CRO Sounds Cautionary Note
October 24, 2007
As we noted in an article yesterday, the annual Regulatory Affairs Professionals Society (RAPS) meeting, held last month in Boston, had a fascinating session on adaptive designs from a regulatory perspective.
Another speaker was Mary Johnson, executive VP of biostatistics at PharmaNet. That contract research organization offers a unified platform of clinical trial technologies.
For her part, Johnson has previously worked as a consultant (for 8 years) at FDA. “What’s very clear in this business is the need to deal with issues up front in the early stages,” she told the audience. “If you don’t, FDA will.”
Johnson reviewed a number of biostatistical scenarios and considerations in trial design. It’s clear she’s been doing this for a long time. Many in the industry presume that adaptive trials will result in smaller numbers of patients—and fewer trials. But Johnson walked the audience through some hypothetical patient-recruitment scenarios in adaptive trials. In one, the sponsor would have needed twice as many patients as anticipated. Says Johnson: “You almost doubled your study size. Be careful what you wish for. You have potentially a much larger study.”
It’s also clear that she thinks adaptive trials are an extremely promising approach. “If we can monitor for early efficacy, we have the option to terminate the study altogether,” she says. “We can select the dose for new studies. If there is lack of efficacy or harm, we can terminate the study for safety reasons, or drop ineffective arms. The end result is really fantastic. We can shorten the time and cost of getting new drugs to market.”
It will be crucial to have access to fresh data, she said. Interactive voice response (IVR) or electronic data capture (EDC) systems to gather the data will be essential. Says Johnson: “You need it quickly and completely and to clean up very fast.”
Johnson doesn’t think it will be eary for all organizations to do adaptive designs. The need to coordinate with external data monitoring committees is one new challenge. “We will need them even more if we are doing sample size estimation,” she says.
She also thinks there may be tension between regulators and sponsors about canceling trials in the interest of efficiency. “The FDA may not want you to stop early,” says Johnson. “They may want the long term toxicity and safety data on your drug.”
She was blunt about the take away message. Talk to regulators about your statistical plan before proceeding with flexible or adaptive approaches.
Here’s how Johnson put it: “Whether FDA would like it or not, you better find out first. The bottom line is to plan ahead. That is my only message. Adaptive designs hold promise for the future. But there are challenges on the sponsor side and on the regulatory side. FDA is cautious. Independent committees are almost always needed. You have to have a third party to keep you honest.”d9A2t49mkex