Cost of delay Equally important in weighing the safety of experimental treatments, though, is the need to consider the cost of delaying
a drug's availability. Alan Garber, director of Stanford University's Center for Health Policy and Center for Primary Care
and Outcomes Research, comments in a companion editorial in HealthAffairs [ http://content.healthaffairs.org/cgi/content/abstract/27/5/w371/] that it is necessary to balance pre- and postapproval data collection methods for a drug with the need for timely access
to a new therapy. For most treatments, FDA demands for more preapproval data will slow the approval process, Garber notes. It may be that the
cost of larger clinical trials is justified for drugs such as COX-2 inhibitors that will be broadly used to treat a chronic
condition, and that a delay in FDA approval of such a therapy may be acceptable because alternative therapies are available.
"But the consequences are much more costly if the drug in question is a unique treatment for a life-threatening condition,"
Garber says. He also points out that postapproval data collection has become much more effective due to electronic health information systems
and improvements in postapproval clinical trials. And data obtained from real-world health care settings are more informative
about how a drug will affect broad subject populations. The balance between pre- and postapproval data collection requirements
thus will shift for different drugs. Useful analysis Because no single approach for detecting drug safety problems fits all medications, the Duke analysis is valuable in demonstrating
how modeling the impact of clinical trial size can better inform study design. If a smaller database has adequate statistical
power to detect the increased risk of an adverse event, it would waste resources to demand larger and more costly development
programs. Overly broad requirements thus could reduce incentives for sponsors to develop new products and could deny important
medical options to patients. Higher regulatory hurdles would have a particularly serious effect on small companies struggling
to bring new drugs to market. But where smaller safety databases lack the power to reliably detect ADEs, the Duke analysis indicates that it may be beneficial
to require more data. Such policies could prevent approval of unsafe drugs or use of treatments in inappropriate subject populations.
While it is important to design clinical trials to assess infrequent ADEs that carry serious health consequences, in the end,
even the largest clinical trials may not detect every adverse event. Preapproval studies can seldom fully characterize the
risks associated with a drug in real-world clinical use nor evaluate long-term effects of a drug prior to market approval.
However, sponsors may obtain more information about a product's safety profile by including high-risk patients in clinical
studies as a way to increase the background risk of an event. Broader consensus on the types of statistical tests used to detect safety issues and the level of acceptable risk for different
types of diseases also could be helpful to FDA and other regulatory authorities. And more standardized reporting of safety
data would make risk information more useful to prescribers and patients. FDA recently published a guidance describing a common
format for reporting adverse events and other safety findings for investigational drugs and biologics as part of ICH efforts
to harmonize drug development and regulatory practices [available at http://www.fda.gov/cder/guidance/8413dft.pdf]. The guidance clarifies the timing for sponsors to submit this information in annual development safety update reports (DSURs).
Jill Wechsler is the Washington editor of Applied Clinical Trials, (301) 656-4634
jwechsler@advanstar.com
| 
