Reconsidering the Way We Perform Clinical Trials
The application of technologies to human disease has enabled an hitherto unprecedented insights into cancer biology and genetics. At least as importantly, the parallel development of targeted therapies has created opportunities for treatments. The combination of increased new genetic knowledge about cancer, an expanding range of therapies that appear to work best when applied to distinct genetic subsets of cancer, and increased clinical trials would appear to present remarkable opportunities to translate knowledge into improved health outcomes.
Unfortunately, the quasi-exponential evolution of knowledge has been offset by impediments on clinical research. In a series of devastating articles, Steensma and Kantarjian have recently outlined the counter-evolution in bureaucratic, regulatory and fiscal constraints on clinical cancer research.1-3 Consider the following evidence.
Steensma and Kantarjian contrast the conduct of a study of the drug azacytidine in 1983, in which the trial protocol involved less than 10 pages, was approved by the institutional review board within one week of filing, and commenced enrollment the following week; a 2012 study of the same drug required more than two years of development and "hundreds of steps involving the US and Canadian national cancer institutes, various co-operative groups, and National Cancer Institute’s Cancer Therapy Evaluation Program (CTEP), and duplicative actions by burgeoning ranks of local potentates at each of the 244 study locations (more study locations than the planned accrual of 240 patients).3" Kantarijan and colleagues cite four sources estimating that the time to initiation of a CTEP or co-operative group clinical trial currently at between 400-700 days.1 For many patients, this is time they do not have.
Steensma and Kantarjian also estimate that increasing regulatory requirements and the commercialization of conduct of clinical trials has led to an increase in costs per patient on a phase IIIA oncology trial from an average of US$5000 in the early 1990s to up to $125,000 in 2013.3 They estimate that per patient costs increased by 70% over the three years from 2008-11 alone.3 These costs contribute to the estimated $1 billion dollars needed to develop cancer drugs—and which must ultimately be recovered by the pharmaceutical industry in any viable business model. Not only does this limit the rate at which opportunities for improved outcomes can be demonstrated, it may make therapies less affordable to the patient and the community by inflating the costs of drug development.
Unarguably, protections are necessary. As one example, the fabrication of results in a study of high-dose chemotherapy in breast cancer, underlines this point.4-5 The devil is in the detail. It has been suggested that increased bureaucracy has not been associated with greater trial quality or safety. Kantarjian et al1 estimate that half of the 300-600 steps involved in activation of clinical trials are unnecessary and may actually be preventing access to clinical trials, thereby harming the very people they were originally designed to protect. Our health system needs to evolve to keep pace with science.
What changes might accelerate progress in translating knowledge into clinical practice?
• Distinguishing trials involving populations with good therapeutic options (for example, adjuvant studies), from those involving populations with few or no available therapeutic options. It seems desirable to have different standards for protections for studies involving patients dying from cancer, for whom experimental therapies may represent their best hope. Such considerations may also apply to rare or understudied diseases (like sarcomas), where a proactive approach may be required to reduce the barriers for clinical trials.
• Reductions in needless bureaucracy, specifically reducing the numbers of the hundreds of steps identified in recent reviews as unconnected with safety or quality. This is not a process which needs to be begun anew: Kantarjian and colleagues note that there have been reviews of clinical trials efficiencies over the past decade that could represent points of departure.
• Benchmarks for efficiency (time to trial initiation, protocol length, length of patient information and consent forms, &c). Such benchmarks ought to include measures of access to clinical trials. There is good evidence that participation in clinical trials is associated with better clinical outcomes, both for the individual, and for improvements for the beneficiary populations.
• Consideration of alternate mechanisms to ensure safety and quality outcomes. For example, is a single standard of protection necessary for studies performed at high volume centres, compared to centres inexperienced in clinical research? Could some of the regulatory burden for trials be transferred from individual protocols to the ethics and review boards at major trials centres? This may include increased institutional oversight and accountability mechanisms, to monitor quality outcomes.
• Novel trial designs that enable more rapid evaluation of novel therapeutic strategies. Such trial designs may include more efficient methods of testing for signals of activity for existing drugs (so called repurposing), basket designs that pool multiple drugs and indications in a single protocol, and Bayesian statistical designs.
• Global strategies (and funding mechanisms) to enable more rapid testing of earlier phase trials, particularly in rare cancers.