The structure of Margolin’ decision to continue development

下面为大家整理一篇优秀的assignment代写范文- The structure of Margolin’ decision to continue development，供大家参考学习，这篇论文讨论了马那克公司开发吡非尼酮。马那克公司建议开发吡非尼酮，因为他们可以使利润最大化。根据FDA的程序，让InterMune授权使用该药物是明智的，这可以降低Marnac的风险。另一方面，即使风险很高，InterMune也能获得可观的预期利润。

Introduction

This report is proposed to develop a decision tree for is the structure of Margolin’ decision to continue development. Decision tree is a flow chart structure with the probability. The structure of the decision tree has the following components [1-3]:

(1) Decision nodes, depicted by squares;

(2) Chance nodes, depicted by circles; and

(3) End nodes, depicted by triangles.

Decision represents the linearized decision rules [2-7], and the result is the leaf of the tree.

Methodology

Originally, the structure of Margolin’s decision to continue development of the non-steroid pirfenidone vs. the steroid can be drawn as a decision tree, shown as follows.

According to the literature [8-15], as the plot shows, it is a tree structure, and there are three decisions in the decision pross. For the steroid, the expected value is 11%*1125.9 + 89%*(-125.1) = 12.51 M, while the expected of the pirfenidone is 14.85%*1515 + 85.15%*(-247.5) = 14.23125 M. Accordingly, the pirfenidone has the higher expected value, while steroid has the higher risk.

The Marnac has handle the risk by generating more safety data. If the cash flow is limited and 200 M is the maximum sustainable value, then Marnac should choose producing steroid because the cost is 125.1 M if it is not approved.

Based on the decision tree, without considering the cash flow limit, Marnac will definitely choose pirfenidone because it has higher expected value and lower risk.

When using the FDA table, I distribute the approval cost to the IND and three phases based on their proportion. The decision tree is shown as follows.

According to the decision tree, there are total 11 decisions in the process. The risk Marnac faces is the lower probability for approval, compared with the previse method (0.08 and 0.14). If some perfect information could be added in the decision tree, such as gather more safety data in the experiment, the Marnac can reduce the risk. When adding this in the decision tree, the probability to pass Phase II may be higher.

The Marnac needs to talk to InterMune because it saw the probability for passing the Phase II is too low, and the expected value is not significant, compared with the original status.

When InterMune considers whether to license pirfenidone, the decision tree is shown as follows (suppose the probability of repeated Phase II is 0.48, same as FDA table)

The expected value is 37.5 M. So, it is suggested the InterMune license pirfenidone. To maximum the profit, the InterMune should level up the probability in the repeated Phase II trial.

The expected money the InterMune will pay is 206.295. While the actual money InterMune made for this is 52.5+18.8 as 71.3, which is truly determined.

The risk of InterMune of licensing pirfenidone is at the repeated Phase II. Since each time for the Phase II experiment, the cost is pretty larger than the expected value, and its probability is relatively lower. If the probability cannot reach the level that FDA satisfies, it will definitely lose money.

The probability of losing money for InterMune is 0.91. It is a high probability. The reason why Marnac and KDL were willing to have InterMune take over the FDA is that the money the InterMune gave to them is likely larger than the one under FDA.

Conclusion

This report is designed to explore the development strategy of Marnac as which kind of drug they should produce. Originally, according to their experience, it might produce pirfenidone rather than steroid since the expected value of pirfenidone is higher and the risk is lower. When using FDA’s new table, Marnac may face a even lower expected value and higher risk, so it is suggested that let InterMune license pirfenidone. According to the decision tree of licensing pirfenidone, it is recommended that InterMune license pirfenidone since the expected value is higher.

Under the FDA process, the probability of approval is 8%, much lower than the original process. To mitigate the risk, the Marnac needs to collect more information to ensure that their experiment could be successful during each phase. For example, they can revise their sampling method, amplify sample size, and correct thee statistic stragety.

According to the report, the Marnac and is recommended to develop the Pirfenidone because they can maximize the profit. Under the FDA process, it is a smart to let InterMune to license the drug, which could lower down the risk of Marnac. On the other hand, InterMune could get a significant expected profit even the risk is high.

Reference

1. Kamiński, B.; Jakubczyk, M.; Szufel, P. (2017). "A framework for sensitivity analysis of decision trees". Central European Journal of Operations Research. doi:10.1007/s10100-017-0479-6.

2. Jump up^ Quinlan, J. R. (1987). "Simplifying decision trees". International Journal of Man-Machine Studies. 27 (3): 221. doi:10.1016/S0020-7373(87)80053-6.

3. Jump up^ K. Karimi and H.J. Hamilton (2011), "Generation and Interpretation of Temporal Decision Rules", International Journal of Computer Information Systems and Industrial Management Applications, Volume 3

4. Jump up^ Wagner, Harvey M. (1975-09-01). Principles of Operations Research: With Applications to Managerial Decisions (2nd ed.). Englewood Cliffs, NJ: Prentice Hall. ISBN 9780137095926.

5. Jump up^ R. Quinlan, "Learning efficient classification procedures", Machine Learning: an artificial intelligence approach, Michalski, Carbonell & Mitchell (eds.), Morgan Kaufmann, 1983, p. 463-482. doi:10.1007/978-3-662-12405-5_15

6. Jump up^ Utgoff, P. E. (1989). Incremental induction of decision trees. Machine learning, 4(2), 161-186. doi:10.1023/A:1022699900025

7. Jump up^ Deng,H.; Runger, G.; Tuv, E. (2011). Bias of importance measures for multi-valued attributes and solutions (PDF). Proceedings of the 21st International Conference on Artificial Neural Networks (ICANN).

8. Klein G (2003). The Power of Intuition. New York: Doubleday. ISBN 0-385-50289-3.

9. Jump up^ Klein G (1999). Sources of Power. Boston, MA: MIT Press. ISBN 0-262-11227-2.

10. Jump up^ Keeney R (2002). Value Focused Thinking: A Path to Creative Decisionmaking. ISBN 0-674-93197-1.

11. Jump up^ Robyn M. Dawes & Bernard Corrigan (1974). "Linear Models in Decision Making". Psychological Bulletin. 81 (2): 93–106. doi:10.1037/h0037613.

12. Jump up^ B. Fischhoff; L. D. Phillips & S. Lichtenstein (1982). "Calibration of Probabilities: The State of the Art to 1980". In D. Kahneman & A. Tversky. Judgement under Uncertainty: Heuristics and Biases. Cambridge University Press.

13. Jump up^ Kane, Becky (8 July 2015). "The Science of Analysis Paralysis: How Overthinking Kills Your Productivity & What You Can Do About It". Todoist Blog. Retrieved 14 May 2016.

14. Goldstine, Herman (1972). The Computer from Pascal to Von Neumann. Princeton University Press. pp. 266–267. ISBN 0-691-08104-2.

15. Jump up^ Taub, Abraham (1963). John von Neumann Collected Works. 5. Macmillan. pp. 80–151.

51due留学教育原创版权郑重声明：原创assignment代写范文源自编辑创作，未经官方许可，网站谢绝转载。对于侵权行为，未经同意的情况下，51Due有权追究法律责任。主要业务有assignment代写、essay代写、paper代写服务。

51due为留学生提供最好的assignment代写服务，亲们可以进入主页了解和获取更多assignment代写范文 提供北美作业代写服务，详情可以咨询我们的客服QQ：800020041。