新薬開発の情報システム

1.1 From Corporate Organization to Organism

There is a new competitive reality in the pharmaceutical industry that has basically emerged in two stages. The first stage was the introduction of new chemical synthesis and assay technologies that increased the screening capacity of a company by several orders of magnitude. The sheer quantities of compounds, the number of disease targets and the variety of assays that researchers must employ in the search for new drugs have created an unprecedented urgency for the development of information systems to store and manipulate data from these sources. Currently, there are many point solutions that have been developed to answer that challenge.

The source of the data flood is not restricted to drug discovery. There is also an ever-increasing sophistication in the pre-clinical and clinical evaluation process. Compared to ten years ago, there has been a tremendous increase in the capacity to profile a compound’s pharmacokinetic and toxicity profile well before it is tested in humans. Whereas in the past, one might test a compound in 3 or 4 in vitro assays before advancing to animal and human studies, one can now run the compound through the gauntlet of dozens of assays. Clinical studies have also added new dimensions of complexity with the advent of pharmacogenomic profiling to stratify test populations and interpret results.

The second stage in meeting the competitive challenge involves not just replacing technology with more sophisticated and data-intensive tools, but the creation of a knowledge environment that allows for the interoperability between all the various groups in discovery and development. This stage implies that the R&D organization itself must be transformed from a collection of independent units (like cells), to an integrated whole; a functioning organism in which all parts are in communication with each other. It will not be a simple task to carry out this transformation.

Partnerships between end-users and their IT counterparts will be essential to re-define the way the company structures its internal collaborations and build information systems that afford the greatest possible utility to carry this out. Some of the major components of a modern pharmaceutical IT infrastructure will include:

• Middleware data integration solutions to facilitate access to often heterogeneous and dispersed information
  

• Analytical applications that foster insight through cross-domain knowledge exploration and real-time collaboration
  

• Research execution systems that enable more efficient logistics support related to research materials, reagents and samples
  

• The ability to integrate pre-clinical research, clinical development and business drivers of corporate success
  

• A collaborative research system to capture project information from personnel and store and disseminate the resulting knowledge.

1.2 The Drivers of R&D Transformation

This transformation is not occurring in a vacuum ? there are several drivers that are compelling pharmaceutical companies to consider re-engineering both IT and business processes.

The pharmaceutical industry environment has changed. There has been, for some time, tremendous pressure to measure success by the number of billion dollar drugs a company has on the market. The rapid success of Celebrex, Viagra and Lipitor has made large pharmaceutical companies think differently about where they focus their efforts and how they spend their resources. Anything that can enhance innovative product development is seen as a strategic advantage. However, the industry’s current productivity level, less than one NCE per company per year for the top 50 pharmaceutical companies, falls far short of the 5-6 NCE’s per year required to meet Wall Street and self-imposed projections for growth. Of course, the innovation deficit need not be made up entirely by the introduction of new products. One may also factor in contributions from new drug delivery systems and formulations (NDDs), new combination therapies and synergies between diagnostic and therapeutic products.

Pharmacogenomics is raising the bar even higher as companies are trying to meet expectations based on older market models. If 5-6 compounds had to be marketed to sustain growth in traditional markets, how many must be discovered and marketed to sustain growth in micro-segmented markets?

The use of mass production technology in chemical synthesis and screening may have improved the situation somewhat, yet it has so far not been able to substantially increase output at the end of the pipeline. Therefore, each company must consider how it can optimize processes to get the most out of the new technologies. The analogy that has been used to describe this situation is that you are driving along on the Autobahn in a Mercedes at 100 miles per hour and somebody asks you to change all four tires. The race toward market dominance will be won by those companies that can change the tires of a moving car.

1.3 The Information Spine

Research and development organizations have been challenged to increase efficiency in order to get more NCEs on the market while holding down costs. The critical ingredients for increased efficiency include information technology tools to improve the decision-making process, allow organization-wide collaboration and an underlying information spine to enable easy access to multiple sources of data.

Today’s pharmaceutical data sources are highly heterogeneous. They have become geographically dispersed, either as part of large multi-national corporations, or through collaborative research. They are dynamic in an uncontrolled way, sometimes with entries being copied or modified several times. There is often little consistency in formats, database management systems, hardware platforms, access methods and data models. At this level of organization, it is no wonder that the industry has not been able to get the most out of the technology that produces the data. At some point, a company will have to evaluate whether it is worth generating more information if there is no infrastructure to extract knowledge and decisions from it.

Data warehousing had been considered by some as a means to deal with large amounts of geographically dispersed information. However, it may in fact not be the ultimate solution to accelerate pharmaceutical discovery. Generally, warehousing does not provide mechanisms to distinguish data based on origin or quality and data mining on these systems still requires knowledge of SQL (used by less than half of one percent of users).

The modern information spine will be constructed of middleware systems that access the different repositories of data without re-casting them in a warehouse. The middleware will apply standard formatting rules defined by the business users that essentially wrap the data and present it to a client side application. Users are presented with consistent, well-integrated data that they can explore with queries that are more natural to learn, while application programmers (as long as they code using an open application programming interface) no longer need to worry about the origin of the data. They each can function within their own expertise, without having to become database gurus. Even though the bottom layer of data consists of many databases in many formats, the information spine enables users and programmers to work as if they are dealing with a single database in a single format. It is this capability that will enhance collaboration and information awareness throughout the organization.

On the client side of the information spine, tools for visualizing and mining large volumes of information allow the analyst to interactively explore structure in the data and discover hidden or unexpected relationships. New software has recently been developed that can handle numerical, categorical, genomic and unstructured text. Rather than replacing SAS or other statistical and mining tools, visualization is used to enhance these capabilities by adding in the power of visual discernment, through the recognition of outliers, patterns and movement. This is the capability that will enhance innovation and discovery.

As the technologies of combinatorial synthesis and high throughput screening advance, an increasing burden is placed on the research infrastructure to support what has essentially become a manufacturing effort. Research execution systems, a new category of tools for discovery, facilitate the process of acquiring reagents, requesting screens and tracking the movement of sample. This is a capability that can ultimately improve the productivity of an R&D operation.

Much of the potential benefit of technological advances is predicated on being able to bring to bear all of the information at a company’s disposal to understand or predict the behavior of a drug against a specific disease, to plan experiments and to evaluate strategy in the context of new results. This knowledge-led R&D process requires that scientists (and business users) formulate questions freely using information from all parts of the company to gain insight into their project. By freeing the users and IS/IT staff from the constraints of existing legacy and domain specific systems, they can begin to think more naturally about information from all sources and use it in creative and innovative ways.

1.4 Changing the Wheels of a Moving Car

Implementation of this new system is a lot different than simply installing a suite of programs. It will involve extensive interviews among the users and management to define current business processes and determine how to design a system to optimize those processes. Multidisciplinary design teams will be required in order to gain from their expertise and their buy-in. Moving from legacy systems to systems based on full collaboration and sharing of knowledge is a major undertaking and one that will require a culture shift in most organizations.

Not everyone can afford to dedicate such effort and resources to changing their informatics infrastructure, particularly smaller companies. A new business model has emerged in Internet space that might be able to address that problem: the Application Service Provider (ASP). ASPs provide remote sourcing and management of technically complex applications using the Internet telecommunications infrastructure. The application of the ASP model to R&D efforts in the life sciences finally enables small and midsize biotechnology and pharmaceutical companies to have the same access to state-of-the-art information technology (IT) as their large competitors. In effect, the ASP model will drive IT innovation into the most innovative segments of the life sciences industry.

1.5 Democratizing Business Strategy

In addition to the science and technology, a company must integrate its understanding of the market, including its competitors, political forces, customer trends and the patent landscape. According to Dr. Joseph Villafranca, Vice President of Biopharmaceuticals at Bristol Myers Squibb, the company is seeking competitive advantage by modifying their research and development organization to do just that.

Bristol Myers Squibb had just undergone a major organizational change in the summer of 1999, organizing by governance rather than by individual projects. In doing so, they are now able to integrate project activities by considering all aspects: science, technology, manufacturing, healthcare systems, customers, competition and products. According to Dr. Villafranca, “It is quite interesting now to go to a meeting where there is representation from the marketing group, representation from clinical, representation from drug discovery, as you are talking about an early discovery program”. He noted that the integration from discovery to marketing provided many unexpected benefits and insights. The intelligence contributed by the business side “came as a surprise to a lot of the discovery scientists, who feel that they are the intellectual engine of the company.”

Knowledge management systems facilitate this type of cooperation by essentially capturing what may be described as meta-data ? the information held within an individual or group. These include lessons learned, such as the optimal way to conduct an experiment, or organize a team or information about competitors and the market. It not only includes recorded knowledge, but also facilitates the flow of knowledge throughout the organization. This may be achieved by maintaining a database of in-house experts that can be searched by topic. Once experts in a particular area are identified, an individual can contact them to obtain vital information first hand.

The improvement of discovery, development and licensing performance can be achieved by systematic assessment of research, technology, patents, companies and changes in those parameters over time. Because the majority of this competitive intelligence information is available only in free text form, which has traditionally has been difficult to explore for associations and trends, leading companies will be able to better position their research and products by employing informatics systems based on text-based search and visualization. In the new R&D paradigm, competitive intelligence and business strategy are the responsibility of everyone from the bench scientist to chief executive officer. Information will be key democratizing business strategy.

1.6 Conclusion

Research informatics has emerged as an essential component of competitive advantage in the pharmaceutical industry. It has been well established that inefficiencies in R&D have largely arisen because of the inability to freely access and share information and the inadequacy of conventional tools to explore relationships between data. As with any business tool, the success of research informatics to eliminate these inefficiencies lies not only in the power of technology, but also in the ability of the individual and the organization to use it as it was intended. Therefore, it will be essential that re-design of information technology progress step-by-step with business re-organization. In today’s data- and knowledge-intensive pharmaceutical industry, they are one and the same.

1.1 From Corporate Organization to Organism
1.2 The Drivers of R&D Transformation
1.3 The Information Spine
1.4 Changing the Wheels of a Moving Car
1.5 Democratizing Business Strategy

2. The Information Spine for P