Using Advanced Simulation in Clinical Operations Jack Porter, President and CEO, ePharmamindshare The introduction of supercomputers and sophisticated simulation software for planning and optimizing clinical trials is rapidly becoming established in the pharmaceutical industry. These simulations are being used for country allocation, site/subject mix, protocol optimization, study troubleshooting, and site identification. A few key statistics that speak volumes about the current performance of clinical trials highlight the fact that supercomputer-assisted simulation is much needed to help accelerate the current performance of clinical trials: • 35% of all investigative sites fail to screen a single patient • 40% of all investigators complete fewer than two patients, causing these sites to become the most expensive • 80% of all patient recruitment power comes from fewer than 30% of investigator sites • 76% of all Phase III studies are more than 90 days late Identifying Current Challenges Current practice, while automated, generates a great deal of data but doesn’t integrate the information in one place, for example, one warehouse. The current process, which employs such tools as electronic data capture, clinical-trial management, and other systems, is still heavily paper based and the forecasts for such key elements as site/investigator selection and patient recruitment are not very useful. The upshot of the current approach to clinical-trial development is that site selection and investigator recruitment are done anecdotally and intuitively, by researchers for the sponsoring company simply talking to country managers and investigators with whom they have worked in the past. This approach ultimately doesn’t allow for the essential evaluation and knowledge-gathering that can contribute to improved studies in the future. Put another way, current practice relies almost entirely on intuition, and constructing clinical trials is way too complex for intuition. It attempts to apply a statistical approach to a problem that isn’t a statistical problem; it’s a multidimensional, multivariant challenge. Changing Current Processes with Simulation Tools What needs to happen is the creation of a new approach to optimizing clinical studies. That can be achieved through full simulation that begins with a comprehensive database of worldwide investigators — about 25,000 doctors — drawn from all of the top sponsoring companies. The database should be a treasure trove of information about investigator performance and such things as what equipment is available to investigators. What also needs to happen is a comparison of specific site protocols against a universe of protocols that determine which investigators performed which trials the best and what characteristics contributed to outstanding performance. This process will allow sponsors to determine what they should look for in investigators for their trials and what their ideal investigator should be. Other simulations will allow for comparing countries and specific sites within countries against protocol requirements to determine where studies should be run, as well as matching groups of people against the requirements to decide where substantial populations of potential patients reside. Although simulation is currently being applied principally in selecting sites and investigators, the capabilities of simulation are much broader and will almost certainly make themselves felt in the future. For instance, simulation will be used for allocating monitors to investigators, forecasting the quality and reliability of data generated, forecasting the regulatory process around the study, and eventually will extend to marketing, sales, and other aspects of commercialization. The advantages of simulation are numerous. Thanks to supercomputers, the ability to run literally thousands of simulations at once and to track multiple variables simultaneously permit examination of dynamics that are just impossible to see in data, such as erroneous trial data that may creep into study results. Simulation also enables advanced learning to identify best practices, thereby permitting modifications that produce even more effective simulations. Although simulation is currently being applied principally in selecting sites and investigators, the capabilities of simulation are much broader and will almost certainly make themselves felt in the future. clinical-study technologies Cegedim Dendrite, Bedminster, N.J., enables sales, marketing, clinical, and compliance solutions for the global pharmaceutical industry. For more information, visit dendrite.com. June 2007 VIEW on Clinical Services

Download PDF