Effects of a Province-wide Change in Gestational Diabetes Mellitus Screening Policy on Treatment and Newborn Birth Weight

Gestational diabetes mellitus (GDM) screening and diagnosis practices vary internationally. The two most used approaches are the International Association for Diabetes in Pregnancy Study Group (IADPSG) criteria, which use a one-step screening process, and the Carpenter-Coustan (CC) criteria, a more conservative two-step approach. Observational studies have suggested that using the more inclusive IADPSG criteria may reduce adverse perinatal outcomes, but randomized controlled trials have not demonstrated this benefit. Prior trials have been limited by their pre-post study designs that compare outcomes before and after screening policy changes without adequately controlling for temporal trends and other confounders. This study uses an interrupted time series (ITS) design to evaluate changes in the incidence and trends of GDM-related outcomes before and after a 2010 screening policy change in British Columbia, which shifted from two-step CC screening to a mixed one-step and two-step approach interpreted using IADPSG criteria. Data were obtained from the British Columbia Perinatal Data Registry and included all births at ≥28 weeks’ gestation between 2004 and 2019. Exclusion criteria were inactive insurance, multiple gestation, pregestational diabetes, late antenatal care, and lack of recommended GDM screening. The primary outcomes included the incidence of medication-treated GDM, lifestyle-treated GDM, and large for gestational age (LGA) infants. Secondary outcomes were small for gestational age (SGA) and health care utilization, measured by at least one endocrinologist visit. Incidences and trends of the outcomes were established across 2 time periods: prior to 2010, when higher glycemic thresholds using the two-step CC approach were standard (consistent with current ACOG guidelines and common US practice), and after 2010, when lower glycemic thresholds based on one-step IADPSG criteria were implemented. Notably, the adoption of IADPSG standards did not result in complete elimination of the 50 g glucose challenge test, resulting in a hybrid screening approach. The ITS method was used to estimate both immediate incidence changes as well as changes in trend attributable to the policy shift, independent of underlying temporal patterns. Covariate analysis accounted for potential confounders, including a 2016 guideline recommending labor induction for GDM patients. A total of 463,879 pregnancies were included. Between 2005 and 2019, the incidence of medication-treated GDM increased from 1.4% to 4.4%, while lifestyle-treated GDM rose from 6.4% to 10.6%. ITS analysis showed an immediate 1.85% increase (95% CI, 1.19%-2.51%) in lifestyle-treated GDM incidence following the policy change, with no subsequent change in trend. Medication-treated GDM incidence showed no immediate increase but demonstrated a slightly increased trend of 0.23% per year (0.09%-0.37%). There were no immediate or trend changes for incidence of LGA or SGA infants or endocrinology visits. However, the preexisting downward trend in LGA infant diagnosis slowed after 2010, resulting in a higher overall incidence than would have been expected without the policy change. Sensitivity analyses using narrower time windows attenuated some findings, including loss of significance for the medication-treated GDM trend. After adjusting for increased cesarean delivery and labor induction rates following 2016 guideline changes, there was no observed change in cesarean births, inductions, or preterm deliveries. Overall, adoption of the more inclusive IADPSG criteria was associated with increased diagnosis of lifestyle-treated GDM without significant improvement to infant birth weight outcomes, outside of a small increase in LGA incidence based on trend changes. These findings align with recent randomized controlled trials and challenge the use of lower GDM diagnostic criteria for improving perinatal outcomes. The authors encourage a risk-benefit analysis that accounts for patient experience, greater health care utilization, and cost. Study strengths include its large population-based design and control for temporal trends, while limitations include the inability to fully disentangle one-step and two-step screening methods after the policy change. (Abstracted from Diabetes Care. 2025;48(8):1410–1417. doi:10.2337/dc25-0480)