
Rosalind Cheetham explains why correctly forecasting clinical trial recruitment levels should be a priority for the pharmaceutical industry.
“Traditionally we would finalize the protocol, get sites to commit and then we would hold an investigative meeting.”
-Rosalind Cheetham
Ever since James Lind’s demonstration in 1747 showing that citrus fruits could cure scurvy in sailors, clinical trials have been part and parcel of understanding how external substances – drugs or otherwise – can be applied to affect the biological state. In fact, the only thing that has significantly changed since then is who the facts need to be proven to; from the individual to the commercial, the citrus fruits of the 18th century have become today’s blockbuster drugs.
Yet with pipelines drying up and innovation sorely needed, the pharmaceutical industry is having to contend with rising costs, patent battles and a plethora of rights and welfare red tape just to reach the stage of clinical trials. Someone who knows the contour of this land better than most is Ros Cheetham, VP of Neurosciences MDC and Medicine Development Leader for GlaxoSmithKline (GSK).
“First and foremost,” begins Cheetham, “whenever we think about anything to do with recruiting participants into clinical trials, everything that we do has to be primarily focused on protecting the rights and welfare of potential trial participants and everything that we do must be grounded in that fact.
“Taking that into account, in terms of sites being overly optimistic about their ability to recruit, I think conducting clinical trials is a very complex issue and there are multiple factors that impact both a site’s ability to successfully conduct a trial, as well as trial participants’ willingness and eligibility to participate in one.
“As a sponsor of a study, we often seek study feasibility information from sites using protocols that are in a fairly early draft format, so they may not be complete at that time in terms of their inclusion and exclusion criteria. We do find that those initial recruitment estimates that sites make can be very heavily affected by later, additional or changed inclusion criteria in the protocol. As such, we sometimes go with incomplete information or information that is later changed that makes it more difficult for sites.”
Appreciation
However, the list of reasons for a participant’s willingness to enroll is often exhaustive and unpredictable. A change in the number of visits a group might have to make could have a significant impact on people, as could the length of a visit. For example, if visits were initially little less than an hour, and then rose with the inclusion of tests and examinations to two or three hours, that could also act as a deterrent to people participating in the trial.
When companies initially go to sites, they can find themselves with issues caused by not managing to access the most appropriate site personnel to make those recruitment estimates. By contrast, some sites may have different inclusion criteria that make participants easier to recruit. All of these can, and often do, lead to inaccurate forecasts.
“There are plenty of variables,” continues Cheetham, “and one of the things that we are starting to do is to hold an investigative meeting before the protocol is final. Traditionally, we would finalize the protocol, get sites to commit and then we would hold an investigative meeting. This either means that they have to live with the protocol the way it is or we’re forced to make amendments, which are very expensive and not the most cost or time effective way of doing things.
“However, if you hold an investigative meeting before the protocol is final, where the inclusion and exclusion criteria, required tests, examinations and visit frequency are discussed in detail – and where there’s peer discussion around the practicality of the trial – there’s still an opportunity for the sponsor to change that protocol and to positively impact the quality of that final protocol. In doing so, it impacts the site’s ability to forecast recruitment more accurately and then deliver on those forecasts.”
If this is the case, then it raises the inevitable question of why this hasn’t it been done in the past. “I think it’s a combination of things,” admits Cheetham, “but to do it this way round does require an infusion of money to cover the cost of that meeting before you’ve made your final decision on participation. It could also be viewed as potentially slowing down the process of starting a trail, but I think that those are very short-term gains where you might lose in the long run by it taking longer to recruit or being more expensive to recruit because of the need to add additional sites or make later changes to protocol. Those upfront costs, and potentially more time, are outweighed by the potential benefits. Perhaps we haven’t always appreciated that.”
Recruitment
On that note, appreciation certainly seems to be the key when looking to improve on recruitment forecasting and an ability to recruit people more efficiently. The first in a long line of tools that companies feel can aid in estimating more efficiently is at the site or sponsor level; a strong, historical database that documents success in recruiting for other similar trials seems to be the most fundamental tool in this context.
“If that’s not available in-house,” offers Cheetham, “then it is possible to purchase such databases or access to databases commercially. I think we often fall into the trap of assuming that recruitment will be linear across the duration of a study – and historically that’s not the case at all. I tend to use the analogy of recruitment usually following the shape of a hockey stick. Initially sites are slow to get off the ground and, often down to a lack of familiarity with the protocol, some things need to be ironed out so the initial recruitment is usually quite slow.
“But once sites get into the recruitment phase and gain more familiarity with the protocol – in turn allowing them to see the results of some advertising for participants – then it accelerates rapidly and you get this hockey stick type of picture. There are tools with which you can predict non-linear recruitment in that sort of paradigm. From that, you can help sites better understand what their workload is going to be at a particular point. It also helps sponsors plan for the high workload at the end of the study to get the data in-house – but a tool where you can plot variable recruitment is likely to give you a far more accurate prediction of the likely recruitment rate.
“Along a slightly different track,” continues Cheetham, “something that we have looked into on a number of occasions and I believe can be extremely helpful is getting sites to participate in a feasibility protocol. Normally, we just ask sites to fill in a questionnaire about their expected recruitment, but you can do something that would require IRB and ethical committee approval whereby sites would record high-level details about potentially eligible subjects over a two- to four-week period. For example, if you were conducting a study in bipolar disorder, you would ask them to record some very basic anonymous information for each subject with bipolar disorder that they witnessed over a four-week period.
“From that, you can make an assessment on how many of those patients would have been likely to be eligible for the trial and may have agreed to participate in the trial. That certainly gives you more accurate information. Again, there is a cost and time element associated with that – but it’s still undeniably more accurate than the historical side of things.”
Quality
Discussing the possibility of identifying sites that will result in a higher enrolment of higher quality subjects remains a subjective area. First of all, a sponsor needs to define precisely what constitutes a high quality site and quality trial patient. The problem stems from the fact that people often have different views on the subject; potential trial participants, their disease and sustainability for the trial, an ability on behalf of the investigational site to fulfill good clinical practice and an accurate report of their data in a timely manner are all factors that fluctuate dependent on the person in question.
Another inherent challenge a sponsor needs to consider is how high it is willing for recruitment to be at any one single site; if anything over 50 percent of trial participants emerge from one specific site then a bias sentiment could begin to prevail. Statistical considerations that must then be included to get the full picture. What companies such as GSK are trying to avoid is sites that enter no subjects at all. Surprisingly, sites that enter just one or two participants are also a problem because they have the potential to introduce variability into the results.
“I think that to identify excellent sites requires in-depth discussions with the appropriate site personnel and those discussions should, if possible, be face-to-face. Moving away from the impersonal paper-based, or even internet-based questionnaires, to a much more personal discussion where you can really discuss their experience, their past trial performance and honestly assess their ability to participate in the trial in an extremely transparent manner. Ultimately, it comes down to a relationship.
“Those face-to-face discussions, especially with investigative sites that you haven’t worked with before, are very important. For sites that you have managed to work with in the recent past, a telephone conversation would probably suffice. There are definitely some variations on a theme here, but I think you need to promote some personal contact to have those honest discussions.”
Obviously there is plenty for companies to keep in mind, but all of this becomes magnified once it reaches a global scale. Without a doubt, the biggest benefit of conducting global trials is the possibility of opening up access to larger populations of potential trial participants. The biggest challenge in working on the global basis is the dealing with the different medical practices and treatment paradigms used in other countries.
“A careful discussion with the sites about whether their patient population looks the same as what you’re expecting in other countries and whether their normal standards of treatment are the same, or very similar, is important, so that you’re getting a more consistent population of patients being entered into the trials. Again, there’s plenty of scope for clear discussion with the local staff who will be speaking with sites about expectations, what the requirements for those sites are, sharing definitions around quality and encouraging them to have those face-to-face or telephone discussions with sites to truly understand their patient base, treatment practices, trial experience and ability to comply with all the trial requirements.
“Certainly, our philosophy is to conduct global trials and to use clinical sites in countries where not only are all these factors fulfilled, but where we also don’t see barriers later on to potentially market the medicine if we get to that point. Global trials open up vast possibilities but the same issues apply, and it comes back to clear communication and a definition of exactly what you’re wanting, and if possible, that early opportunity to look at the protocols and give specific feedback about what will and won’t work in their country.
“It can be as simple as a rescue medication that is widely available and used in the US that a company may not have, but may have a very acceptable alternative. Making the protocol able to accommodate both those alternatives early in the process is going to make things easier for sites to recruit successfully.”
Biography
Ros Cheetham is Vice President for Neurosciences MDC and Medicine Development Leader for Glaxosmithkline.