YouTube Facebook LinkedIn Google+ Twitter Xingrss  




GUEST COMMENTARY

By Judith S. Hurwitz

August 13, 2002
| A quarter of a century ago, the companies that represented the backbone of commerce in the United States were just beginning to automate their operations. These were companies focused on financial services, such as Citibank, John Hancock, and American Express, as well as manufacturing companies, such as General Electric and General Motors.

Early on, they employed armies of professionals to handcraft systems that would become the differentiator against competitors; the idea of buying packaged software was inconceivable. How could a third party without in-depth knowledge of a company's unique value proposition provide software that would fit its needs?

Today, the life science industry is beginning to flex its economic muscle in much the same way as the insurance and manufacturing industries did in past decades. These companies are beginning to position themselves as a key economic engine of the next decade. At the core of the life science revolution is IT. There is a discrepancy, however, between the emerging technology requirements and the preparedness of the technology industry to meet the needs of this market.

Hundreds of firms offer products to life science companies, including packaged software, development tools, portals, and gene chips. Though there are some very sophisticated offerings within this group, such as those from ID Business Solutions Ltd. and Spotfire Inc., many of the products on the market have significant problems. Three in particular stand out:

  • Many products are actually derivative research projects that were not designed as commercial software products. Often, the underlying code is unstructured and hence difficult to modify and scale. There is a huge difference between the way software is constructed when it is designed as a software product and when it is a consulting engagement. When writing software for a single engagement, less time is spent structuring a product to withstand the demands of a large number of customers over time. A successful application for one customer cannot necessarily be generalized.

  • Many products are narrowly designed to appeal to a single requirement, such as functional genomic modeling tools for target identification or compound registry-tracking tools. Though specialization is needed to accomplish a particular goal, customers typically need to extend this capability to adjacent areas. Most narrowly defined tools do not have published interfaces or APIs (application programming interfaces), and thus tend to be islands of functionality — disconnected from other stages of the discovery or development process.

  • Few products are designed in a modular manner so they can interconnect with other related products. Each of these products is therefore an island of automation.

For the life science market to be able to exploit technology and transform the efficiency and effectiveness of bringing promising new drugs to market, massive changes will be needed. Software for life sciences must be built on a modular framework that looks at the life cycle, from the discovery of promising drug targets through the development and delivery of those drugs.

What will the future of commercial software for life sciences look like? The following six criteria should guide technology providers:

  • Modular, Web-services orientation: A component-based architecture should underlie any technology in life sciences. This approach will enable companies to add pieces as needed and integrate across the development life cycle. Web services offer an approach to applications based on the

At the core of the life science revolution is IT. There is a discrepancy, however, between the emerging technology requirements and the preparedness of the technology industry to meet the needs of this market.
concept that applications should be created by linking a series of self-contained business services together. Some lower-level services are designed to be horizontal services used by most applications. These would include middleware, such as security services, communications protocols, rules engine, and various data management services. Other services are more specialized. For example, there could be a high-throughput screen service or a simulation for testing toxicity of a compound. Each of these business services or components is created independently of any specific application. Therefore, there are no dependencies between components. Each of these components includes a well-defined interface so that it can more readily be connected to another component, or even to an entire application. These services can be linked together and linked with existing applications to form virtual applications.

  • Interoperability: Beyond XML (extensible markup language) and SOAP (simple object access protocol), applications will have to be able to address a peer-to-peer design to allow instantaneous correlations and communications.

  • Usability: The ability to visualize information within the appropriate context is essential. The more visual the presentation of information, the more likely it can be used effectively for decision-making.

  • Scalability: Many products currently available are not designed with scalability in mind and have poor visualization in their user interfaces. As visualization plays a bigger role, sophisticated data storage and system management functions will have to be developed in conjunction with a highly visual/graphical/analytical user interface. With the trend toward personalized medicine, the need to scale in terms of application sophistication and data management will continue to grow.

  • Meta-data core: At the core of most products must be the notion of a meta-data environment. More than any other discipline, life sciences is predicated on understanding the transition of data throughout the different stages of drug discovery and development. Data must be managed universally across applications and platforms.

  • Discovery orientation: In traditional industries, such as financial services and manufacturing, transaction processing was the core engine for management. In contrast, the life sciences are about discovery, where the customer rarely knows the outcome in advance. A new way of looking at the interaction of the scientific and medical process in conjunction with data is required to truly capture the power of technology.

People are questioning whether a market for software for life sciences exists. It is a fair question, given the state of many products available. But once the ideas outlined above are melded into the products and services offered by companies, a viable and profitable industry is destined to emerge. *



Judith S. Hurwitz is a principal at strategy consulting and research firm CycleBridge Technologies in Newtonville, Mass. She can be reached at jhurwitz@cyclebridge.com.






For reprints and/or copyright permission, please contact  Jay Mulhern, (781) 972-1359, jmulhern@healthtech.com.