Given a dataset with multiple likelihood ratios (LRs) for predicting Heart Failure:
- Previous Heart Failure = LR 2
- Classical Exam Findings = LR 4
- Previous Hypertension = LR 6
Assuming LR(Previous Heart Failure) and LR(Classical Exam Findings) are independent, I understand that a combined LR can be obtained through simple multiplication (i.e. 2 x 4).
However, LR(Previous Hypertension) and LR(Previous Heart Failure) cannot necessarily be deemed independent as Hypertension is a risk factor for Heart Failure.
Is their a method to combine LR(Previous Heart Failure) and LR(Previous Hypertension) into a similar combined LR?
We have a large amount of clinical data to calculate any additional data that is needed to combine dependendent LRs, but I have no idea how to approach this.
The end application is too generate a patient specific LR by combining the LRs from multiple variables (such as hypertension and heart failure).
Thanks