Maternity versus non-maternity healthcare settings for the diagnosis of gestational diabetes or diabetes in pregnancy.

rationaleSubstantial uncertainty exists regarding how healthcare system factors, especially care setting types, affect the uptake and implementation of gestational diabetes or diabetes in pregnancy screening in practice. Evidence suggests that the care environment may influence diagnostic processes.

objectivesTo assess the effectiveness of maternity and non-maternity healthcare settings for the diagnosis of gestational diabetes mellitus or diabetes in pregnancy. SEARCH

methodsWe searched CENTRAL, MEDLINE, Embase, and Cumulative Index to Nursing and Allied Health Literature (CINAHL) for randomised controlled trials (RCTs) up to April 2026, and for controlled non-randomised studies of interventions (NRSIs) up to December 2025. We also searched CENTRAL to identify ongoing trials. ELIGIBILITY CRITERIA: We included RCTs and planned to include cluster-RCTs and NRSIs, but we did not find any. Eligible participants were pregnant women being screened for gestational diabetes mellitus (GDM) or diabetes in pregnancy. We defined maternity healthcare settings as those specialising in care for pregnant, birthing, and postpartum individuals, with appropriate screening protocols, and non-maternity healthcare settings as general medical environments lacking a focus on pregnancy and women's care. We considered any screening format (e.g. risk-factor-based screening) and any diagnostic test (e.g. one- and two-step oral glucose tolerance tests (OGTT)). OUTCOMES: Critical health service outcomes were compliance with screening and follow-up guidelines, diagnosis of gestational diabetes or diabetes in pregnancy, maternal emergency department visits during pregnancy, maternal antenatal hospitalisation, length of maternal hospitalisation, length of neonatal hospitalisation, and care accessibility. Critical maternal outcomes included hypertensive disorders, maternal death, intensive care admissions, hyperglycaemia requiring pharmacological treatment, infections, and postpartum glucose intolerance, including progression to diabetes, within 15 years after birth. Critical child-related outcomes encompassed perinatal death, neonatal intensive care admissions, preterm birth, neonatal hypoglycaemia, large-for-gestational-age, shoulder dystocia, and birth trauma. Further critical and important outcomes not included in the summary of findings tables can be found in the full text. RISK OF BIAS: We used the Cochrane RoB 2 tool to assess bias in the RCTs. We had planned to use ROBINS-I version 2 to assess bias in the NRSIs if we had identified any. SYNTHESIS

methodsWe used standard Cochrane methodology. We planned to synthesise results for each outcome using meta-analysis where possible (inverse-variance or Mantel-Haenszel approaches; fixed-effect model). We used GRADE to assess the certainty of the evidence for each outcome. INCLUDED STUDIES: We included one RCT with 781 participants: 390 were randomised to screening in a maternity healthcare setting and 391 were randomised to screening in a non-maternity healthcare setting. Except for screening compliance, trialists evaluated all outcomes using a per-protocol analysis instead of by original randomised groups. This approach created substantial disparities in sample sizes between settings and largely nullified the methodological protections typically provided by a randomised design. The trial evaluated screening for GDM only. We did not identify any ongoing studies or studies awaiting classification. SYNTHESIS OF

resultsHealth service outcomes Screening in a maternity healthcare setting may increase compliance with screening and follow-up guidelines (risk ratio (RR) 1.69, 95% confidence interval (CI) 1.53 to 1.87; 781 participants) and may improve care accessibility, assessed as time to access antenatal diabetes care, measured in days (mean difference (MD) -4.70, 95% CI -6.03 to -3.37; 690 participants), both with low-certainty evidence. We are very uncertain about the effect of screening in a maternity healthcare setting versus a non-maternity healthcare setting on diagnosis of gestational diabetes or diabetes in pregnancy (RR 1.10, 95% CI 0.60 to 2.01; 690 participants; very low-certainty evidence) and length of neonatal hospitalisation (MD 0.04, 95% CI -0.57 to 0.65; 690 participants; very low-certainty evidence). No direct evidence was available on screening for diabetes in pregnancy as distinct from GDM. No data were reported on maternal emergency department visits, maternal antenatal hospitalisation, length of maternal hospitalisations, and costs. Maternal outcomes We are very uncertain about the effect of screening in a maternity healthcare setting versus a non-maternity healthcare setting on pre-eclampsia (RR 1.24, 95% CI 0.56 to 2.75; 690 participants; very low-certainty evidence). The following outcomes were not reported: eclampsia, HELLP syndrome, maternal death, admission to intensive care, hyperglycaemia requiring pharmacological treatment, infections, and postpartum glucose intolerance. Child-related outcomes We are very uncertain about the effect of screening in a maternity healthcare setting versus a non-maternity healthcare setting on: stillbirth or foetal death (RR 0.91, 95% CI 0.08 to 9.93), admission to neonatal intensive care (RR 1.01, 95% CI 0.59 to 1.71), preterm birth (RR 1.06, 95% CI 0.41 to 2.71), neonatal hypoglycaemia (RR 0.91, 95% CI 0.23 to 3.59), large-for-gestational-age (RR 0.73, 95% CI 0.51 to 1.04), shoulder dystocia (RR 0.15, 95% CI 0.01 to 3.70), and birth trauma (RR 0.91, 95% CI 0.08 to 9.93), all with 690 participants each and very low-certainty evidence. An effect estimate for neonatal death was not estimable (maternity setting: 0/475; non-maternity setting: 0/215). AUTHORS'

conclusionsThere is insufficient evidence to determine whether screening for gestational diabetes or diabetes in pregnancy in maternity healthcare settings compared with non-maternity healthcare settings improves health service, maternal, or child-related outcomes. Low-certainty evidence from a single study indicates that screening in maternity healthcare settings may result in a reduction in time to access antenatal diabetes care and may increase compliance. The effects on other outcomes remain unknown. Conducting studies using intention-to-treat analyses could bridge existing research gaps, enhancing effect estimates and certainty.

fundingThis Cochrane review was funded by the World Health Organization (WHO) and the UNDP-UNFPA-UNICEF-WHO-World Bank Special Programme of Research, Development and Research Training in Human Reproduction (HRP). REGISTRATION: Protocol available via PROSPERO: https://www.crd.york.ac.uk/PROSPERO/view/CRD420251108692.