How To Improve Clinical Trials – Some Good Old-Fashioned Wisdom
Intel cofounder Gordon Moore predicted in 1965 that the number of transistors on an integrated circuit would double approximately every two years. What came to be known as Moore’s Law has basically held true ever since.
The reciprocal effect of this increased microchip performance is decreased computer costs. Similar economic gains have been made in other information-intensive industries (e.g. telecommunications). Unfortunately, the same relationship does not exist for clinical trials, which have witnessed steady cost increases of 7.4 percent annually over inflation for the past 20 years — despite the advent of various technologies (e.g. electronic health records and electronic data collection) that significantly reduce the amount of time and effort required to gather clinical data.
In a discussion paper (April 13, 2012) presented to the Institute of Medicine’s Forum on Drug Discovery, Development, and Translation, Duke University Medical Center researchers Judith Kramer, M.D., and Kevin Schulman, M.D., note that the problem with escalating clinical trial costs is the result of sponsors, CROs, regulatory agencies, and providers failing to evolve the clinical research business model to keep pace with advances in technology. As a result, rising clinical trial costs have played a pivotal role in skyrocketing drug development costs.
Performing clinical trials as they have been done for decades is no longer acceptable. Technological and methodological advances have afforded sponsors the opportunity to increase the quality of data collected while decreasing clinical time lines and costs. In order for the industry to continue to the extent it has, sponsors must evolve.
It was with this mindset that Life Science Leader magazine and NextDocs, a clinical content document management solutions provider, convened an intimate thought-leadership roundtable event in New Jersey with a goal of providing you with some best business practice information on how to improve clinical trials. In addition to myself, the participants included Mark Anderson, executive director, global head of clinical data management, Alexion; Greg Fiore, M.D., founder of SSI Strategy consultancy; Reinilde Heyrman, M.D., VP, chief clinical development officer, IKARIA; Mitch Katz, Ph.D., executive director medical research operations, Purdue Pharma; and Eva Krusinska, Ph.D., principal at PharmaLand Executive Consulting Services. Together they have a combined 125+ years of industry experience. And though you may not recognize some of their current employers, their past career experiences include more recognizable company names such as Abbott, Acorda, Amgen, Covance, Daiichi Sankyo, Eisai, Forest Laboratories, GSK, Merck, Pharmacia & Upjohn, Schering- Plough, and The Medicines Company. What follows is some good old-fashioned wisdom on how you can improve clinical trials — and it doesn’t require you to spend a fortune on the latest whiz-bang gizmo or gadget.
Are You Asking The Right Questions?
One of the first questions I posed during the group discussion was fairly straightforward: How can we improve clinical trials? According to Katz, a good place to start is by reducing the complexity of clinical trials. “We have used clinical trials to try to answer too many questions,” he states. “Louis Lasagna [respected expert in clinical pharmacology] taught us that if you ask too many questions, you are going to lose the focus of your trial.” Anderson agreed, adding, “If we don’t get that under control, we could be asking a lot of questions that really don’t gain us much information about the actual effectiveness of the drug or what could be commercialized as we go forward.”
The initial thought behind asking lots of questions in a study was that it would reduce costs, because it would decrease the likelihood of having to do an additional study. Though intuitively this makes sense, it does not fit with the reality of what happens in the field. “What we see coming back from our sites is that the more data you ask for, the more involved the trial will be,” says Heyrman. “This is not conducive for either patients or science, and it results in higher costs.” (For an additional example of how industry practices are not always conducive to patients, see sidebar, “Cookie-Cutter Informed Consent Creates Confusion, Not Comprehension.”) She advises always keeping in mind exactly what is needed for the successful registration and launch of a drug — probably in more regions of the world than just one. A simple technique taught to Heyrman by one of her mentors regarding clinical trial development was, “Ask yourself what exactly you think you’re going to do with every single piece of data. If you can’t answer that straightforwardly, the next question should be: ‘Why exactly then are you even collecting it?’” For example, while working at a different company more than 10 years ago, Heyrman relates that one of the clinical trial procedures involved the collection of concomitant medications. “While we found out what concomitant medication was being used, we never collected the dose or administration route,” she states. This changed when she went to another company that considered it heresy not to collect dose, mode of delivery, and frequency of concomitant medications, though she admits that in the absence of an adverse event, little, if any, data analysis was ever conducted.
Much of the “collect-and-monitor-everything” mantra that has existed in clinical trials was driven by the legacy of Big Pharma deep pockets. Katz believes industry is shifting toward simplifying clinical studies, and this shift is being driven by smaller companies with limited resources making smarter decisions. “Because you have one shot,” he says, “you have to pick the best choices for your primary and secondary efficacy end points.” That being said, during some of Krusinska’s recent consultancy work she found the reverse to be true. “A client wanted to conduct a trial with six ‘arms,’” she states. “They wanted to compare each ‘arm’ with everything. It wasn’t a large pharma company with deep pockets, but a very small start-up that wanted to get everything done in one trial.” According to Anderson, this propensity in smaller companies to try to collect every piece of ancillary information persists out of the fear that a regulatory agency might, at some point, ask for it. He doesn’t believe this to be adequate justification for overcomplicating a trial and reminds, “If you fail to ask a question, you can always do another clinical trial.” In other words, don’t fall into the trap of trying to capture everything just to remain bound to the clinical trial time line. To ask better questions requires greater deliberation in the clinical trial design process at the beginning of the study design. This requires slowing down the overall process if you really want to speed up completing a well-designed study.
Speed Up By Slowing Down
A recent study of 343 businesses, published by the Harvard Business Review in conjunction with the Economist Intelligence Unit, revealed that businesses embracing business-accelerating initiatives in order to gain an edge ended up with lower sales and operating profits than those pausing at key moments. Over a three-year period, firms that slowed down improved top-line performance — averaging 40 percent higher sales and 52 percent higher operating profits. Fiore advocates applying these same philosophies in the clinical trial space. “I like the point about asking questions about what is going to be done with each data element,” he says. “By virtue of that conversation, it slows down the up-front design process.” Fiore believes this is just the starting point and supports applying QbD principles to the clinical trial design process.
A recent article by Frederic Sax, M.D., Quintiles’ global head for integrated drug development, provides a succinct framework for applying quality by design (QbD) principles to clinical trials — Plan- Do-Check-Act. The “Plan” phase requires what he calls design diligence. Heyrman believes it is this planning phase that serves as one of the biggest roadblocks to improving clinical trials. “If you really want the trial to go quicker, it actually requires more thinking up front,” she says. The challenge is the changing of the mindset, because it feels like you are starting slower. “It is so much easier to quickly design a trial. Get it all started. Then make amendments one, two, three, and so on along the way,” she states. “But how much money did we allocate by making the changes?” Heyrman notes that taking this approach is usually driven by what she describes as the “all-encompassing, first-patient-in syndrome” and doesn’t result in the trial being completed any faster. However, it does result in costing more. It may seem cliche, but Anderson advocates spending 90 percent of your time on the planning component of the trial and 10 percent on the execution. In the planning phase, Katz suggests gaining a greater understanding of the practice of medicine in the field when conducting global clinical trials. “On the front end, we need to involve the right people, who may or may not be key opinion leaders, but who know the practice of medicine in the relevant regions and what is really feasible when designing the study from a practical standpoint.”
Anderson suggests a good place for starting the trial design is to begin with the drug label and work your way back. “I’ve got to make sure I have a good product package insert in order to give me a good drug label,” he says. “This is going to allow me to get my product on the market. Further, a good label provides for a point of differentiation, which leads to greater commercial viability.” To that end, Krusinska advises you to talk to the commercial side and get them involved early in the process. “Don’t apply this approach to individual trials,” she states. “You have to involve the commercial team across clinical development plans.” Fiore adds, “We also need to think about reimbursement, so as to build in pharmacoeconomic elements as early as practically possible.” Katz agrees. “Where we have to get a little smarter is planning for third-party payer involvement. We need to partner with them before we even start Phase 3 trials.”
The consensus is that the good old-fashioned wisdom on how to improve clinical trials can be described by the word simplify and the following principles:
- Ask the right questions.
- Spend more time on planning so as to come up with designs that make sense.
- Gain alignment between industry and third-party payers for reimbursement purposes.
- Gain alignment between industry and regulatory agencies so you provide them with what they want.
In the words of Katz, “We have to get back to the basics.”
Simplifying Clinical Studies Provides Spirited Debate
During our recent clinical trial thought-leadership roundtable, one thing became clear — simplifying the clinical trial process is a desirable goal. To do so, members of the roundtable suggested that we need to ask the right questions — sparking the following spirited debate.
Katz: “So we put everything but the kitchen sink into a trial, and we realize at the end of the day we’ve collected a lot of rubbish. It causes us a lot of discrepancy resolution problems and costs us a tremendous amount of money. You’ve got to get simple.”
Krusinska: “You were mentioning complex designs leading to operational difficulties. It’s not only operational difficulties. Actually it does not work for science, either. If I have 200 end points, what value does it bring statistically?” Anderson: “What we need to look for from the beginning is — are we asking the right questions to start with?”
Katz: “But the questions have to be aligned with what the regulatory agencies are looking for. The reason we collect everything is because we anticipate that eventually the FDA is going to ask us a question about the data.”
Heyrman: “But honestly, you can give them the answer, ‘We didn’t collect that. Sorry.’”
Fiore: “As long as you demonstrate that you thought about it and had good reasoning.”
Anderson: “But the fear is, they’re not going to approve it, and you’ll have to do another study. That’s the issue companies struggle with in designing studies.”
Katz: “Agencies sometimes give contradictory recommendations or direction. They tell you one thing once, and then eventually, as they learn more, they’ll change their minds. It became very obvious to me that these agencies need to partner with industry. The regulatory agencies may not have the bandwidth to fully understand the implications and consequences of their changing decisions.”
Anderson: “We’ve got to change our way of thinking. It would be great for our government and other regulatory agencies around the world to say, ‘We want it this way.’ We are a conforming industry. If you tell us something, you’d be surprised how fast we’re willing to get behind it and align to it. I have yet to see something where guidance comes out and somebody says, ‘It is too difficult to comply with that guidance, so we’re going to submit the information the way we’ve collected it.’”
Fiore: “We tolerate extreme subjectivity in the industry, and Mark [Anderson], with what you’re saying, it would make sense if there were more clarity, because this subjectivity fuels the ‘I need to collect more, I need to be ready for any question a regulatory authority asks for.’”
Heyrman: “It’s really not the health authorities; that’s just a nice smoke screen we use to protect ourselves. If my scientist cannot explain to the data manager why we need to collect this piece of data, it probably isn’t worth collecting. If you can’t explain to somebody not in your field why you want to have what you want to have, question yourself as to whether you really need to have it.”
I have heard similar debates in other forums, with some folks advocating for specific regulatory clarity as to what is required. This is not likely going to happen — and for good reason. Though the FDA’s remit is safety and efficacy, it also needs to create an environment where people feel comfortable being able to question and challenge convention in order to spark innovation. Rather than seeking greater regulatory specificity handed down from the FDA, instead, the goal should be to strive to create a greater degree of collaboration and transparency between industry and regulatory agencies.
Cookie-Cutter Informed Consent Creates Confusion, Not Comprehension
Modality teaching theory suggests that people have different learning styles — visual (sight), auditory (sound), or kinesthetic (touch/physical activity), and that people learn best with educational programs customized to meet their optimal style. In traditional education environments, teaching has been implemented kinesthetically from kindergarten through third grade, visually from grades four to eight, and through auditory means (i.e. lectures) from ninth grade through adulthood. This cookiecutter approach has maintained widespread popularity despite numerous studies reporting its lack of effect.
Recent research found that interactive informed consent produced statistically significant higher test scores than the standard paper consent.
Though most learners use some combination of the three styles, one or more is usually dominant and defines the best way for a person to learn new information. Why then do we as an industry take a cookie-cutter approach to providing potential clinical trial participants with lengthy informed consent documents, filled with legalese, and expect them to truly understand what it means? Informed consent serves as the cornerstone of human-research subject protection. Yet proof of comprehension of the documents that people just signed is not required and is rarely obtained.
Next to the oil and tobacco companies, the pharmaceutical industry is one of the largest industries employing lawyers in the world. Pharmaceutical company legal departments charged with minimizing their company’s exposure have traditionally provided clinical trial informed consent documents in a onesize- fits-all format. This may satisfy lawyers, but it doesn’t really satisfy anyone else, and it does little to accommodate different patient learning styles, let alone facilitate their understanding. Mitch Katz, Ph.D., executive director medical research operations, Purdue Pharma, says, “It doesn’t inform the patient. It basically scares the living daylights out of them.” Some advocate e-consenting as a solution. “The beauty of the e-consent is that the patient can go home, get on a website, and read it again,” says Katz. “It gives you an opportunity to think about it more.” However, e-consenting doesn’t go far enough if it is only taking the lengthy document and putting it in an electronic format. “The biggest excitement I have had in the past five years has been my ability to now make information visual for people so they can make better and faster decisions on study data,” says Mark Anderson, executive director, global head of clinical data management at Alexion. “Information can be processed quickly on a regular basis, and there is the ability to generate graphs and other visual displays with some of the new software tools that really allow people to access specific patient data and explore interesting data trends.” How to get that information translated back to the patient and caregiver so they can make better decisions going forward is the next step Anderson believes will result in real breakthroughs.
Recent research found that interactive informed consent produced statistically significant higher test scores than the standard paper consent. Using an iPad, the interactive informed consent provided subjects with multiple options for hearing, viewing, and reading material. They can take a test and then get immediate feedback. Essentially, they have the ability to complete any form at their own pace. Subjects using the iPad spent nearly 10 more minutes viewing the informed consent than those using paper (22.7 minutes with the iPad vs. 13.2 minutes with paper). Even more telling is the fact that overall satisfaction and enjoyment slightly favored the interactive iPad presentation, despite taking nearly twice as long to complete.
I recently had the opportunity to speak with an executive of a top-10 pharmaceutical company testing an e-consent system similar to that described above. One of the initial findings was that it didn’t have the ability to print out the informed consent forms. When I inquired if the patient was able to take the iPad home to view the e-consent at their leisure or open up the document online via a home computer, the answer was “No,” illustrating why providing the patient with a printed version was so important. As research has demonstrated a greater level of engagement by patients using an interactive format, it is not clear why we want the patient to be engaged only when they are at the clinical trial site and not when at home. When you consider how the interactive approach accommodates multiple learning styles and improves comprehension, why then are we still sending the patient home with a printed version and expecting them to be able to adequately convey the consent information to friends and family in a single, noninteractive format? Greg Fiore, M.D., founder of SSI Strategy consultancy, says informed consent can’t be just patientcentric, but must be people-centric. “If you want to improve clinical trials, let’s start by improving the informed consent education process so as to increase comprehension and decrease confusion,” Fiore says.