Automated Insulin Delivery in Pregnancy With Type 1 Diabetes: An Integrative Review of Clinical Evidence, Socio-Technical Implementation and Equity.

Automated insulin delivery (AID) has moved rapidly from experimental use to emerging clinical implementation in pregnancy complicated by type 1 diabetes. Randomised trials now show that selected systems can improve pregnancy-specific continuous glucose monitoring outcomes, particularly time in range and overnight glycaemic stability, but the evidence is system-specific rather than class-wide and remains underpowered for definitive maternal and neonatal endpoints. This integrative review synthesises clinical trial, observational, qualitative, implementation, policy and equity evidence to examine how AID can be translated into safe, equitable and sustainable perinatal diabetes care. We argue that AID in pregnancy is best understood as a socio-technical intervention: its effectiveness depends on the interaction between algorithm design, pregnancy physiology, user-device behaviours, education, workforce capability, antenatal workflows, delivery planning, postpartum dose reduction and access to specialist support. We summarise system-specific evidence for CamAPS FX, MiniMed 780G and Tandem Control-IQ, review gaps in maternal and neonatal outcome data and propose a practical implementation framework spanning patient, device, clinician, service and health-system levels. The review also presents a reproductive-stage care pathway from preconception to postpartum and identifies equity risks across high-income and lower-resource settings. The field has moved beyond asking whether selected AID systems improve glucose metrics; the next responsibility is to determine how services can deliver these benefits safely, affordably and equitably for the pregnant people most likely to benefit.