Parental comorbidity and medication use in the USA: a panel study of 785 000 live births

Pharmacoepidemiology and costs of medications dispensed during pregnancy: A retrospective population-based study

OBJECTIVE

To describe the pharmacoepidemiology and costs associated with medications dispensed during pregnancy.

DESIGN

Pharmacoepidemiological study and cost analysis.

SETTING

Queensland, Australia.

POPULATION

All women who gave birth in Queensland between January 2013 and June 2018.

METHODS

We used a whole-of-population linked administrative dataset, Maternity1000, to describe medications approved for public subsidy that were dispensed to 255 408 pregnant women. We describe the volume of medications dispensed and their associated costs from a Government and patient perspective.

MAIN OUTCOME MEASURES

Prevalence of medication use; proportion of total dispensings; total medication costs in AUD 2020/21 ($1AUD = $0.67USD/£0.55GBP in December 2022).

RESULTS

During pregnancy, 61% (95% CI 60.96-61.29%) of women were dispensed at least one medication approved for public subsidy. The mean number of items dispensed per pregnancy increased from 2.14 (95% CI 2.11-2.17) in 2013 to 2.47 (95% CI 2.44-2.51) in 2017; an increase of 15%. Furthermore, mean Government cost per dispensing increased by 41% from $21.60 (95% CI $20.99-$22.20) in 2013 to $30.44 (95% CI $29.38-$31.49) in 2017. These factors influenced the 53% increase in total Government expenditure observed for medication use during pregnancy between 2013 and 2017 ($2,834,227 versus $4,324,377); a disproportionate rise compared with the 17% rise in women's total out-of-pocket expenses observed over the same timeframe ($1,880,961 versus $2,204,415).

CONCLUSIONS

Prevalence of medication use in pregnancy is rising and is associated with disproportionate and rapidly escalating cost implications for the Government.

Multi-Drug Featurization and Deep Learning Improve Patient-Specific Predictions of Adverse Events

While the clinical approval process is able to filter out medications whose utility does not offset their adverse drug reaction profile in humans, it is not well suited to characterizing lower frequency issues and idiosyncratic multi-drug interactions that can happen in real world diverse patient populations. With a growing abundance of real-world evidence databases containing hundreds of thousands of patient records, it is now feasible to build machine learning models that incorporate individual patient information to provide personalized adverse event predictions. In this study, we build models that integrate patient specific demographic, clinical, and genetic features (when available) with drug structure to predict adverse drug reactions. We develop an extensible graph convolutional approach to be able to integrate molecular effects from the variable number of medications a typical patient may be taking. Our model outperforms standard machine learning methods at the tasks of predicting hospitalization and death in the UK Biobank dataset yielding an R² of 0.37 and an AUC of 0.90, respectively. We believe our model has potential for evaluating new therapeutic compounds for individualized toxicities in real world diverse populations. It can also be used to prioritize medications when there are multiple options being considered for treatment.