Jason Street
Decision Analysis
11/19/2014
Merck KL-798 (A) Case

I made a few assumptions in creating my tree based on the reading for this case. Firstly, I set the initial investment for Merck in the project to be $35 million, which consists of the $30 million licensing payment plus the $5 million cost to Merck for Phase I. Secondly, we were given that Phase I had a 60% chance of success, so I show the other 40% as losing all money invested in the project at that point. Similarly for Phase II, there is a 50% chance of losing the $40 million invested in it. Based on the initial analysis in the tree shown in Exhibit 1, I found the expected value of purchasing KL-798 to be -$1.16 million, so I would not recommend pursuing this investment. If we relax the assumption that we will always follow Mr. Merck’s advice and pursue the investment even if it doesn’t provide some medical benefits, then we see a much large expected value of $104.5 million (Exhibit 2). The way I depicted this decision was by turning Phase 2 into a decision node, so we would be able to choose the most likely outcome from Phase 2. I performed a sensitivity analysis on the first decision tree that yielded an expected value of -$1.16 million. I was not sure about how to use all the probabilities as inputs for the sensitivity analysis (particularly where we needed to account for 4 probabilities that need to sum up to 1). Out of the inputs I was able to analyze, I found that the highest sensitivity lies in whether the Phase I trial is successful. This makes sense since there’s just a 60-40 chance of achieving success in Phase I, so with much higher probabilities the expected value of investing should increase greatly with it. The next highest sensitivity is if testing for KL-798 goes to Phase III and is FDA approved for obesity only. Lastly, I looked at the value of perfect information if we have Foresight Consulting predict