parity, distance to referral facility, and previous pregnancy complications, AI-based predictive risk assessment remained significantly associated with faster referral (adjusted β = −21.7 minutes; 95% CI: −31.8 to −11.6; p < 0.001). AI-based predictive risk assessment was associated with significantly faster referral of high-risk pregnant women in primary midwifery care. AI-supported risk stratification may serve as a decision-support tool to strengthen early identification and timely referral while maintaining professional clinical judgment.]]>
