Predicting the Probability of Success in Clinical Drug Development

Although we have made several strides in clinical drug development, there are still several factors that impact our

Uma Arumugam

drug development endeavours as posited in the past. However, the encouraging news is that the probability of success (POS) in clinical drug development is higher than previously estimated. This is welcome news for investors and drug developers reaffirming their confidence in the return on their investment, specifically for ventures that bootstrap in order to increase the POS, given that they have limited human and capital resources.

In particular, we have had tremendous innovation in clinical trial design methodologies in the last decade, especially with the introduction of adaptive trial design and master protocols (umbrella trials, basket trials, and platform trials). These, along with advancements made in the execution of these clinical trials with the use of technology (electronic data capture, electronic health records, and digitization), have paved the way for more efficient processes to be implemented and help us realize the benefits much more rapidly. However, with clinical trials becoming more complex, increasingly resource and capital intense, and time-consuming, there still lingers the question on the outcome of these trials in terms of success, despite our best efforts. For example, in oncology, it is well recognized that more than 95% of the drugs/compounds that have a demonstrable effect on animals, fail in Phase I clinical trials in humans, indicating that most preclinical models of cancer are inadequate. In addition, most of the anticancer drugs either have no effect on the overall survival of the cancer patient or may provide an increase in few months in overall survival. This dismal rate of success leads to the lingering question of the POS that every investor and drug developer would like to know prior to developing a portfolio of assets or a particular compound for a specific therapeutic indication.

Estimates of POS in clinical trials have been published in the past with some variances in their approach and results (see here and here). Last April, Chi Heem Wong and others at MIT published a revised estimation by reconstructing drug development pathways, which is believed to be a more accurate estimate of the POS in drug development. One of the interesting findings is that they obtained higher POS estimates for all phases relative to the estimates published before. Additionally, it has been shown to result in a lower estimated drug development cost, especially in Phase III, where the cost of conducting a trial dominates those of other phases.

A closer inspection of this data reveals an increasing trend post-2013, which was preceded by a decade of decrease in successful outcomes. This has mainly been attributed to careful licensing of compounds and better identifying potential failures, thus leading to higher productivity. They also propose two other possible reasons for the trend: one, the increased use of biomarkers to target drugs at patients who are more likely to produce a positive response; the other, the new wave of medical breakthroughs. They also found that clinical trials using biomarkers for patient stratification have higher success rates, especially in the area of oncology.

Analyzing and interrogating both the scientific and operational data in order to assess the POS from all the accumulating clinical trials in an ongoing fashion will likely yield significant insights. Undertaking an assessment of the POS of an asset can be coupled with the product profile and safety assessments of assets/compounds prior to launching a clinical trial. As we make progress and gain additional actionable intelligence, we can perhaps develop a framework that can eventually be used as a guidance tool for designing clinical trials. Obviously, this will require resources to be allocated, which in the long term will have a profound impact on streamlining the process and, most importantly, reducing the cost of drug development.

Uma Arumugam, MD, is Director, Clinical R&D, Early Phase Services, ICON plc