Another Crack Appears in the Crumbling Clinical Trials Consensus

Right at the start of the year, the calls for radical change in the way that drug development is organized are emerging thick and fast in Europe. The first week back at work saw a conference in the European Parliament where speaker after speaker denounced the failings of the current model and the rigidity of the authorization and delivery systems in the face of new science and new technology.

The speakers ranged from the health minister of Belgium to the head of the European Medicines Agency, and from key figures in the European medicines industry to leading academics and experts on health technology assessment. But the central message was the same: opportunities for better care are being missed, and shouldn't be. 

The flavour of the discussions can be accurately inferred from the contributions of Pier Franco Conte, an oncology professor at the university of Padua, well-known in European academic and medical circles. Moderating one of the debates, he pointed to the challenges facing clinicians and researchers: a proliferation of increasingly sophisticated data and tools, but a persistent incapacity to interpret and use them.

For Conte, who is at the forefront of the growing understanding of cancer and the potential of targeted therapies,  there is a pressing need to review the way drugs are developed and delivered to patients, because despite all the assets theoretically available, there is a continuing and increased incidence and mortality from the disease.

He set out a list of changes he believes could improve the situation-and they all tend towards a disestablishment of what has, for decades, been received thinking. He recommends a switch by regulators away from obsessive attachment to statistical significance in clinical trials, and towards the concept of clinical value. It no longer makes sense, he said, to consider positively a randomized trial large enough to demonstrate a p-value difference, while rejecting a single-arm trial showing a clinically worthwhile endpoint in a population selected for prognostic and predictive characteristics. 

Classic endpoints should be adapted and valued according to the specific setting of the disease and drug under study, he maintains. He offered the example of trials with immune checkpoint inhibitors, where neither response rates nor progression free survival (and not even median overall survival) are good predictors of patient outcomes. In his view, landmark analyses such as the percentage of patients alive at a fixed time point better describe the efficacy of the treatments.

The risk-benefit ratio also requires some serious nuance as a requirement for approval. A positive ratio based on the primary trial endpoint and the toxicities or quality of life of the patient can be misleading, because it represents the average ratio for the average patient. This can be completely useless in large trials with a sufficient statistical power to demonstrate small differences. In these cases, the number needed to treat and the number needed to harm should be taken into account by regulators.

Conte is equally harsh on persistent divergences of product evaluation. A significant number of drugs approved by FDA and EMA score very poorly on scales elaborated by scientific societies such as ASCO and ESMO, which take account of parameters related to the setting. "There is no relation between drug prices and these values", he pointed out. "This misalignment between regulators and scientific societies is not justifiable or sustainable, and shared values are now eagerly needed."

Most tellingly, perhaps, Conte points to "an increasing gap between evidence provided by pivotal trials and real world needs." He attributes this to a range of factors: patient selection ("elderly/comorbidities/co-medications are all underrepresented or not allowed in trials"); or modern technologies are not easily transferable to the real world setting.

Sometimes the place in therapy is outdated: the new treatment has been tested in a patient population which has not been pre-exposed to state-of-the-art treatments or compared to a suboptimal available comparator, says Conte. "Pivotal trials in most of the cases cannot address some important clinical questions such as sequences of treatments versus best treatments upfront." 

And safety and efficacy data from trials "are based on limited timeframes", while in the real world setting patients are often treated for much longer. He insists that long-term efficacy and safety data derived from real-world patients are necessary to better understand the risk-benefit ratio of innovative drugs.

His remarks found an echo among fellow speakers, and one concrete result so far is the creation of a manifesto among many of the contributors 'for a new approach for better medicine in Europe'. The next-and more difficult-step will of course be turning manifesto ambitions into on-the-ground changes in decision, regulation, or practice. That will provide plenty to be watching out for as the year evolves.

Peter O'Donnell is a freelance journalist who specializes in European health affairs and is based in Brussels, Belgium