Ensuring the safety of medicines during pregnancy is really important, but it's tricky because pregnant women are often left out of clinical trials. This makes it hard to figure out how drugs might affect them. A recent study tried to solve this problem by using advanced AI and machine learning to predict potential risks. They looked at real-world data from the FDA to see which drugs might cause adverse reactions during pregnancy.

The researchers explored different methods to classify drugs into high or low risk categories. They compared traditional machine learning models with more complex ones, like Directed Message Passing Neural Networks (DMPNN). The DMPNN model, which combines information from the drug's molecular structure and other details, showed the best performance.

But how do we know it works? They tested the model using data from other sources like SIDER and DailyMed. It passed with flying colors, showing it could accurately predict risks in new scenarios. The researchers also checked 22 oral hypoglycemic drugs and found potential red flags for pregnancy-related side effects.

This study suggests that AI can be a powerful tool to make drug safety assessment much better. It gives doctors and pregnant women important information to decide on the safest medications.

Predicting Pregnancy Drug Risks: Can AI Help?

Actions