According to a recent article from the Food and Drug Law Institute (FDLI), the Food and Drug Administration (FDA) in recent years “has struggled with an increasing number of disagreements among its scientific staff and there seems to be less unity around hard decisions and more instances where FDA staff labor to collaborate and work toward a consensus opinion.”
The article, published in FDLI’s Food and Drug Policy Forum, explained that, “a growing number of these disagreements are being brought before FDA’s public advisory committees for mediation.” The agency’s efforts to resolve scientific differences and to take final actions are “causing it to miss review goals that it commits to as part of the Prescription Drug User Fee Act (PDUFA).”
Part of the reason for these disagreements is the “broader political environment in which FDA must operate,” such as “Congressional and media scrutiny.” The political culture “spooks managers away from challenging scientists who cling to outlying conclusions, even in cases where those opinions are based on flawed analysis.” As a result, these internal disagreements are aired out during FDA public advisory committee meetings. The resulting transparency “insulates managers from charges that they “suppressed” dissenting views or intervened to guide the decision to a particular outcome.” This allows managers to let FDA’s external advisors sort out these disputes, rather than work to reach a definitive, internal consensus through the agency’s scientific process.
Structure of FDA
While this “practice may be politically expedient,” Scott Gottlieb, M.D., noted that, “it is an inefficient way to run a scientific enterprise and it also reduces FDA’s ability to leverage the collective wisdom of its scientific personnel.” Dr. Gottlieb, a practicing physician who has served in various capacities at the FDA and was a policy adviser at the Centers for Medicare & Medicaid Services, asserted that, “complex scientific views are the product of agreement formed among experts with diverse training, who hail from different disciplines and backgrounds.”
However, he noted that at FDA, “experts from different scientific disciplines work in silos,” and when “experts from these different silos disagree, there is a waning expectation that they should sort through their differing interpretations of data to reach a scientific accord.” Consequently, Dr. Gottlieb noted that the disagreements are too easily “turfed” to advisory committees to be resolved. Accordingly, he pointed out that while the “political culture is pressuring the agency toward these ends, the agency’s organizational structure may also be contributing to a lack of collaboration by discouraging opportunities for collaboration inside the drug review process.”
To examine the FDA’s organizational structure, Dr. Gottlieb described how FDA’s Center for Drug Evaluation and Research (CDER) adopted a “matrix” organization to improve accountability in its review process. Under this approach, separate offices were created to house personnel from the different scientific disciplines that were important components to the review of new drugs.
In a matrix organization, work is divided into projects. Dr. Gottlieb noted that, “the goal of moving toward a matrix was to enable specialization and create discrete project-managing units. This structured organizations around the task being accomplished. Each project is a distinct program and is looked after by its own project manager. The teams all have functional level employees who report to their respective functional managers and also to the project manager. At FDA, the functional managers are the leadership of the different therapeutic divisions.”
Among its advantages, a matrix is supposed to be an effective configuration for a project-based organization, enable resource sharing, cost reduction, and better coordination between multi-disciplinary and geographically dispersed teams. Although the structure of FDA’s current drug review process has merits, Gottlieb recognized that “it can also reduce collaboration and stymie decision-making, contributing to the agency’s challenges.”
For example, Gottlieb noted that inside FDA, among the distinct offices that were created as part of CDER’s configuration toward a matrix was an independent office for safety surveillance, chemistry, biostatistics, biotechnology, new drug quality, pharmacogenomics, and modeling and simulation, among others.
Under the resulting structure, the scientists from different offices that are part of the resulting matrix are “loaned” to the therapeutic divisions to work on the review of a new drug. Inside FDA, these loaned scientists are referred to as “co-locates.” They are supposed to work alongside the drug review teams inside CDER’s different therapeutic divisions. These “co-locate” scientists function as consultants to the clinical reviewers who are ultimately in charge of the drug file. While “this structure has brought managerial efficiency to CDER,” Gottlieb recognized that it also created “review silos, where scientists from different disciplines become segregated, and their input sometimes marginalized.”
FDA’s “weak” matrix needs to be reconfigured to give the management of the review team more control over key components of the scientific review process.
Gottlieb noted that part of the problem is that FDA’s organizational structure is largely what would be described in business literature as a “weak matrix.” Among other things, this means that the project manager has limited authority. “At FDA, the role of the project manager is to facilitate all aspects of the review. He or she is charged with moving along different parts of the process: the clinical review, the pharmacological review, the statistical review, as well as special assessments such as analyses of teratogenicity or carcinogenicity, or the analysis of the proper dose of a new drug.”
But in practical terms, “the weak matrix means that FDA project managers have limited dominion over key aspects of the review. Key disciplines involved in the review aren’t accountable to the project manager, or the division director. This weak structure also makes it harder to organize collaborative projects or even team meetings. The project manager doesn’t have strong authority when it comes to managing the collaboration between the different scientists involved in a drug’s review.”
He further noted that, “while the project managers report to their particular therapeutic division chiefs, most of the scientists involved in key parts of the review report up through different offices. In this structure, FDA staff is not attached to temporary management teams in a strong fashion. They might have multiple project teams that they are part of at any given time.”
Adding into the structural problem is the political scrutiny surrounding drug review decisions. Gottlieb noted that “individual scientific opinions become hardened” because “professional staff members who have a point of view that deviates from the prevailing opinion (especially when that minority view hews in the direction of a greater adherence to caution and a claim to promoting drug safety) have less incentive in this environment to work toward a scientific consensus.”
In addition, Gottlieb also noted the challenge of having various scientific disciplines and their respective silos all function as consultants to the review process. In most cases, they will complete their portion of the review, and forward on to the project manager a report summarizing their analysis and conclusions. He asserted that this process erodes the vigorous debate and collaborative process necessary for building consensus in science and medicine. Gottlieb added that the FDA matrix “can serve to isolate pertinent expertise, encourage the formation of obstinate opinions, and become a barrier to active collaboration.”
As a result, the mediation of conflicting views often can’t be achieved at the review level. It requires issues to be elevated to forums like advisory committees or regulatory briefings. While airing disagreements publicly affords the agency a measure of political cover from charges that it squelches minority-held views, Gottlieb asserted that the entire process is “inefficient and inconsistent with the way that scientific opinion is formulated in other endeavors.”
Work rules that FDA put into place to improve retention of FDA’s best scientists may inadvertently impede scientific collaboration and need to be reconsidered.
Next, Gottlieb explained how the matrix structure also “complicates the process for drug developers who try to schedule meetings with FDA or get timely feedback.” This is problematic sponsors must be in touch with each component of the review staff since different review silos are taking on a more prominent role in the drug review process, and at times, they are formulating opinions that are at odds with the clinical review team, which impedes the completion of a drug review.
Consequently, drug developers must interface with the project manager but also try and establish separate and direct dialogue with staff from the other review silos. “This obviates the whole concept of a matrix organization where the project manager takes responsibility for coordinating all the aspects of the review, and the primary review team is responsible for identifying and resolving issues with the application.” It also creates a process where participants in a review team are not only separated geographically, but they may rarely meet as a full team during a drug review.
The challenge of bringing together review teams is also compounded by the flexible work rules that allow reviewers to work off campus, making it more difficult to organize in-person meetings between review staff or hold face-to-face meetings with sponsors to provide feedback. FDA has a “maxiflex schedule.” Scientists can spend up to two days a week working from home and if someone works the required 80 hours in eight days, then they could get two days off. Additionally, reviewers telecommute.
However, FDA’s flexible work rules exist so that the agency could offset any incentive to leave and to remain competitive with academic institutions and drug companies for new talent. Additionally, the flexible work rules were a way to mitigate departures after CDER moved to the new White Oak campus, where the compute on I-495 is awful. Nevertheless, Gottlieb noted that the “flexible work rules make it difficult even for FDA to schedule internal in-person meetings.
“A final problem with this organizational structure is that there is no clear authority inside the matrix charged with resolving scientific disagreements. For example, there is no formal Chief Medical Officer (CMO) role or office that can mediate between different scientific opinions. The leadership in the OND is positioned to take on some of this role, but they do not have dominion over all of the parts of the matrix. Nor do they presently have the staff, or the mandate, to play this function. To be truly effective, the function would need to reside above the level of the matrix, presumably inside the immediate Office of the Center Director.”
“In recent years, the number of staff involved in the drug review programs has almost doubled; the scientific work has become more complex, involving more discrete disciplines; and the political scrutiny of FDA has become intense. As a result
of these and other challenges, FDA has gone in the direction of creating more discrete review silos—more offices, more scientific nodes and a more complex matrix. The agency might need to take on reforms that begin to move the structure of its drug review process in a very different direction that breaks down these silos.” Based on this analysis, Gottlieb makes four recommendations that FDA should consider.
Recommendations
Group more of FDA’s review staff around areas of therapeutic focus rather than scientific discipline in order to prevent fragmentation of its scientific staff.
One alternative is to consolidate the experts from different scientific disciplines back into the review divisions, rather than maintain them in the current silos. They would be organized as part of a single, multidisciplinary division. This structure could foster a more collaborative culture, where everyone is a more permanent member of the same team and make experts a more integral part of the clinical review teams rather than acting merely as consultants to that process, as they are today. Gottlieb also noted that if FDA’s scientific experts were more permanently embedded into the review teams, it might also foster more creative approaches to drug development.
Organize review staff based on a matrix that is focused around projects rather than review functions in order to foster more scientific collaboration.
One alternative to the matrix is the project-based organization (PBO). In a PBO, Gottlieb explained that core business processes are organized within projects rather than functional departments. Project managers typically have very high status and direct control over business functions, personnel and other resources. PBOs organize their structures, strategies and capabilities around the needs of projects. The requirements of individual projects often cut across conventional boundaries.
A PBO would allow therapeutic divisions at FDA to assemble teams based on the particular needs of a drug review. The training and expertise of people who are part of a review team could be closely matched to the challenges posed by a particular application. Projects might not comprise single drug reviews. They could also comprise classes of drugs or even specific indications. This would allow review teams to stay together over the course of multiple drug reviews and develop expertise in more narrow areas. Gottlieb noted that this organizational approach has become more common in the pharmaceutical industry as drug companies move away from their strong matrix configurations.
In the case of FDA, a PBO can be a hybrid with a functional matrix structure, where professional staff members from different disciplines are housed within a clinical division for longer durations. This would be instead of today’s approach, where they are assigned on the basis of individual drug applications. In this structure, there is still a management tie to their office, in addition to the therapeutic division.
For training purposes, they can maintain an affiliation to a functional office head. But they become a more integral part of the divisional staff and its review teams, as opposed to merely acting as consultants to review staff. The management of all review activity would fall under the direction of the therapeutic division head. The project manager would gain more ability to integrate the activity of different scientists into the review process and manage the components of the review. Gottlieb stated that this would make it “easier to organize meetings and to surface and mediate scientific disagreements since the disputes and their resolution would occur within the confine of a single, collaborative team.”
Create an Office of the Chief Medical Officer as part of the drug review structure, to mediate conflicting opinions between review staff.
FDA might also consider the creation of a discrete chief medical officer role for mediation of conflicting opinions between the scientific silos. This can take the form of a new office that would sit above the OND. The new office could be housed in the immediate Office of the Center Director, where a CMO could be properly empowered to help mediate disagreements that might arise not only within the OND but also between OND and other offices such as the Office of Surveillance and Epidemiology.
Assemble review teams that work on multiple projects and build cohesion across several reviews, rather than rely on temporary work groups that are brought together for each individual application.
Finally, if FDA’s drug review leadership continues to believe that its current matrix organizational structure is the best compromise between competing alternatives, it could align review staff more firmly around the projects that they are tasked with working on, rather than organizing them principally around their scientific disciplines. One alternative might organize staff based on project-based teams that are brought together for more sustained periods of time. One way would be to allow review teams to work on multiple drug applications before the teams are disassembled and members assigned to new projects.
Conclusion
In the end, Gottlieb noted that FDA may face practical challenges migrating from its current configuration. While changing the current matrix configuration may meet some resistance from inside FDA, he recommended that sponsors should insist on an analysis of FDA’s current management configuration by a leading management-consulting firm in order to stimulate improvements in FDA’s organizational structure, This could become a condition for re-authorizing the Prescription Drug User Fee Act in 2012 (PDUFA V). An objective analysis of FDA’s organization might identify ways to improve the current management paradigm, or migrate to a different orientation altogether, for the ultimate benefit of patients.
Ultimately, Gottlieb concluded that, the current political environment celebrating the bickering around difficult FDA reviews needs to be changed. “Quality science requires experts from various disciplines to work together in a collaborative fashion and reach an informed consensus.” This is the way physicians come to resolution on hard medical questions. This is how good science is conducted. Accordingly, Gottlieb asserted that, “when the drug agency is repeatedly unable to come to a common view on difficult questions, it is a good sign that FDA is not developing its opinions in a framework that makes optimal use of the agency’s expertise.”