Teamwork + Trust
The development of anti-TNF therapy
Is the current method of conducting clinical trials adequate for determining the impact of candidate drugs on a disease process like rheumatoid arthritis?
A trial is by definition a very artificial entity in that you have exclusion and inclusion criteria, which define populations very rigidly. Some people would argue that defining patients before you enter them into a trial often means that you’re loading that trial in favor of the patients that are most likely to respond, or patients that are least likely to show side effects...
Already the regulatory authorities are mandating so-called phase four studies. Once the drug is licensed, companies are still required by regulatory authorities to keep information about adverse events, for example. For expensive drugs, I think more and more payers are insisting on some kind of evidence of effectiveness in the real-life situation…
In a world where resources are limited and health budgets are under strain,… some kind of objective evidence that they are doing good and not harm is part and parcel of what I call post-marketing surveillance. That’s a different question from whether you can have better trials. And the answer to that is yes.
Traditionally in most diseases you start off with patients that are the sickest, as was the case with anti-TNF drugs. It isn’t the best population to see the best result in, but these people are in a terrible mess usually by the time they get into the trial because they’ve failed everything else, they’re often debilitated and sick as a result. Their resistance to infection is low because they’ve become immobile or they’ve put on weight—all the other factors that encourage what we call co-morbidity.
Obviously they are often not the best candidates to include in a trial. But if you have an agent of unknown safety, usually ethical issues demand that you start to gain full consent from a patient population where it’s unlikely that you’re going to do them harm. That’s a difficult question.