Objective: To develop and validate a prediction model for adverse pregnancy outcomes at the time of gestational diabetes (GDM) diagnosis to stratify care.

Research Design and Methods: A prediction model development and validation study was conducted on data from a retrospective population-based study. The outcome to be predicted was a composite for adverse pregnancy outcome: hypertensive disorder of pregnancy, large-for-gestational (LGA) age neonate, neonatal hypoglycaemia requiring intravenous therapy, shoulder dystocia, perinatal death, neonatal bone fracture and nerve palsy.

Results: Predictors of the composite adverse pregnancy outcome were: fasting and 1 hour glucose from the diagnostic OGTT, gestational age of GDM diagnosis, previous macrosomia, previous pre-eclampsia or eclampsia, previous LGA, ethnicity, weight at GDM diagnosis and parity. The apparent performance of the model was reasonable: c-statistic was 0.691 (95% CI, 0.662 - 0.720) and the calibration plot showed near perfect agreement between predicted and observed risks (Figure 1).

Conclusions: A promising prediction model for adverse pregnancy outcomes in women with GDM was developed. Model estimation using the Least Absolute Shrinkage and Selection Operator method is currently underway to correct for over-optimism. External validation and decision curve analysis will determine suitability for clinical application.


S. Cooray: None. J. Boyle: None. G. Soldatos: None. J. Zamora: None. B.M. Fernandez-Felix: None. J. Allotey: None. S. Thangaratinam: None. H. Teede: Other Relationship; Self; Roche Diagnostics France.


National Health and Medical Research Council of Australia (APP1151242); Australian Academy of Science; Diabetes Australia Research Program; Medical Research Future Fund of Australia; Australian Government Department of Education and Training (7167_2019)

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at http://www.diabetesjournals.org/content/license.