First, trials determining the risk-benefit balance of therapies must be larger. Improvements in existing standards of care are likely to be incremental (10%-25% relative reduction in events), and large studies with numerous events are needed to identify moderate treatment effects. It is important that these trials use clinical, not surrogate, end points.

Second, trials must be simpler. The cost of large, complex trials—often hundreds of millions of dollars—is a significant hurdle, in both initially bringing a treatment to market and subsequently conducting comparative studies. Moreover, restrictive inclusion criteria and excessive exclusion criteria often limit clinicians’ abilities to extrapolate findings to a broader, heterogeneous population. Data collection during the trial may also be excessive, and this too adds to the cost. According to the Tufts Center for the Study of Drug Development, the typical clinical trial in 2012 involved 13 end points, 169 case report form pages, and 175 days of on-site monitoring.¹

Third, for most therapies, studies must be randomized. Estimates of effect size in large observational studies may be precise but remain fundamentally hampered by bias and confounding that can be controlled only through random allocation.

Simplifying approval processes, reducing the burden of data collection, using existing registries for data collection, and at the same time maintaining protection of subjects are top priorities for the future.