As Published In February 2007

LIFESCIENCES

Found in Translation
By Jeff Erickson

Windber research pioneers the next generation in translational medicine.

When patients diagnosed with cancer or heart disease walk into the Windber, Pennsylvania-based Windber Medical Center, they ask questions about choice of therapies and associated risks. Now, armed with a new take on translational medicine and powerful software built on Oracle Database 10g, Windber clinicians can provide better answers—and more-effective care.

Translational medicine has been defined as the process of taking results from a laboratory experiment and converting them into information that can be used for patient care. Windber Research Institute (WRI), a partner of the Windber Medical Center and a pioneer in data-intensive biomedical research, is now turning this system on its head.

"We focus on working with clinicians to identify the problems they see on a day-to-day basis, when the patient is sitting in the room with them, and then move the problem back to the research bench," says Michael Liebman, executive director of WRI. "Now, when a researcher brings all his or her expertise and technology to bear on a research problem, the solution has an immediate use in the clinic." The benefits of WRI's approach are enormous: patients benefit faster from cutting-edge research, and researchers get invaluable clinical data to support, clarify, and refine further research.

A Range of Datasources

The quality and quantity of WRI's patient data sets it apart from most other research institutes. WRI benefits from close ties with Walter Reed Army Medical Center, which collects a high volume of information about patients such as details about a patient's exposure to chemicals and whether the patient smokes or drinks alcohol. These researchers also have access to a huge set of clinical data, including tissue samples, genomics, proteomics, and imaging data from mammographies and ultrasound, PET, CT, and MRI scans, as well as pathology images. "We've built a patient-centric data model that integrates all of that data and makes it available to our researchers," says Liebman.

Snapshot Windber Research Institute www.wriwindber.org Annual revenue: US$1.6 billion Employees: 55 Year started: 2001 Oracle products: Oracle Database 10g Other products: InforSense informatics technologies, Concentia Image repository InforSense www.inforsense.com Year started: 1999

"We have a lot of different disciplines looking at any given problem," says Liebman. "At the same time, we're looking at the clinical data. We're looking at the socioeconomic data. We're looking at the psychological information. We're looking at how patients respond to a range of treatments. We're even looking at things like how patients respond to being told that they're high-risk or low-risk," he adds.

WRI then takes this data set and adds one more dimension. "We treat patients as a collection of longitudinal information," say Liebman. "A disease is a process that takes place over time. Our patients are changing while they have the disease. It's an important aspect to look at, but very few places are doing it. We need to understand how patients' aging processes will impact their responses to a therapy or their suscepti-bility to other diseases."

A Single Integrated View

The challenge has been to build an IT support infrastructure that provides a "plug-and-play" environment for researchers, with workflow that can continually change and evolve as researchers work with a mix of clinical science, molecular science, and clinical practice.

"How do you combine data with high-throughput experimental techniques like genomics, genetics, and proteomics?" asks Jonathan Sheldon, chief scientific officer at InforSense, a pioneer in integrative analytics technology and WRI's key partner in its translational medicine initiative. "How do you bring those technologies together with the clinical data and make the results mean something to researchers?"

Working with InforSense, WRI has adopted an approach that provides a single integrated view of its various data resources, combined with analytical tools, so that the clinician and the molecular scientist or the epidemiologist can collaborate within a single informatics environment. "With InforSense on Oracle, the researcher can integrate these different types of data and do that in a very ad hoc, flexible way," says Sheldon. "He or she can bring in the clinical data, merge that with the genome-expression data, and carry out various analytics or modeling to look for combinations that have a clinical significance and utility," he says.

Once a researcher has identified a gene that, for example, suggests whether a patient will respond to a particular treatment, the system delivers that information back to the clinician at the center in the form of risk analysis algorithms.

"To make the information usable to the clinician, the InforSense solution for WRI takes complex analytical workflows, wraps them up, and deploys them as a Web portal that hides all the complexity," says Sheldon. In the Web portal, researchers and clinicians can browse and drill down into data, choosing from hundreds of dimensions, to identify patient populations for further exploration. The populations can then be used as a starting point for Web-deployed workflows that enable complex analysis of experimental data. As data and analysis techniques change, new workflows can easily be published into the portal, enabling informatics staff to support a broad, disparate, and evolving user community within a single technology framework.

"Even though the researcher may have built the most-complex workflows that use many different kinds of data sets, the clinician can hit a few buttons and the system runs the analytics in the background," says Sheldon. "The results are returned back into a portal that has a very intuitive interface for the clinician."

Choosing Oracle

WRI and InforSense selected Oracle Database 10g as the engine that is powering their next-generation solution for several reasons. First, WRI has to seamlessly integrate data from diverse sources, including research institutes, hospitals, government agencies, pharmaceutical companies, and healthcare providers—all with disparate IT systems.

Just as important, WRI and InforSense needed a solution that helped ensure the security of medical data. By enabling in-database analytics, Oracle Database 10g minimizes movement of data in and out of the database, significantly reducing security risks. "For security and compliance reasons, that was a key issue for WRI," says Sheldon.

ROI for the Research Center

As biomedical information grows exponentially more complex, translational research is testing the limits of IT solutions. Sheldon believes enterprise software that incorporates IT best practices with research institute best practices will become a must. "Instead of writing custom Perl scripts to get the analysis you want, which can take hundreds of days of coding for a single study, you can install the latest software and get the same capability in as little as a couple of weeks," says Sheldon. And unlike custom scripts, enterprise software is reusable. "Use the software to analyze a study on, say, lymphoma; then simply drop in a different disease area and the workflow would run on that data and return relevant results," he adds.

Moving Forward

For More Information Oracle for Life Sciences Oracle Database 10g

With the help of InforSense informatics technologies and Oracle Database, WRI has built an IT backbone for a new kind of physician/patient decision support system. Now WRI, a nonprofit, is starting a for-profit venture to assist other medical centers in doing the same. It has already begun commercial ventures in the U.S. and the Netherlands.

"Translational research is a very hot area," says Sheldon. "Everybody's talking about it, but there is very little out there that really does accomplish it. WRI and InforSense have designed the system with clinicians' needs in mind," he emphasizes. "That's really why the collaboration between InforSense and WRI has been such a success, and why other medical centers are taking notice."

"InforSense workflow technology using Oracle Database enables us to build solutions that are flexible enough to support the dynamic and iterative thought processes of our scientists and our clinicians," says Liebman. "Combined with the ability to then deploy these findings throughout our research teams, this technology enables our scientists to translate their research into real decisions that can impact patient care. We see this solution as the 'command-and-control center' for our clinicians to give patients the best possible answers and the best possible care."

Trendwatch Q&A: Understanding Patient-Driven Translational Research PROFIT spoke to Patrick F. Terry, an entrepreneur and pioneer in patient-driven translational research. In 2000, Terry and a group of leaders in the fields of genomics, bioinformatics, and biotechnology founded a pioneering clinical genomics company. Terry has also founded a series of international nonprofit life science research organizations. His perspective is forged from personal and family experiences with genetic disease and cancer and extensive work in the patient advocacy community. PROFIT: What is patient-driven translational research? TERRY: Translational research is fundamentally about integration of a results-focused delivery system into a single bidirectional pathway—a meaningful and clinically significant results-driven enterprise that impacts real-world patient care. It can be oriented quite easily as an information system, because ultimately our genetic code is an elegant informational storage, retrieval, and adaptive export system. All of the analogies and fundamental laws of information systems can apply to this patient-centric biomedical concept. Quantification of biological measurements across the dynamic range of biology with an entirely new level of precision equates to a tremendously powerful data set and potential insights into disease pathology, mechanisms of action, rational interventions, and clear capture of cause-and-effect influences like never before in our history. As a result, the opportunities for improving the effectiveness of interventions—diagnostics, drugs, therapeutic regimes, and clinical trials—that will deliver actionable information to patients at a faster pace are enormous. PROFIT: What's the importance of getting patient information to researchers, rather than just having researchers think up an experiment, do it, and then get the results back into the clinical setting? TERRY: There's the power of having longitudinal information and data capture on patients and disease—patients that are traversing the treatment paradigm. Capturing that data over time is extremely important because there is valuable information embedded in both space and time. What often happens, especially in the rare-disease community where a lot of my work happens, is that disease gets mischaracterized. A disease is usually characterized in the literature from a few case studies, and they're usually extreme case studies. So the valuable components of longitudinal and rigorously collected natural history across diverse patient populations is the data that the entire research community and the policy community will be struggling with over the next decade, because we've built systems to disconnect patients from their data or patients from their sample. And the challenges will be that you can do these real-time translations and you can deliver this valuable information to research participants, but you have to counterbalance that against some serious questions—is the result reproducible, is the finding real, is it generalizable, how do you test this hypothesis? There are new systems for dealing with the data generation that can happen in real time and impact the patient sitting right in front of the physician or researcher. PROFIT: What are some of the processes that can overcome those challenges? TERRY: There's a federal law called HIPAA, which is a set of rules that protects personalized, identifiable health information. Covered entities that are either healthcare deliverers or receive federal dollars must comply with HIPAA. But as a patient group and a patient research foundation, we're a noncovered entity, so we don't have to play by the HIPAA rules, and we've built a trust model with the patient community wherein the patients consent to a research protocol and also consent to be recontacted. So if a third-party researcher discovers something or wants new information or has a new hypothesis, we operate as kind of a contract research organization. We can recontact the patient and manage all of the consenting and collection of information or biological data and deliver that to the research community. This allows for the dynamic linkage of longitudinal, environmental, molecular, and personal medical information to be captured and monitored over long periods of time. PROFIT: What are the key benefits to patients and researchers? TERRY: Collaboration. The power of the collective pulling in unison could be the most powerful benefit. For Big Science to deliver big results, we all need to work together to make translational research medicine a reality in our lifetime.