To address the evidence gaps in drug safety for pregnant women due to their exclusion from clinical research, emphasizing the importance for both maternal and fetal health.
Key Findings:
79% of women take at least one medication during pregnancy without sufficient evidence, with specific examples of medications.
Only 4% of clinical trials in the past decade included pregnant women, illustrating the scale of the issue.
Machine learning can help identify risk-stratified patients and improve drug safety recommendations, with examples of successful applications.
Interpretation:
The use of machine learning and large datasets can potentially close the evidence gap for medication safety in pregnant women, but requires careful integration with causal inference methods to ensure clinical relevance.
Limitations:
Dependence on large datasets which may not always be available, and potential biases in data collection.
Risks associated with 'black box' models that lack transparency in decision-making, which could undermine trust in findings.
Conclusion:
Addressing the exclusion of pregnant women from clinical trials is crucial for improving medication safety and outcomes, and immediate action is needed to implement these findings in clinical practice.
