Getting patients and doctors into clinical trials has become the most delay ridden aspect of the drug discovery and development
process over the past 10 years. Patient and investigator recruitment and retention are often cited as the most costly and
time consuming aspects of trial operations.1 However, there is good news for managers seeking new ways to facilitate the overall clinical trial process, especially enrollment
of subjects. Simulation exercises, made possible today by customizable decision support technologies and well-defined predictive
analysis strategies, can help clinical trial managers cut costs, reduce risk, forecast outcomes, and increase efficiency.
Simulation solutions
Photography: Getty Images Illustration: Paul A. Belci
|
Increasingly, pharma companies are turning to simulation—a merger of strategy, analysis, and technology that enables them
to run through virtual scenarios of their major initiatives. A new application of this technology is in the critical area
of clinical trials. The end tools allow the user to forecast the resources needed, predict and adjust the anticipated costs,
establish realistic goals, measure and mitigate their risks, and in the end make better business decisions. For example, the
solution can be used in conjunction with existing or expected recruitment data to help predict how patient recruitment sites
will perform based upon a predetermined protocol using a set of input templates.
In this context, the protocol is a study plan on which all clinical trials are based. The plan is carefully designed to safeguard
the health of the participants as well as answer specific research questions. By definition of the U.S. National Institutes
of Health, a protocol describes the following:
- What types of people may participate in the trial
- The schedule of tests, procedures, medications, and dosages
- The length of the study.
Decision support technology can run multiple replications of many different scenarios that take into account the clinical
trial process's variability and complex resource interdependence. These scenarios generate realistic data on how a company's
subject recruitment process will perform. This includes enrolling recruiting sites, setting up training sites, getting subjects
recruited, dosing them, seeing them through the entire cycle, and then closing out the clinical trial.
Implementing decision support
Companies that want to implement a simulation strategy must first determine if they have a need. To gain the greatest benefit,
a company engaging in simulation should have more than one clinical trial per year. It should also be a relatively sizable
mid cap or a larger firm that runs clinical trials frequently to justify the time and effort required for implementation.
Second, trial managers should communicate with other key participants and decision makers early in the study to determine
if there is a specific need to complete a trial on time and within a certain budget—and what those parameters are. It is critical
that trial managers talk not just to the doctors and scientists conducting the study but also to the senior management in
the company and other trial managers.
Performance Measures Plot
|
The next step, data evaluation, has a significant impact on how simulation is carried out in the clinical trial. Companies
and their consultants or technology providers must review the available portfolio data and determine if it is sufficient to
support full predictive analysis. It's important to know how long a particular site takes to get up to speed, when it's fully
trained, when it starts enrolling subjects, and what possible delays could arise.
Many times individual sites have their own data, or the pharmaceutical company has collected data from past experiences. Relevant
information includes site type, delays to recruitment of the site, delays to training of the site, different cycle times,
and more. The accuracy of the data regarding delays to site qualification and rate of subject enrollment is critical to the
success of the trial. The risks of using data that is based on estimates rather than actual results can be mitigated by expanding
the bounds of the distributions that are used for these entries in the simulation.
