Jill Wechsler

FDA believes that new analytical and diagnostic technologies open the door to developing more accurate and predictive measures of how a test product may affect physiological or pharmacological responses. In vitro diagnostic tests could identify patients at risk for adverse events as well as those most likely to benefit from therapy. New safety biomarkers may provide early indications of a compound's toxicity and avoid wasted research efforts. And markers of drug metabolism may accelerate individualized drug dosage that can yield safer and more effective treatments.

As the most expensive aspect of drug development, a main Critical Path objective is to encourage more productive research practices. This involves moving away from long and costly empirical testing and adopting trial designs based on a better understanding of underlying disease states and drug mechanisms of action. FDA wants sponsors to conduct more "learning trials" to explore dose–response relationships and yield more information about product performance. Instead of just gathering data on how a population responds to a test product, the aim is to gain information on how variations in disease states, organ systems, and genomes can make individuals respond differently to treatment.

The list of proposals for creating more efficient clinical trials ranges from fairly practical approaches, such as establishing common standards for case report forms and for collecting clinical trial data, to devising more sophisticated research models. FDA would like to see consortiums formed to clarify statistical methods and propose standards for designing non-inferiority trials and studies with enriched patient populations. Agreement is sought on rules for revising trials based on interim results and for using historical data and animal studies to better predict human response. Groups also could address how to handle studies compromised by subject attrition and missing data.

Tailored study protocols

Another important objective is to develop study protocols tailored to specific diseases or conditions. New trial designs could improve understanding of how treatment response relates to outcomes for cancer patients. More analysis on whether pathogen levels in the blood accurately correlate to infection may improve development of vaccines and antibiotics. Standardized measures of patient symptoms and pain scores would help researchers measure a variety of indicators.

New approaches to data pooling and statistical modeling also hold promise for improved decision-making in the clinical research arena. One project in the works, for example, aims to establish a large database on cardiac arrhythmic risk to address the significance of QT interval prolongation. Pooled data on historical controls could reduce the size of control groups in clinical trials, while combined adverse event data on products or diseases could avoid safety problems in research.

Computer models of human physiology also may yield information about likely long-term performance of implanted medical devices, and clinical trial simulation could improve decisions on drug dosing. FDA and sponsors have discussed establishing a public database on unsuccessful clinical trials as a way to identify patterns associated with failure and help sponsors avoid repeating past mistakes. Databases on rare diseases, including patient histories and biomarkers, could provide virtual historical control groups, identify study subjects, and improve the design of clinical programs.

Encouraging collaboration

The Opportunities List aims to spur companies and research organizations to jointly tackle these challenges. FDA recognizes that government agencies and individual sponsors cannot address this broad range of activity, and that the agency's role is to help define how partnerships may work and to evaluate results and develop policy guidances on new approaches that researchers find to be effective.