How PrecisionTree Helped a Biopharmaceutical Company Decide between Multimillion Dollar Investments
- Industry: Pharmaceutical
- Product(s): PrecisionTree
- Application: Using Decision Trees in Biopharmaceutical investment decision-making
While employed at a major Biopharmaceutical company facing a choice between competing product investment strategies, Richard Bayney and his colleagues used Palisade’s PrecisionTree software to construct a multi-phase decision tree to bring clarity to the analysis and help facilitate a complex decision.
While many software tools can help solve complex decision problems, I have found Palisade’s DecisionTools Suite to be highly adaptable to the idiosyncratic drug development and selection problems I encounter in the Biopharmaceutical industry.Richard Bayney, Project and Portfolio Value Creation
The Biopharmaceutical industry faces expensive, complex, multi-phase decision-making, punctuated by high technical and regulatory risks as well as uncertain commercial outcomes. Not surprisingly, data gathering and validation become the focal point for much discussion and debate, especially when an analysis of decision alternatives between competing investments results in a marginally dominant policy suggestion. Invariably, decision-makers request more data – not all of which may be decision-relevant – resulting in protracted and inefficient decision-making.
In this vein, the challenge to transparent, defensible decision-making under conditions of risk and uncertainty rests in testing the sensitivity of data ranges to a dominant policy suggestion. In some cases, incrementally small changes to assumptions and data result in a different policy suggestion, while in other cases, a policy suggestion may remain unchanged by large variances in data. In any event, decision trees facilitate good discussion around phase-specific investments and risks as well as uncertain commercial value, and enable decisions to be taken on the basis of phase-specific, risk-adjusted value. The most popularly used metric for such decisions in the Biopharmaceutical industry is risk-adjusted Net Present Value (eNPV).
In the case discussed here, a major Biopharmaceutical company faced a decision to invest $325 million in a lead molecule (ABC) for the treatment of Alzheimer's Disease (AD), or $550 million in a different lead molecule (XYZ) for the treatment of Mild Cognitive Impairment (MCI) and General Anxiety Disorder (GAD). NB: To preserve the confidentiality of the project, the targeted diseases along with their investment profiles have been altered significantly.
Both compounds had clinical and regulatory risks associated with their development and regulatory approval, and because of the rapidly changing competitive environment, there were huge commercial uncertainties associated with their market uptake.
Determining (and reinforcing) the dominant strategy
With the use of Palisade's PrecisionTree, Bayney and his colleagues constructed a multi-phase decision tree for each decision alternative. The strategies mapped out included failure and success for each compound ABC (for AD) and XYZ (for MCI + GAD). The inputs used included information and data regarding probabilities of clinical and regulatory success of the various outcomes of each strategy, probabilities of high and low success associated with uncertain commercial outcomes, and commercial values associated with uncertain commercial success (Figure 1). These data and assumptions were elicited from the respective Subject Matter Experts (SMEs) within the Project Team.
“While the policy suggestion in favour of compound XYZ targeting MCI + GAD was easy enough to determine on the basis of its higher eNPV, i.e. $290M versus $245M for the competing decision alternative of compound ABC in AD (Figure 1), the company was still hesitant to commit to a decision until a thorough examination of the sensitivity of the policy suggestion to combined clinical and regulatory risks and commercial uncertainties was conducted,” Bayney says. “This sensitivity analysis helped the company in its belief in the SME judgments and, afterwards, in its deliberations.”
One such analysis (Figure 2) was conducted on high AD commercial success and high MCI commercial success to determine the breakpoints in each commercial range, i.e. the point at which a rational decision-maker would be indifferent between the 2 decision alternatives. Other sensitivity analyses (not shown) included searching for decision breakpoints between (a) technical and regulatory risks associated with AD and MCI + GAD as well as (b) technical and regulatory risks and commercial uncertainties associated with the 3 diseases.
With the aid of PrecisionTree, the analysis demonstrated the following:
- How to dissect a complex investment decision in a logically consistent manner.
- How to conduct a defensible project and program evaluation.
- How to challenge SME judgment and evaluate the sensitivity of the dominance of a policy suggestion to risks and uncertainties.
- How to facilitate good decision-making by striving for approximate data correctness as opposed to absolute precision.
Bayney used other decision tree software earlier in his career, and then switched to PrecisionTree and the DecisionTools Suite for their simplicity, flexibility, and user-friendliness. He employs other tools in the DecisionTools Suite such as Evolver, @RISK, and RISKOptimizer in his consulting and teaching work.
“While many software tools can help solve complex decision problems, I have found Palisade’s DecisionTools Suite to be highly adaptable to the idiosyncratic drug development and selection problems I encounter in the Biopharmaceutical industry. In particular, I’ve found PrecisionTree to be a superb tool for live working sessions with Project Teams containing multi-disciplinary experts,” he explains. “PrecisionTree offers compelling graphical outputs that facilitate communication with senior decision-makers on the most complex of investment issues.”
To learn more about Richard Bayney and Project and Portfolio Value Creation, visit www.ppvc.net. or contact him directly at email@example.com