Shire Makes Case for Patient-Centric Protocol Development

Optimized Model Sticks to Scientific Merits and Mitigates Risk by Reducing KOL Influence

By Marc Dresner, IIR

In reviewing my notes from the 2011 Partnerships in Clinical Trials conference in preparation for this year's event, I came upon an intriguing and somewhat controversial proposal by Shire Clinical Operations Director Joe Kim that I thought warranted a revisit in this forum.

The workshop session, titled "Trial Optimization Strategies in a Fully Outsourced Model," outlined a process for enabling teams to retrospectively build a business case for patient-centric clinical trial protocol development.

With the increasing focus on personalized medicine, this presentation may be even more timely today than when it was originally presented.

My summary...

“The protocol must fit the patient,” said Shire's Joe Kim (pictured left). “Recruitment does not begin with outreach; it begins with getting the protocol right.”

Kim also suggested sponsors should reconsider the extent to which protocol development depends on input from KOLs.

“KOLs are necessary and I wouldn’t say we need to get rid of them,” he qualified, “but their experience—while valuable from a scientific standpoint and from a patient care perspective—boils down to opinion, albeit expert opinion, and it’s not always correct.”

Literature is an excellent supplement, Kim added, but it may likely include past mistakes, which, if left unidentified and unresolved, the trial would be doomed to repeat.

Lastly, Kim pointed out that the final protocol is never really final, because it is revised and amended throughout. “You’ll never get rid of amendments,” he stressed, “so we’re left to the whims of expert opinion, baked-in past mistakes and a process inherently conducted on the fly.”

To remedy the situation, Kim proposed adopting a model that holds only to scientific merit and safety build-ins, removing any layers for which there is no data support—a minimizing endeavor where all but the proven essentials are stripped away and nothing short of the critically important is added.

To illustrate, he presented data from Centerphase, which had analyzed a protocol for his team. “Very easily I can pull the analysis out, pop it into a slide and say here’s what’s happening to your protocol. For criterion one, we get 65% drop-out; criterion two, nearly 80%...So you can start having a meaningful discussion about the relative importance of criterion two to the scientific merit of the protocol,” he explained.

“We can also look at sequential loss, overall loss,” he continued. “We can get actual numbers around a particular site and identify the available pool of patients.”

Kim noted that many suppliers today offer this capability. He also emphasized that data reporting and visualization tools have significantly improved. “You can quickly and easily pull together a case to make to senior management. Turnaround time is much shorter and costs are far lower than in the past.”

So what implications does this tack have for protocol development? “We’re never going to get away from the old model, but we have to add data into this and triangulate and compromise where we can,” Kim said.

“We can then hopefully arrive at an optimized protocol with all of the opinion and misconception removed from the equation because we’ll have data to say whether this decision is going to improve recruitment or not. Even if senior management says, ‘I don’t care, keep it in,’ you’ve articulated the risk.”

Kim added that as organizations become more adept at optimizing protocols, it stands to reason that the number of amendments—at an average cost of roughly a half-million dollars each—will be reduced. In short, Kim argued that it’s far less expensive to have a qualified provider conduct this kind of analysis than it is to pay for an amendment.

With this in mind, Kim outlined an approach for building a retrospective business case to convince management at sponsor organizations to implement the model.

(Editor’s Note: Prior to the workshop, attendees were given an email assignment related to how this model might fit in to their own protocol development processes.)

Kim first identified a protocol that had been amended for I/E criteria. He then obtained the pre-amended protocol and had it analyzed for elements containing I/E criteria. Once these were identified, he measured the number of patients that were enrolled due to the amendment.

To be clear, this was NOT the number of patients that were enrolled post-amendment, but those that fit the new I/E criteria, and how many were consequently enrolled. He then extrapolated the potential enrollment savings had the I/E criteria been optimized up front. I.e., How many more patients could we have been recruited over the course of enrollment?

On top of that, Kim calculated the cost to process the amendment, which he admitted was relatively easy because Shire outsources the entire process. “You basically look up the CRO change order and see how much it cost,” he explained. “It’s not just direct IRB fees; it’s everything involved. If you don’t outsource and you use an internal CRO, you need to make judgments about the FTEs required to process an amendment—and that’s fairly easy to find, too. You just talk directly to the study manager or to your reg department to get a sense of FTE costs.”

For example, Kim supposed one had a study of n=400, which took 10 months to enroll and included an amendment at month six (four months of enrollment under the revised protocol). Looking back, you find that there were 30 patients enrolled in the first six months that also met the new I/E criteria in the amendment.

“Do a little math and you can say we enrolled an additional 7.5 patients per month under the new amendment,” he estimated. “Extrapolate that back and it comes to a total of 45 patients over months one through six. So 45 over 400 is greater than 10%, which means I could’ve shaved off more than a month had we had the enrollment up front.”

“By doing all of this upstream,” he added, “you could determine exactly what the cost was. It’s a slam-dunk case. I could’ve saved half a million dollars and enrolled it full a month earlier if I had only spent $15k analyzing the protocol.”

One qualification: If you send the pre-amended protocol to be analyzed and they don’t find anything, this could mean that the amendment was introduced in vain. “Perhaps a decision was made to change the I/E criteria under the assumption that the change would produce more patients, when in fact it didn’t. I think it’s uncommon, but possible,” he concluded.

Final note: Kim referenced the potential of electronic health records, but for the purposes of this summary, I decided to table that topic for another day.

EDITOR’S NOTE: The 21st annual Partnerships in Clinical Trials conference and exhibition—the world’s premier gathering for clinical development leaders and practitioners—takes place March 4-7 in Orlando, FL.  As a reader of this blog, when you register to join us an mention code XP1700BLOG, you'll receive a 15% discount off the standard rate!

For more information or to register, please visit www.clinicaltrialpartnerships.com.

ABOUT THE AUTHOR/INTERVIEWER
Marc Dresner is IIR USA’s sr. editor and special communications lead. He is the former executive editor of Pharma Market Research Report, a confidential newsletter for market research professionals specializing in the pharmaceutical, biotech and healthcare industries. He may be reached at mdresner@iirusa.com. Follow him @mdrezz. Meet him at the 2012 Partnerships in Clinical Trials conference in Orlando this March!