1. The Role of Simulation in Representing Real World Risk In BiopharmaR&DPortfolio Optimization Davis Walp Head, Value Based Solutions Quintiles
2. Disclaimer The views and opinions expressed in the following PowerPoint slides are those of the individual presenter and should not be attributed to Drug Information Association, Inc. (“DIA”), its directors, officers, employees, volunteers, members, chapters, councils, Special Interest Area Communities or affiliates, or any organization with which the presenter is employed or affiliated. These PowerPoint slides are the intellectual property of the individual presenter and are protected under the copyright laws of the United States of America and other countries. Used by permission. All rights reserved. Drug Information Association, Drug Information Association Inc., DIA and DIA logo are registered trademarks. All other trademarks are the property of their respective owners. 2
3. 3 6 Drug F Drug H eNPV Drug A Drug D 3 Drug I Drug C Drug B Drug E Drug G - 40 Pr(Successful Launch) The Old-Fashioned Approach to Portfolio Construction Set an eNPVthreshold, and exclude any compounds that fall below Drug A and Drug F are the only candidate compounds with positive ENPVs eNPV> 0 eNPV< 0
4.
5. The Solution Create a statistical model of the portfolio that will enable simulation of various changes to the portfolio and the resulting impact of portfolio events on the likelihood that the portfolio will drive business objectives 6
9. Commercial PerformanceTake random draws of all uncertain variables and observe the model outcome (iteration) Repeat numerous times (1,000+) to assure a sufficiently large number of observations Characterize the range and distribution of outcomes observed
10. Example: Time-in-Phase Simulation and Impact on Launch Assumption 8 Phase I Phase II Phase III Phase I Phase II Phase III Drug Drug Traditional Time In Phase Time In Phase Time In Phase Time In Phase Time In Phase Time In Phase Time In Phase Drug A 24 mo 30 mo 42 mo Drug A 24 mo 30 mo 42 mo Point Estimation Each iteration draws randomly Frequency Frequency Frequency Frequency Frequency Frequency from a probability distribution Time - in - Phase (days) Time - in - Phase (days) Time - in - Phase (days) Time - in - Phase (days) Time - in - Phase (days) Time - in - Phase (days) Iteration Phase 1 TIP Phase 2 TIP Phase 3 TIP Iteration Phase 1 TIP Phase 2 TIP Phase 3 TIP Simulation of 1 19 31 39 1 18 31 38 Time In Phase Results to 2 26 28 47 2 26 29 47 Reflect Real 3 23 34 40 3 23 34 40 World Variability 4 27 28 42 4 27 28 43 … … 10,000 Mode = 24 Mode = 30 Mode = 42 10,000 Mode = 24 Mode = 30 Mode = 42 Illustration of time in phase simulation for a single product Launch Expectation T0 T0- 2mo. T0+5mo T0+1mo T0+1mo Each iteration of the simulation will feature a different launch date and total development cost. Total portfolio composition will also vary based on simulated project failures
11. Differing Development Timelines Mean Different Commercial Trajectories 9 Early MM $3000 $2000 $1000 Illustrative On time Late 2015 2017 2019 2021 2023 2025
12. 10 Drug A 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 Drug B Revenue 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 Drug C Development Costs Revenue 2012 2013 2014 2015 2016 2017 2018 2019 2020 COGS Development Costs Revenue Promotional Spend COGS Development Costs Net Income Promotional Spend COGS Net Income Promotional Spend Cash Flow Cash Flow Net Income NPV NPV Cash Flow NPV Distribution (ex.) Range of Potential Outcomes Uncertain Variables Distribution (ex.) Range of Potential Outcomes Uncertain Variables Bernoulli Discrete: Success (1) or Failure (0) Clinical Trial Outcome Bernoulli Discrete: Success (1) or Failure (0) Clinical Trial Outcome Triangular Continuous: Any positive length of time Time - in - Phase Triangular Continuous: Any positive length of time Time - in - Phase Triangular Continuous: Any positive dollar value Development Costs Triangular Continuous: Any positive dollar value Development Costs Triangular Continuous: Any non - negative dollar value Commercial Performance Triangular Continuous: Any non - negative dollar value Commercial Performance Simulation Enables Sophisticated Metrics for Evaluating Tradeoffs Probability that Company Will Meet or Exceed a Specific Revenue Threshold in a Particular Year Company Portfolio + Inline Brands Company P&L Probability that Company Will Launch x NMEs in a Given Year or Timeframe Probability that Company Will Deliver A Certain EPS Target to Company in a Specific Year Management can better gauge the value of an in-license opportunity by evaluating how business performance metrics change when the project becomes part of the portfolio
14. Expected Return / NPV Volatility (Risk) . . . Enables Tradeoff Analysis in the Context of Portfolio Risk and Return illustrative Key Questions Efficient Frontier (theoretical) Is a large increase in expected returns worth a significant increase in risk burden? Scenario A: Add a Risky Program to Existing Portfolio Existing Company Portfolio
15. Expected Return / NPV Volatility (Risk) . . . Enables Tradeoff Analysis in the Context of Portfolio Risk and Return illustrative Key Questions Efficient Frontier (theoretical) Is a large increase in expected returns worth a significant increase in risk burden? Scenario A: Add a Risky Program to Existing Portfolio Existing Company Portfolio Is a large reduction in portfolio risk worth a marginal reduction in portfolio ENPV? Scenario B: Add a Low Potential Program to Existing Portfolio
16. Other Comments Key takeaway: Portfolio prioritization by ranking ENPVs of individual compounds will NOTgenerate an optimal portfolio! ENPV analysis of individual compounds is a critical step and a necessary, but not sufficient, step for any total portfolio analysis Portfolio optimization takes individual compound assessment one step further Portfolio management via diversification is not a novel concept Arguably counter-intuitive that portfolio fitness can be improved by including negative-ENPV compounds Incorporating portfolio interdependencies, conditional success rates, etc. will exacerbate the insufficiency of static portfolio prioritization Simulation based portfolio optimization is a strategic exercise, but supported by a quantitative and robust foundation 14
