Comparison of Rcausal and Tetrad

• 

  FGS in Rcausal: Drug Y1 does not causal for X2, X3, X4, and X5

• 

  FGS in tetrad: Drug Y1 does not causal for X2, X3, X4, and X5

• 

  FCI in tetrad: Drug Y1 does not causal for X2, X3, X4, and X5

• 

  FCI in Tetrad:Drug Y1 does not causal for X2, X3, X4, and X5

• 

  PC in Tetrad: Drug Y1 does not causal for X2, X3, X4, and X5

Inspiration

Adverse drug events are confounded by 1. patients using multiple drugs 2. Multiple drugs causing the same event 3. Adverse events and morbidities having the same phenotype

What it does

Compare Rcausal and Tetrad using the same data

How I built it

Processed data and ran algorithms

Challenges I ran into

Processing data Rcausal takes away R features such as tetrad

Accomplishments that I'm proud of

Getting data into R without clipboard

What I learned

Causality can not be determined with blinded EMRs. Adverse events had more relationships to each other than the drug being evaluated. There may need to be temporal and historical data as well as natural language processing.

What's next for Comparison of Rcausal and Tetrad

Evaluating a larger dataset and adding temporal sequencing with the dataset

Built With

• rmarkdown