The performance of marginal structural models for estimating risk differences and relative risks using weighted univariate generalized linear models
Austin PC. Stat Methods Med Res. 2024; Apr 24 [Epub ahead of print].
Aims — To determine the test characteristics of algorithms using hospitalization and physician claim data to predict gestational diabetes (GDM).
Methods — Using population-level healthcare administrative data, we identified all pregnant women in Ontario in 2019. The presence of GDM was determined based on glucose screening laboratory results. Algorithms using hospitalization records and/or physician claims were tested against this gold standard. The selected algorithm was applied to administrative data records from 1999 to 2019 to determine GDM prevalence in each year.
Results — Identifying GDM based on either a diabetes mellitus code on the delivery hospitalization record, OR at least 1 physician claim with a diabetes diagnosis code with a 90 day lookback before delivery yielded a sensitivity of 95.9%, specificity of 99.2%, and positive predictive value of 87.6%. The prevalence of GDM increased from 4.2% of pregnancies in 1999 to 12.0% in 2019.
Conclusions — Algorithms using hospitalization or physician claims administrative data can accurately identify GDM
Shah BR, Booth GL, Feig DS, Lipscombe LL. Can J Diabetes. 2023; 47(1):25-30. Epub 2022 Jul 9.
The ICES website uses cookies. If that’s okay with you, keep on browsing, or learn more about our Privacy Policy.