Why Smart Statistics Can Save Pharma

January 15, 2015
Phil Birch, DPhil

Applied Clinical Trials

Fundamental weaknesses of modern clinical development can be resolved through recent statistical advances.


Inefficient use of statistics is a detriment to the drug development process. The majority of dose finding trials do not test enough doses (often just two or three) to accurately determine dose-response. Many of the models used to estimate the optimal dose for Phase III are prone to uncertainty. Over- and under-powered trials waste resources and patients’ time.

Those fundamental weaknesses of modern clinical development-which in many cases undermine good candidate drugs and thus good companies-can be resolved through recent statistical advances.

MCP-Mod, for example, addresses weaknesses in design and analysis of dose finding studies, a common issue that can lead to selection of a suboptimal target dose for the pivotal study. The methodology increases the range of doses studied and considers several dose-response models before selecting the best one. This approach leads to a more accurate estimation of the dose to move to Phase III.

Regulators and industry have voiced support for MCP-Mod’s implementation. The EMA, in its first qualification opinion ever issued on a statistical method, recently qualified MCP-Mod as a more robust approach to dose finding. Shortly before the EMA issued its opinion, an industry consortium formed in order to produce a standardized and validated toolset for executing MCP-Mod in trials. The ADDPLAN DF Consortium, founded by Novartis, Janssen, Eli Lilly, and ICON plc’s Aptiv Solutions, has grown to also include Pfizer and Roche.

Recognizing the flexibility of another smart statistical approach, adaptive design, the Consortium has expanded its focus beyond refining a standard operating platform for MCP-Mod to incorporating an adaptive functionality into MCP-Mod. The broad aim of the collaboration between these pharmaceutical companies is to use sophisticated adaptive designs, such as those for dose finding, to improve clinical decision-making at critical points in the development process, whether in exploratory, late, or post-market phases of development.

Having matured from an operational standpoint and in terms of explicit regulatory support, methods such as MCP-Mod and sophisticated adaptive designs are no longer optional to implement. These methods, as well as other smart approaches to trial design and execution, are boardroom strategies that increase the bottom-line productivity and valuation of portfolios.

At the portfolio level, these strategies exert de-risking effects that prevent waste and provide the best opportunity for approval and commercialization. Portfolio-level benefits can only occur with early decision-maker buy-in, driven by a clear process for change, internal regulatory support, and technology to maintain trial validity and integrity-which is not a standard operational endeavor. CROs, through strategic alliance partnerships, can provide insight, expertise, and technology to help medical and clinical teams operationalize these innovative strategies to enhance pipeline decision-making.

A number of top 10 pharmaceutical companies have placed improved statistics at the center of go-to-market strategies. Midsize and small companies are following suit at an accelerating pace. Sponsors that couple smarter statistics with early health economic modeling and patient-centric or risk-based monitoring have the best chance of separating their company from the pack against a backdrop of increasing development costs and pressure to demonstrate value.


Author Information

Phil Birch, DPhil (Phil.Birch@iconplc.com) is VP Innovation Strategy, Alliance Partnerships at ICON plc.

Related Content:

News | Blogs