Experts call for better clinical trials, improved reporting

EASD 47th Annual Meeting

LISBON — When it comes to clinical trial data, all that glitters is not gold, according to Rury R. Holman, FRCP, who discussed the balance of benefits and harms of when conducting clinical trials. Holman believes there is a need for a systematic approach to assessing clinical trial data and improved reporting.

Currently, there is a lack of consensus for the burden of truth required for drug licensing for clinical use or the withdrawal of a drug from the market due to safety concerns, according to Holman, of the Oxford Center for Diabetes, Endocrinology and Metabolism, University of Oxford, UK. The FDA recently required that all new antidiabetic drugs be assessed for safety with respect to cardiovascular safety outcomes; this strategy represents a disease-specific attempt to tackle the current issue. However, this strategy poses cost and timing concerns, Holman said.

‘Seeking the truth’

“Seeking the truth is what [it is] really about,” Holman said during a press conference. “Decisions in medicine are never absolute; they are made on the balance of probabilities. You take experience, information and guidance available, but often have to go beyond that.”

Robert M. Califf

For this reason, Holman said there is a need for a systematic approach to assessing rapidly evolving data from sources with variable provenance.

Observational data can identify signals of potential good or harm, but cannot assign causality, he explained. In addition, meta-analyses are only as good as the trials included; a trial that is not conducted to tackle the issue of concern or is not reported results in a flawed meta-analysis.

“There has been a rapid increase in the exploratory analyses of large-scale medical record databases. These are very helpful because they generate ideas [and] suggestions, and these issues of interest can then be explored further in approved trials. But these initial findings, as I will call them, are often reported, for whatever reason, as potential medical breakthroughs or new safety concerns and they pose inappropriate hopes and fears,” Holman said.

Such articles often give equal or greater prominence to a so-called new, exciting finding than to properly designed trials, he noted. A more rigorous approach for opportunistic studies of large-scale medical record databases is needed, in addition to reporting guidelines for the media, he said.

Improving clinical trials

According to Holman, in the last century there were three outcome studies in diabetes of note. During the first 10 years of this century there were about six; in 2011 there are 14 outcome studies in diabetes and researchers look to recruit more than 110,000 patients. The problem is that each of these trials will look at one drug and its safety profile for one purpose, he noted.

“This means that we are competing worldwide for patients and we are using money to run highly expensive trials. We need a better solution to look at several drugs in one trial,” Holman said.

This concept of the “pragmatic trial” is one that is advocated by Robert M. Califf, MD, of the department of cardiology at Duke University Medical Center.

“If we continue to do trials the way we are currently, they will take forever, we will not optimize the information and we will spend a lot of money collecting data, most of which is not very useful,” Califf said during the press conference. “What we need is more pragmatic networks that are efficient and reduce the cost of trials by a large amount.”

Califf discussed five steps that may improve the risk-benefit balancing of medications and improve patient outcomes.

The first step is a switch from pharmacologic studies to systems biology studies that involve both humans and preclinical models. The second step is performing particularly focused studies to demonstrate that the target endpoint was met. The third recommendation, which has become a standard due to regulatory efforts, is the ruling out of adverse events. Conducting global mega trials is the fourth step, tying in with the pragmatic trial approach. Lastly, Califf called for intelligent post-marketing surveillance and the democratization of information. – by Stacey L. Fisher

For more information:

• Califf RM. Defining the balance of risk and benefit for commonly used drugs.
• Holman RR. Seeking the “truth.” Warning signals from clinical trials – all that glitters is not gold. Presented at: The European Association for the Study of Diabetes 47th Annual Meeting; Sept. 12-16, 2011; Lisbon.

Disclosure: Dr. Holman has received research support from Amylin, Bayer, Merck and Novartis; attended advisory boards with Amylin, Lilly, Merck, Novartis and Novo Nordisk; and given lectures supported by Bayer, Lilly, Merck and Merck Serono.

Alan J. Garber

Clinical trials provide a focused answer to a severely limited question in a fixed and defined population. Since clinical patients present a wide spectrum of presentations and characteristics, clinical trials may not adequately address all clinical possibilities. To answer those unresolved questions or unstudied populations, we rely upon clinical judgment or data base analyses which are less perfect than trials.

– Alan J. Garber, MD, PhD
Endocrine Today Chief Medical Editor

Disclosure: Dr. Garber is a consultant/advisory board member and on the speakers’ bureau for GlaxoSmithKline, Merck, Novo Nordisk and Daiichi Sankyo.

Follow EndocrineToday.com on Twitter.