Gestational diabetes mellitus (GDM) is a common pregnancy complication associated with adverse short- and long-term outcomes for mother and child. Major international bodies recommend universal GDM screening—commonly a one-step 75 g oral glucose tolerance test (OGTT) at 24–28 weeks' gestation. Despite these recommendations, implementation in low- and middle-income countries, including many sub-Saharan African settings, often remains selective, delayed, or absent. This study audited routine antenatal care (ANC) records from three tertiary public hospitals in Addis Ababa to generate context-specific evidence on GDM screening uptake and adherence to guideline-recommended practice.
A facility-based, cross-sectional retrospective chart review was conducted at three public hospitals in Addis Ababa (Tikur Anbessa Specialized Hospital, Gandhi Memorial Hospital and Abebech Gobena MCH Hospital) between March and May 2025. These tertiary hospitals were purposively selected for their high ANC volume and provision of maternal health services.
Eligible records included pregnant women with at least two ANC visits at the same facility who were either ≥28 weeks' gestation or had delivered at the participating hospitals within the study period. Women with pre-existing type 1 or type 2 diabetes were excluded. Rather than sampling, the study used a census of all eligible electronic medical records during the period, totaling 1,050 records (350 per facility).
Trained data collectors abstracted socio-demographic, obstetric, and GDM screening data using a structured checklist. The primary outcome was evidence of any glycemic testing during pregnancy (OGTT, fasting blood sugar [FBS], random blood sugar [RBS], or hemoglobin A1c). Secondary outcomes included adherence indicators: who was screened, timing of screening relative to the guideline window, and the screening test used. Laboratory information systems were cross-checked to corroborate test requests and results.
Data were collected with ODK, validated and exported to SPSS 25. Continuous variables were summarized with means (SD) or medians (IQR); categorical variables with frequencies and percentages. χ² tests compared facility-level proportions. Binary logistic regression estimated crude and adjusted odds ratios (COR, AOR) with 95% confidence intervals for factors associated with being screened. Missing values for some continuous covariates were imputed using facility-specific medians or means. The Hosmer–Lemeshow test indicated acceptable model fit.
Of 1,050 antenatal records reviewed, 139 documented any GDM screening, for an overall uptake of 13.2%. Screening rates varied significantly across facilities (range 5.1%–18.0%). The median gestational age at screening was 34 weeks (IQR 30–37); only 12 of 139 screened women (8.6%) were tested within the recommended 24–28 weeks window. The one-step 75 g OGTT was the predominant screening method (122/139; 87.8%). Nearly half (65/139; 46.8%) underwent more than one test. Guideline-concordant screening—OGTT performed at 24–28 weeks—was documented in 11/139 (7.9%). Screening practice was almost exclusively risk-based (99.3%), with universal screening documented in only one case (0.7%).
Among the 139 women who were screened, 57 (41.0%) were diagnosed with GDM. Screening triggers that showed higher proportions of GDM diagnosis included polyhydramnios, prior GDM, previous stillbirth, and pre-pregnancy overweight/obesity. For example, polyhydramnios was the most frequently recorded indication for testing (58/139; 41.7%). Management after diagnosis was split: approximately half of diagnosed women (49.1%) were initiated on pharmacotherapy, and the remainder received lifestyle-based management.
In multivariable analysis, facility site was strongly associated with screening. Compared with facility A, women at facility B had significantly lower odds of being screened (AOR 0.23; 95% CI 0.12–0.42), while those at facility C had higher odds (AOR 2.73; 95% CI 1.36–5.46). Older maternal age increased the odds of screening (AOR 1.12 per additional year; 95% CI 1.07–1.16). Multiparity was associated with higher odds of screening compared with nulliparity (AOR 2.83; 95% CI 1.63–4.92). Earlier entry to ANC was also independently associated with higher likelihood of screening: each additional week delay in first ANC contact reduced the odds of screening by about 5% (AOR 0.95 per week; 95% CI 0.92–0.99). Gravidity was not included in the multivariable model due to collinearity with parity. Number of ANC visits showed a negative crude association but was not independently associated with screening in adjusted models.
This audit in three tertiary public hospitals in Addis Ababa found low overall GDM screening uptake (13.2%), predominately risk-based testing, and substantial delays in testing relative to guideline recommendations. Although the OGTT was the principal test when screening occurred, only a small fraction of women were tested within the recommended 24–28 weeks window. High diagnostic yield among the screened (41.0% diagnosed) likely reflects selection of higher-risk women for testing under a risk-based approach. Facility-level differences and patient characteristics—older age, multiparity, and earlier ANC initiation—were associated with higher screening rates, suggesting both health-system and patient-level determinants influence implementation.
These findings align with prior reports from sub-Saharan Africa describing selective screening, variable methods, and delayed testing, and they underscore the implementation gap between international recommendations for universal, timely screening and routine practice in resource-constrained settings.
In the three audited public hospitals in Addis Ababa, GDM screening was infrequent, delayed, and largely risk-based despite guideline recommendations for universal screening using a 75 g OGTT at 24–28 weeks. Screening uptake varied by facility and was associated with older maternal age, multiparity, and earlier ANC initiation. The results support the need for facility-specific implementation strategies and interventions addressing patient- and system-level barriers to achieve universal and timely GDM screening in routine antenatal care.
Note: The source article provides detailed tables and statistical estimates referenced in the Results and modeling sections. Additional implementation recommendations and broader contextual discussion beyond the reported results were not included in the source and therefore are not reported here.