Accuracy of clinical risk factor-based models as a screening test for detecting gestational diabetes mellitus in a low-resource setting

TL;DR

This study evaluates how well clinical risk models can screen for gestational diabetes in a low-resource setting, aiming to reduce the need for glucose tests.

Contribution

The study provides new evidence on the accuracy of clinical risk models for gestational diabetes screening in a Nigerian population.

Findings

Three clinical models showed high sensitivity but low specificity for detecting gestational diabetes.

Negative predictive values were high, suggesting models can safely identify low-risk women.

Gestational diabetes prevalence was 19% using IADPSG/WHO criteria in the study population.

Abstract

Screening and diagnosing gestational diabetes mellitus (GDM) usually requires
a 2-hour, 75 g oral glucose tolerance test (OGTT), which can be challenging
for both patients and healthcare systems. Alternative clinical risk
factor-based models have been suggested but have not been extensively
tested, particularly in low-resource countries. This study aimed to evaluate
the accuracy of these risk factor-based models as screening tools.

This prospective cohort study involved 400 consenting pregnant women
receiving antenatal care in Lagos, Nigeria. Participants were evaluated for
GDM risk using three clinical models and underwent universal screening and
diagnosis at 24 to 28 weeks with a single-step, 2-hour 75g OGTT, using
IADPSG/WHO criteria. The Receiver Operating Characteristic (ROC) curve was
used to assess the accuracy of the risk factor-based models.

The mean age of the subjects was 31.0 ± 5.3 years. The prevalence of
GDM, according to the IADPSG/WHO 2013 criteria, was 19.0%. Using the
clinical risk score models developed by Naylor and cols., Caliskan and
cols., and Phaloprakarn and cols., positive risk scores for GDM were found
in 85%, 67.3%, and 93.8% of subjects, respectively. The sensitivity,
specificity, and accuracy of these models ranged from 71.1% to 96.1%, 6.7%
to 33.6%, and 23.8% to 40.8%, respectively. However, the negative predictive
values were relatively high, ranging from 83.2% to 88%.

The clinical risk factor-based prediction models evaluated in this study may
effectively identify women at low risk for GDM who can be exempted from the
2-hour OGTT.

Figures