FDA New Drug Product Reviews

FDA’s mission is to protect and promote US public health. One of the ways it achieves this mission is by authorizing new medical products to be marketed based on the data submitted in a marketing application to assess their safety and effectiveness [5]. In 2019, FDA began rolling out a new interdisciplinary approach to its review of marketing applications for new drugs, with the key new feature being the required use of an integrated, collaboratively written review document and a more interdisciplinary, issue-focused review team approach [Reference Woodcock, Stein and Bugin6]. This approach enhances the traditional multidisciplinary review approach that involved scientists from multiple disciplines working together in a separate but coordinated fashion to assess the safety and effectiveness of a new drug product by building in earlier and more frequent collaboration and a joint review document written as a team.

Multiple FDA working groups gathered internal and external stakeholder feedback on the traditional review approach to marketing applications and identified several problems, including redundancy within review documents and an insufficient focus on the key review issues associated with marketing application reviews [Reference Hearns-Stewart, Farley and Lee7]. Review issues are the “problems” that the FDA review teams must work through in a cross-disciplinary way to understand and, if possible, solve [8]. These are issues related to a drug’s benefit and/or risk, and they affect approvability or lead to other FDA regulatory actions. Based on these challenges, a new more interdisciplinary approach was developed.

Marketing application review, both traditional and the new interdisciplinary, begins with presubmission activities, such as a pre-new drug application (NDA)/biologics license application (BLA) meeting with the new drug’s Sponsor (or manufacturer). Upon submission of the application, the review team convenes to discuss its suitability for filing (e.g., is it complete?), then typically meets again at a halfway point (i.e., the Mid-Cycle Meeting), and again late in the review cycle, and lastly to wrap up and make any decisions (i.e., Wrap Up Meeting). For some applications, an Advisory Committee Meeting is held. NDA/BLA are the regulatory terms for the marketing application for a new drug (see Fig. 1). In the interdisciplinary review, review teams conduct several new activities to promote greater integration and focus on review issues. First, a benefit-risk scoping meeting that involves the entire FDA team and senior leadership is held prior to accepting (i.e., filing) the marketing application, where review issues can be identified early and monitored for the remainder of the review. Then for each review issue identified, a joint assessment meeting (JAM) is held. JAMs are problem-focused, interdisciplinary working meetings that involve all team members with a relevant perspective on the issue; these meetings enable the team to identify and then work through the key review issues collectively, increasing collaboration and communication. Integrated JAMs are highlighted in Fig. 1.

Fig. 1.

Integrated review approach: New drug application/biologic licensing application (NDA/BLA) are the focus of the review teams’ assessment. New integrated meetings are indicated: Joint assessment meetings (JAMs) are the problem-focused interdisciplinary working meetings prescribed in the new interdisciplinary review [9].

In both traditional multidisciplinary reviews and new interdisciplinary reviews, FDA’s review teams use a conceptual framework, known as the structured benefit-risk framework (BRF), to guide the assessment of benefit and risk for the proposed products [8]. It is through the examination of benefit-risk concepts that FDA review teams identify review issues and formulate an integrated assessment of benefits and risks, ultimately determining whether the benefits outweigh the risks in the context of the proposed treatment (see Fig. 2). The review issues encountered by teams assessing the benefits and risks of a NDA tend to concern the clinical evidence and uncertainty related to efficacy and/or safety signals, such as study endpoints, design of or results from clinical trials, dosing and pharmacokinetics, and subpopulation factors. In the new interdisciplinary reviews, review team assessments are documented in an integrated way in the new integrated review template, which replaces multiple separate, discipline-specific review templates that existed in the traditional multidisciplinary review.

Fig. 2.

Benefit-risk assessment: Review issues are considered by individual team members and also collectively before determining that the benefits outweigh the risks, representing a favorable decision [Reference Hearns-Stewart, Farley and Lee7].

FDA’s implementation of the interdisciplinary assessment of FDA new drug marketing applications has taken a phased-in approach to enable careful evaluation of its new features as its process and integrated review template are rolled out. An update on this phased approach was recently provided during a virtual public workshop in October 2020 [10], in which FDA highlighted internal and external support for the enhanced collaboration and communication that the new approach was promoting, but noted concerns with transparency in review documentation on reviewer disagreements, the independence of reviewers, and the need for more information regarding development program issues. These concerns echoed those discussed in publications earlier in the year regarding the lack of transparency and knowledge loss in the new approach [Reference Woodcock, Stein and Bugin6,Reference Herder, Morten and Doshi11]. This clearly highlights the need for a robust method of evaluating integration in the new review approach to weigh what is gained against what might be lost.

IPO Approach to Modeling Integration

To conduct this comparative evaluation, we needed a general way of modeling integration. Although integration is widely regarded as a critical feature of interdisciplinarity [Reference Klein, Repko, Newell and Szostak2,Reference Repko and Szostak12,Reference Bergmann13] and has received close attention in several literatures (e.g., philosophy of science [Reference Brigandt14], biology, [Reference Green, Wolkenhauer, Pulverer, Schnapp, Rembold, Breiling, Senyilmaz Tiebe and Breithaupt15] education [Reference Akkerman and Bakker16], cross-disciplinary theory [Reference Boix Mansilla, Lamont and Sato17,Reference Hoffmann, Thompson Klein and Pohl18], organizational psychology [Reference Salazar, Lant, Fiore and Salas19]), there are few theoretical accounts presenting a systematic framework that can be deployed to model the integration present in complex processes, such as NDA reviews. O’Rourke et al. is an exception [Reference O’Rourke, Crowley and Gonnerman20]. It develops a philosophical framework and an associated input-process-output (IPO) model for cross-disciplinary integration that are grounded in a close and wide-ranging review of the literature on integration, especially in philosophy of science and interdisciplinary theory; the model is designed to synthesize insights into the nature of integration found across a large swath of this literature (see Fig. 3). Further, it is designed to be a general model of integration systematically applicable across a wide range of differences, such as the differences obtained between a multidisciplinary review and an interdisciplinary review [Reference Laursen and O’Rourke21]. Both types of review may exhibit integration, and the model used should be able to account for that [Reference Newell22]. Finally, the IPO character of the model in O’Rourke et al. aligns closely with the structure of NDA reviews − there are clear inputs, a review process, and judgments about the efficacy/safety of the drug as output. The model thus allows for systematic identification of integration exhibited by NDA benefit-risk reviews.

Fig. 3.

Schematic diagram of input/output process model of integration [Reference O’Rourke, Crowley and Gonnerman20].

The O’Rourke et al. IPO model presents integration as a “generic combination process, the details of which are determined by the specific contexts in which particular instances of integration occur” (p. 67) [8,Reference O’Rourke, Crowley and Gonnerman20]. In this “generic” process, inputs are combined in the production of outputs. This is an application of the widely used IPO model, which has proven very useful across a wide variety of domains [Reference Kozlowski23,Reference Funnell and Rogers24]. The utility of the IPO model from O’Rourke et al. derives from the analytical framework it offers that can be used to organize “one’s thinking” about the inputs, outputs, and integration process variables of the model and “facilitate comparison” across these categories (p. 68) [8]. The model is sensitive to qualitative and quantitative characteristics of inputs and outputs, accommodating concrete and abstract elements and recording the number of elements and degree of difference among them. O’Rourke et al. understand integration to be a process, and so the process elements are central to the model, and in particular integrative relations. The integrative relations involved in the process activities, or changes to inputs, were assessed as either Synthetic or Combinatorial, relying on descriptions of these relationships from O’Rourke et al.:

Synthetic − brings together inputs in some way for irreversible integration. Combinatorial − an assembling or combining of inputs but is of low change to the inputs (e.g., stacking).

Integrative relations contrast with disintegrative relations, which reverse integration present in the inputs. (Preservative relations are third relation, but they do not figure into our analyses.) The disintegrative relation involved in the process activities is analytic:

Analytic – changes aimed at breaking an input down into its constituent parts or to differentiate between the inputs.

The model views integration as consisting in placing inputs into integrative relations (e.g., subsumption, serialization) in the production of outputs, where the number of inputs will typically exceed the number of outputs. The final pieces of this model are parameters (e.g., scale, comprehensiveness) that allow the model to be applied across a wide range of contexts (see Fig. 3). A contextualized IPO model, created through the coupling of the O’Rourke et al. cross-disciplinary integration IPO model with the structured BRF for new drug reviews, would further enable data collection and analysis for this study (see Fig. 4). Contextualization was possible through the FDA researcher’s experience, which was related to process knowledge of how FDA Review Teams operate, familiarity with the context of drug development, and awareness of benefit-risk review issues. This experience also helped with the coding and the analysis/assessment of different dimensions of the inputs, processes, and outputs.

Fig. 4.

Contextualized input-process-output (IPO) benefit-risk framework: This contextualized IPO model merges the key dimensions of the benefit-risk framework (i.e., evidence and uncertainty related to review issues that influence the assessment of benefits or risks) and examples of how the inputs, process activities, and outputs would appear in the context of a new drug product marketing application assessment.