perspec tives
nature publishing group
conference proceedings Workshop on Predictable Outcomes September 30–October 1, 2008 Washington, DC
Workshop on Expediting Patients’ Access to Medicines March 30–31, 2009 Surrey, UK
Expediting Patients’ Access to Medicines by Improving the Predictability of Drug Development and the Regulatory Approval Process L Liberti,1 A Breckenridge,2 HG Eichler,3 R Peterson,4 N McAuslane1 and S Walker1 Ideally, well-designed global pharmaceutical development programs that include simultaneous submissions to multiple regulatory agencies can result in predictable regulatory approvals and expedited access to medicines for all patients. The workshops described herein investigated current trends in development and submission strategies along with regulatory review performance data in order to consider whether barriers to predictable expedited approval outcomes can be overcome through innovative clinical development approaches and a better understanding of review processes and procedures, particularly as these relate to the assessment of a product’s risk–benefit profile. Many pharmaceutical development programs are based on a global strategy that includes the option of simultaneous dossier submissions to key regula-
tory agencies, with the expectation of a streamlined approval process. The International Conference on Harmonisation of Technical Requirements for Registra-
tion of Pharmaceuticals for Human Use (ICH) has made tremendous strides in harmonizing regulatory quality, safety, and efficacy requirements for medical product registration across Europe, the United States, and Japan, and the expansion of ICH to include the Global Cooperation Group and the establishment of regional harmonization initiatives have also served to facilitate the goal of simultaneous global development and submissions. Despite these advances, however, a perception exists that simultaneous submissions may face unpredictable approval delays, sometimes associated with unexpected requests for additional clinical data or region-specific information that increase the chances of rejection or dissimilar review outcomes from different regulatory agencies. Therefore, companies and agencies are addressing ways to improve the consistency of the submission dossier. These include innovative trial designs and new technologies that have the potential to better establish a new product’s efficacy and safety profile, provide data relevant to all reviewing agencies, and accelerate the review process. Two workshops organized by the Centre for Medicines Research International Institute for Regulatory Science (the Institute)—Predictable Outcomes (September 2008) and Expediting Patients’ Access to Medicines (March 2009)—brought together experts from regulatory agencies, pharmaceutical companies, and academia to examine the practicalities of achieving simultaneous submission as part of a global development program, as well as the critical success factors necessary to achieve predictable review outcomes.
1Centre for Medicines Research International Institute for Regulatory Science, London, UK; 2Medicines and Healthcare products Regulatory Agency, London, UK; 3European Medicines Agency, London, UK; 4Department of Pediatrics, Faculty of Medicine, University of British Columbia, Vancouver, British Columbia, Canada.
Correspondence: L Liberti (
[email protected]) doi:10.1038/clpt.2009.179 Clinical pharmacology & Therapeutics | VOLUME 87 NUMBER 1 | JANUARY 2010
27
perspec tives This report summarizes the key findings from workshop presentations and syndicate session discussions. Supplementary information can be found at http:// www.nature.com/cpt. Simultaneous submissions and predictable outcomes
Making the review of medicines more predictable. In the current pharmaceutical industry regulatory climate, authorities must balance ever-increasing pressures to expedite their constituency’s access to new medicines along with the pharmaceutical industry’s need for incentives for innovation against statutory requirements to comprehensively review safety, efficacy, and product quality assessments, meeting an expectation by all of shortened regulatory review times.1 A present reality is that, in spite of constant increases in industry research and development expenditures, there has been a steady decrease in the number of new molecular entities submitted for review, with only 10% of compounds that attain the first-human-dose milestone reaching the market.2 Thus, the introduction of measures to improve the likelihood of making a safe and effective innovative medicine available to the public in a timely manner remains an important common goal for agencies and sponsors. The regulatory approval system should be predictable for medicines developed in accordance with current ICH guidelines. However, it continues to take an average of 14 years for a pharmaceutical company to bring a product to the market,2 during which time regulatory guidelines and scientific advice may have changed to keep pace with scientific progress. These changes may be reflected in additional regional regulatory requirements late in the development stage that result in unexpected prolongation of development programs and extended regulatory review times and ultimately may be a factor in inconsistent review outcomes. In reviewing the critical success factors that contribute to the predictable development and regulatory review of new medicines, workshop participants identified early and frequent sponsor–regulator dialogue as a major positive contributor. Developed to encourage the open exchange of information, the Institute’s Scorecard 28
Project provides a structured communication mechanism through which pharmaceutical companies and regulatory agencies can evaluate one another’s processes and work products on dossier-specific activities. Results from a small pilot study of the scorecard indicated that companies would like to have improved access to agency reviewers and to see agencies develop processes to enhance regulator–sponsor transparency in the decision-making process; agencies would like companies to improve their extent of presubmission communications to better inform the agency of the nature of the dossier.3 The workshop participants agreed that feedback mechanisms such as the scorecard were most helpful when the results of the discussions were openly communicated and used as tools to effect change. The Scorecard Project is entering its next phase of development, i.e., validation of the tool with additional agencies and sponsors. Benefit–risk models can also form the basis for more open discussions between agencies and sponsors and should be
Discovery to development
Proof of concept
Phase III
Submission
applied as early in the development process as possible to enhance predictability of the review outcome. Although benefit–risk evaluation cannot be reduced to quantitative algorithms, the adoption of a structured, systematic, and transparent approach to such assessments would allow the equivalent, ongoing global assessment of a product across regions. Ongoing development and refinement of the current EMEA benefit–risk template4 integrating experiences from other models such as the Institute’s benefit–risk assessment framework, could serve as an assessment and communication model for other jurisdictions.5,6 A complete list of critical factors that can result in successful simultaneous submissions with lower risks of divergent outcomes was identified at the workshop (Figure 1). It is organized by development stage, and the key factors include early planning for target labeling and obtaining appropriate data to differentiate the new product from existing therapies. This planning should take into consideration
Early differentiation Input from key stakeholders Having agreed-upon surrogates and biomarkers Good safety screening
Clinically relevant dosages/forms Address of liabilities of established therapies Regulatory validation of tools Interaction with regulatory agencies Advice from key opinion leaders Focus on products with an acceptable safety profile
Choice of comparator Significant resource investment for long-term outcome studies Realistic labeling that accurately reflects the product’s evolving benefit–risk profile Adequate trial recruitment
A clear, complete, easily navigated dossier Validated tools that can demonstrate differentiation Meeting unmet medical need
Figure 1 Critical success factors in drug development organized by stage. VOLUME 87 NUMBER 1 | JANUARY 2010 | www.nature.com/cpt
perspec tives
Table 1 Measures to improve the predictability of regulatory review Prospective measures
Retrospective measures
Identify cases in which there are no regulatory precedents for the introduction of new technologies and concepts. The phase III time frame could be used to educate agencies about new or complex compounds and identify data gaps required for the agency to make an informed decision.
Conduct open, frank discussions between companies and agencies following a dossier review.
Create a new category for compounds or applications that employ new design paradigms. This could create a new pathway to allow additional facilitated discussion and an opportunity for continuous, flexible, and broader dialogue. Improve submission quality. After 10 years’ experience with the ICH CTD, some companies are still being criticized for the quality of their submissions. Continually develop the current EMEA benefit–risk template: focus on critical issues, determine value, make it a model for other jurisdictions, acknowledge that output may be qualitative or at best semiquantitative, and have a goal of standardization. Be aware that multiple filters on regulatory advice could create different interpretations.
Utilize feedback mechanisms (from internal reporting, scorecards, etc.) to detect procedural flaws, communicate internally between different units, and bring about change. Peer reviews, quality management audits, and benchmarking are key feedback mechanisms. Change the status of the Institute’s Scorecard Project from retrospective to prospective. The next phase of study should include an appropriately large data set consisting of multiple companies (Institute membership companies) and multiple dossiers across therapeutic areas. It also must extend beyond the current participants to include emerging regulatory agencies. Unsuccessful dossiers should also be included in the study. Redesign scorecards to be more straightforward and easier to use. A realtime, online evaluation with easy-to-use drop-down menus is one option to consider for scorecard assessment; data would immediately be uploaded to a central repository. Then the data could be useful in performance management.
CTD, common technical document; EMEA, European Medicines Agency; ICH, International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use.
input from all stakeholders, including health technology assessors. This latter group is almost always sequenced in the decision-making process after regulatory approval, with the usual result that data collected during the development program are insufficient to permit a valid health-economic analysis, resulting in the product’s omission from national or private insurer formularies. Table 1 provides prospective and retrospective measures identified by workshop participants that can be taken to improve the predictability of a pharmaceutical development program and its subsequent review. Regulatory divergence. Even when the goal of simultaneous global submissions is achieved, regulatory agencies may render differing decisions. The most obvious and extreme such divergence in these decisions is between approval and nonapproval. Other examples of disparity include receiving a broad vs. a narrow indication, differences in the types and numbers of claims in labeling, and differences in postapproval commitments. There may also be divergences in the requirement for risk-management plans, the types of risk-minimization tools required, and the detail of safety information and its prominence in the labeling. In deconstructing the documented divergences, it is important to understand whether they were based on real differences in a benefit–risk threshold
or on evidentiary standards. The reasons for divergent evidentiary standards fall into several categories. The first are the well-accepted population differences detailed in the ICH E5 guideline7: intrinsic genetic differences in population biology that influence drug activity or safety profiles (e.g., β-receptor responsiveness, HER-2 activity) and extrinsic differences in health-care environments such as the infrastructure to deliver quality health care. Additional causes for divergent outcomes are the influences of cultural or political approaches to health care; for example, when two regulatory groups, despite exhaustive mutual consultation, continue to maintain differing opinions as to whether a development program should employ placebo or active controls. The workshop participants developed recommendations for methods to avoid divergent outcomes (Table 2). Strategies for improving the successful development of medicines
Enhanced clinical trial design. As many as 40% of compounds in phase III development ultimately fail to achieve clinical trial objectives,8 and these late-stage failures contribute to the steady rise in research and development costs, which are predicted to exceed US$80 billion by 2011 (ref. 2). In the learn-and-confirm paradigm recently adopted by some sponsors, confidence in and knowledge
about a compound’s mechanism of action, safety, and differentiation are captured much earlier in the clinical development process. In this model, smaller, leaner phase II studies address early proof of concept; later phase II trials confirm activity, characterize dose response, and contribute to an understanding of pharmacokinetics/pharmacodynamics; and phase III trials are streamlined and focused, with a low rate of failure. Modeling, through the analysis of existing data, is also used to quantify therapeutic activity, predict and characterize safety, and provide the basis for differentiation, potentially accelerating the time to dossier submission while maintaining an appropriate benefit–risk balance throughout the development cycle. Biomarkers and personalized, expedited medicines. Biomarkers have become the basis of critical metrics to demonstrate a compound’s efficacy and/ or safety as well as to identify the patient populations most likely to respond or to be at the lowest risk for safety issues. They have assumed center stage in many clinical trial designs as advances in imaging and other biometric and biochemical technologies have improved the type and quality of measurable characteristics. In Europe, biomarkers have had a major impact on the development of new oncology therapies, with 27% of cancer treatments approved between 2000 and 2008 being indicated for patients with specific
Clinical pharmacology & Therapeutics | VOLUME 87 NUMBER 1 | JANUARY 2010
29
perspec tives Table 2 Recommendations for mitigating the risk of divergent regulatory outcomes Recommendation
Methods
1. Develop effective and efficient processes by which regulatory authorities strive to harmonize views on the adequacy of a sponsor’s development plan and provide feedback; to obtain timely marketing authorization for the indication being sought; and also to support a timely and favorable health technology assessment.
Survey regulatory authorities and pharmaceutical companies on their experience and lessons learned regarding the different methods of obtaining regulatory guidance from more than one regulatory authority: through joint, serial, or parallel advice. Evaluate the joint scientific advice process for lessons learned; the results should provide a foundation of best practices, especially in simplification of logistics and faster process.
2. Establish agreed–upon frameworks for benefit–risk assessments to improve the underlying science supporting benefit–risk decisions; improve the process, reliability, predictability, and quality of these decisions; and create greater alignment and clearer communication among stakeholders.
Develop a survey for regulatory authorities and industry to gather specific data on which factors most influence the ultimate benefit–risk evaluation. Include the following fields to assess multiple dimensions of benefit–risk assessments: evidentiary standards, societal values, decision-making processes, comfort with uncertainty (in risk or in benefit), frameworks for benefit–risk assessment.
3. Develop a global toolbox for risk management plans. Tools actually selected and used to mitigate risk may be highly dependent on the health system, societal values, and other factors in the region or country of interest.
Determine which tools can be used most effectively in each region/country. Perform a survey among regulators and industry to evaluate best practices for use of tools in different regions/countries.
Perform a survey of regulatory authorities and industry to obtain data on 4. Evaluate the impact of local requirements on regulatory approval. Assess the degree to which, if at all, local requirements such as bridging studies have impact of specific local requirements, exploring societal values, scientific validity, and other factors. had an impact on approval in different regions/countries. 5. Develop and harmonize guidelines for evaluation of new therapies for specific diseases.
genetic biomarkers.9 In the United States, the US Food and Drug Administration website hosts a list of more than 100 examples of products whose labeling is associated with use in populations with 28 different genomic biomarkers.10 Although post hoc analyses have been part of some regulatory reviews of medicines with specific activities or contraindications associated with pharmacogenetic variations, ideally the efficacy and safety analyses should be performed on treatment results from patients who have been prospectively stratified into groups with positive or negative genomic findings, where all patients are included in the analysis of its results. Pharmaceutical research is gradually shifting to form the basis of personalized medicine, in which smaller, focused efforts to validate specific targets and patient selection on the basis of these characteristics could increase overall efficacy rates or reduce adverse events. This targeted approach to development provides the evidence base that is key to improving the probability of technical and regulatory success, thereby expediting access to therapies with optimized benefit–risk profiles. This new paradigm has already shown a positive impact on therapies developed for the treatment of differentiated hema30
Identify the specific factors that can support or limit homogeneity from one country or region to another.
tological cancers, for which the 5-year survival rate two decades ago approximated zero and today approaches 70% (ref. 11). The use of screening tools will potentially identify the pool of patients who will best benefit from a specific treatment, and this improved efficacy will probably reduce barriers to reimbursement/access and provide a more rational basis for determining pricing, and it should ultimately improve health outcomes. Improved efficacy or enhanced safety among targeted patient cohorts could also potentially result in streamlined development programs with smaller, less costly, and more timely clinical trials and expedited regulatory review, which together will reduce the barriers to new therapies. Surrogate end points. A surrogate end point is a biomarker that is intended to substitute for a clinical end point and that is expected to predict clinical benefit or harm based on epidemiological, therapeutic, pathophysiological, or other scientific evidence.12 In several priority disease areas such as oncology, surrogate end points, including imaging, relative response, time to progression, and progression-free survival, are accepted as part of the regulatory review, albeit with regional differences in their acceptance or weighted importance.
However, there are multiple potential issues of concern to regulators surrounding the use of biomarkers and surrogate end points in phase III trials; these issues relate to the variability of multifactorial disease, the heterogeneous risk levels of the treated population, the confounding effects of multiple therapies, the extrapolation of results to other medicines with the same or differing mechanisms of action, and the requirement for direct rather than surrogate measures of safety. The development of global standards for approving and analyzing biomarkers, imaging protocols, and surrogates remains a critical need. Conclusions
New approaches to global development and simultaneous submissions, enhanced clinical design, and the use of tools such as biomarkers and surrogate end points are evolving rapidly and may result in greater predictability in the pharmaceutical development process, the convergence of regulatory outcomes, and improved targeted therapies with better benefit– risk profiles. Workshops such as the ones offered by the Institute provide an important neutral forum in which representatives of industry, regulatory authorities, and academia can meet to discuss issues and challenges surrounding these
VOLUME 87 NUMBER 1 | JANUARY 2010 | www.nature.com/cpt
perspec tives developments such as the establishment of standardized benefit–risk models, the validation of tools and measurements, and the mitigation of cultural differences in the development and review process. This work must continue to be addressed by all stakeholders to assist companies, regulators, health-care providers, and payers in their efforts to deliver optimal, expedited health outcomes to all patients. SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature. com/cpt ACKNOWLEDGMENTs The authors acknowledge Patricia Connelly for writing and editorial assistance provided in the development of this manuscript. The views expressed in this article are the personal views of the authors and may not be understood or quoted as being made on behalf of or reflecting the position of the EMEA or one of its committees or working parties. CONFLICT OF INTEREST The authors declared no conflict of interest.
© 2010 ASCPT
1. Eichler, H.G., Pignatti, F., Flamion, B., Leufkens, H., Breckenridge, A. Balancing early market access to new drugs with the need for benefit/ risk data: a mounting dilemma. Nat. Rev. Drug Discov. 10, 818–826 (2008). 2. Centre for Medicines Research. Centre for Medicines Research (CMR) International Fact Book (CMR International, London, 2008). 3. Mallia-Milanes, A., McAuslane, N., Salek, S. & Walker, S. Can standardised scorecards evaluate the quality of industry submissions and their regulatory assessment? [abstract]. Drug Inf. J. 41, 426 (2007). 4. Committee for Medicinal Products for Human Use (CHMP). Report of the CHMP Working Group on Benefit-Risk Assessment Models and Methods (19 January 2007). 5. Walker, S., McAuslane, N., Liberti, L. & Salek S. Measuring benefit balancing risk: strategies for the benefit-risk assessment of new medicines in a risk averse environment. Clin. Pharmacol. Ther. 85, 241–246 (2009). 6. Mussen, F., Salek, S. & Walker, S.R., eds. Benefit– Risk Appraisal of New Medicines: A Systematic Approach to Decision Making (John Wiley, Surrey, UK, 2009). 7. International Conference on the Harmonisation of Technical Requirements for
8.
9.
10.
11.
12.
Clinical pharmacology & Therapeutics | VOLUME 87 NUMBER 1 | JANUARY 2010
Registration of Pharmaceuticals for Foreign Use. ICH Harmonised Tripartite Guideline: Ethnic Factors in the Acceptability of Foreign Clinical Data E5 (R1) . Beltangady, M. Phase III failures: what can be done? Appl. Clin. Trials Online (1 September 2005). Committee for Human Medicinal Products. Reflection Paper on Pharmacogenomics in Oncology (2008). US Food and Drug Administration. Table of Valid Genomic Biomarkers in the context of approved drug labels . Ries L.A.G. et al. (eds). Surveillance Epidemiology and End Results: SEER Cancer Statistics Review, 1975–2002 (2005, based on November 2004 SEER data submission). Biomarker Definitions Working Group. Biomarkers and surrogate endpoints: preferred definitions and conceptual framework. Clin. Pharmacol. Ther. 69, 89–95 (2001).
31
