Challenge
Our client wanted to facilitate medication monitoring through the use of AI. The aim was to identify potential safety signals for heterogeneous patient populations in order to detect any drug-related side effects at an early stage and further improve patient safety.
Approach
We employed causal machine learning (ML) to detect heterogeneous drug-related adverse events. Our approach included:
- Using causal ML to estimate counterfactual values and effect heterogeneity
- Considering cofounders and mediators in the analysis
- Conducting a prototyping process to identify potential safety signals
This approach would allow us to analyze complex drug interactions in real-world patient populations and identify potentially dangerous combinations.
Results
The implementation of such an AI solution can lead to the following results:
- Successful estimation of counterfactual values and effect heterogeneity through causal ML
- Detection of a potential safety signal during prototyping
- Consideration of cofounders and mediators to improve analysis accuracy