SEARCH

SEARCH BY CITATION

OVERVIEW OF SPECIAL ISSUE

  1. Top of page
  2. OVERVIEW OF SPECIAL ISSUE
  3. REGULATORY PERSPECTIVE
  4. DRUG INFORMATION ASSOCIATION (DIA) BAYESIAN SCIENTIFIC WORKING GROUP
  5. ADDITIONAL PAPERS
  6. CONCLUSION
  7. REFERENCES

We are pleased to introduce this special issue exploring the use of Bayesian methods in medical product development and regulatory review. The papers included in this issue are not exhaustive in terms of the areas for which Bayesian methods are used; rather, the issue offers a variety of applications spanning the different phases of product development, highlighting the advantages and potential drawbacks of Bayesian approaches in each application.

The issue showcases four papers from the Drug Information Association (DIA) Bayesian Scientific Working Group (BSWG) and four additional papers selected from the pool of journal submissions thought to be relevant to the special issue topic and of general interest to the readership. This special issue began following a meeting of the DIA BSWG, at the suggestion of Lisa LaVange and via conversations between Lisa and Karen Lynn Price. Upon approval by the editorial board, the special issue began to take shape. We endeavored to include a variety of applications and examples, hoping to illustrate the potential utility of Bayesian methods throughout the product development life-cycle.

REGULATORY PERSPECTIVE

  1. Top of page
  2. OVERVIEW OF SPECIAL ISSUE
  3. REGULATORY PERSPECTIVE
  4. DRUG INFORMATION ASSOCIATION (DIA) BAYESIAN SCIENTIFIC WORKING GROUP
  5. ADDITIONAL PAPERS
  6. CONCLUSION
  7. REFERENCES

While Bayesian approaches to study design and analysis have been endorsed by the Center for Devices and Radiological Health (CDRH) at the US Food and Drug Administration, [1, 2] that has not been the case in FDA's Center for Drug Evaluation and Research (CDER). Bayesian approaches are generally well-accepted and even promoted by regulatory agencies for use in early and exploratory phases of drug development but less so in the confirmatory trial setting. Others external to FDA have called for their use in evaluating drug safety, both pre- and post-approval. [3] This special issue includes papers describing Bayesian applications for safety trials, meta-analyses, and non-inferiority trials – all examples of areas where their utility could be further explored.

Additional information regarding the papers included in this issue is provided below.

DRUG INFORMATION ASSOCIATION (DIA) BAYESIAN SCIENTIFIC WORKING GROUP

  1. Top of page
  2. OVERVIEW OF SPECIAL ISSUE
  3. REGULATORY PERSPECTIVE
  4. DRUG INFORMATION ASSOCIATION (DIA) BAYESIAN SCIENTIFIC WORKING GROUP
  5. ADDITIONAL PAPERS
  6. CONCLUSION
  7. REFERENCES

The DIA Bayesian Scientific Working Group (BSWG), chaired by Karen Price (Eli Lilly and Company) was formed in 2011. The BSWG's mission is to ensure that Bayesian methods are well-understood, accepted, and broadly utilized for design, analysis, and interpretation in order to improve patient outcomes throughout the medical product development process and to improve industrial, regulatory, public health and economic decision making.

The DIA BSWG is comprised of representatives from the regulatory, academic, and industry sectors engaging in scientific discussion/collaboration to:

  • facilitate appropriate use of the Bayesian approach

  • contribute to progress of Bayesian methodology throughout medical product development

The BSWG intends to do this by:

  • Creating a scientific forum for the discussion and development of innovative methods and tools.

  • Providing education on, and promoting the dissemination of, methods and best practices for Bayesian methods.

  • Engaging in dialogue with industry leaders, the scientific community, and regulators.

  • And to foster diversity in membership and leadership.

Bayesian methods provide a framework for leveraging prior information and data from diverse sources to determine probabilities relevant to inferences about issues arising at all stages of medical product development, including realistic quantification of the benefits and risks essential to health economic evaluation. The Bayesian approach provides important information for decision-making and has the potential to play an important role throughout the drug/device development process lifecycle.

Bringing together representatives from the academic, industrial, and regulatory spheres of activity is essential for overcoming hurdles and encouraging regulatory acceptance of the use of Bayesian methods. Although computational challenges remain, software to carry out the necessary calculations is widely available and steadily increasing in capability. A coordinated effort from the spheres of activity is needed to overcome a general lack of education and understanding regarding Bayesian approaches in general and to provide general guidance on many topics related to the effective implementation of Bayesian methods in medical product development.

The guidance document provided for devices submitted through CDRH [1] is a pioneering effort in this direction. Bayesian methods have recently been applied in a variety of ways, such as adaptive trial design, and have been used to provide an objective, data-driven basis for allocating resources among competing products and for guiding when to stop development of a product unlikely to be clinically or commercially successful.

Initially, the working group evaluated important challenges faced throughout the medical product development lifecycle in order to identify areas where Bayesian methods provide particular promise and are in greatest need. There were several areas that were prioritized, including (but not limited to) the topics discussed in this journal issue: educational gaps/needs, meta-analysis and network meta-analysis, safety evaluation, and formal borrowing of historical information. The papers in this issue on behalf of the DIA BSWG are some of the initial deliverables representative of the type of work the group intends to produce in coming years.

The education/survey paper summarizes results of a recent survey completed by the DIA BSWG Education sub-team, chaired by Fanni Natanegara (Eli Lilly and Company) and Beat Neuenschwander (Novartis). It also includes recommendations regarding how to improve the progress on Bayesian applications in medical product development.

The safety sub-team of the DIA BSWG, which is chaired by Amy Xia (Amgen) and Karen Price (Eli Lilly and Company), provided two papers in this issue. The network meta-analysis paper was authored by the safety meta-analysis sub-team of the BSWG, which is chaired by David Ohlssen (Novartis). The primary editor was Professor Bradley Carlin (University of Minnesota). The paper explores the use of Bayesian methods when applied to drug safety meta-analysis and network meta-analysis. Guidance is presented on the conduct and reporting of such analyses, and the work is illustrated through a case-study. While this paper focuses on a safety example, aspects discussed in the paper are readily applicable to efficacy outcomes. The safety trials paper was written by the safety trials sub-team, which is chaired by Karen Price (Eli Lilly and Company), and evaluates challenges associated with current methods for designing and analyzing safety trials and provides an overview of several suggested Bayesian opportunities which may increase efficiency of safety trials along with relevant case examples.

The fourth paper summarizes initial work of the priors sub-team of the DIA BSWG, which is chaired by Nelson Kinnersley (Novartis) and Scott Berry (Berry Consultants). The lead author for this paper is Kert Viele (Berry Consultants). The paper reviews several methods for historical borrowing, illustrating how key parameters in each method affect borrowing behavior, and then compares the methods. The group emphasizes “dynamic” versus “static” borrowing and the decision process involved in determining whether or not to include historical borrowing. The primary goal is to provide a clear review of the key issues involved in historical borrowing and a comparison of several methods useful for practitioners.

ADDITIONAL PAPERS

  1. Top of page
  2. OVERVIEW OF SPECIAL ISSUE
  3. REGULATORY PERSPECTIVE
  4. DRUG INFORMATION ASSOCIATION (DIA) BAYESIAN SCIENTIFIC WORKING GROUP
  5. ADDITIONAL PAPERS
  6. CONCLUSION
  7. REFERENCES

The four additional papers selected from the pool of journal submissions expand the range of applications considered in this special issue.

Gamalo, et al., provide methodological approaches to determine non-inferiority margins that can utilize all relevant historical data through a novel power adjusted Bayesian meta-analysis with Dirichlet process priors. They provide a Bayesian decision rule for the non-inferiority analysis and illustrate via case examples.

Gsponer, et al., present a practical guide to Bayesian group sequential designs where multiple criteria based on the posterior distribution can be defined to reflect clinically meaningful decision criteria on whether or not to stop or continue the trial at the interim analysis. Practical implementation of the approach is illustrated via case examples from different phases, endpoints, and informative priors.

Cruz's paper, “Bayesian analysis for nonlinear mixed-effects models under heavy-tailed distributions”, considers an extension of nonlinear mixed-effects models in which random effects and within-subject errors are assumed to be distributed according to a rich class of parametric models often used for robust inference. The methods are demonstrated via a dataset from a pharmacokinetic study.

Stamey, et al., propose a Bayesian approach to evaluate the impact of unmeasured confounding in cost-effectiveness studies. Simulation studies were conducted to determine the impact of ignoring the unmeasured confounder and to determine the size of the validation data required to obtain valid inferences.

CONCLUSION

  1. Top of page
  2. OVERVIEW OF SPECIAL ISSUE
  3. REGULATORY PERSPECTIVE
  4. DRUG INFORMATION ASSOCIATION (DIA) BAYESIAN SCIENTIFIC WORKING GROUP
  5. ADDITIONAL PAPERS
  6. CONCLUSION
  7. REFERENCES

With these papers it is clear to see the breadth of areas Bayesian methods are being applied, demonstrating the fact that Bayesian methods have an impact throughout the medical product development lifecycle. As the amount and complexity of available data increases dramatically (“big data”), researchers in pharmaceutical development must utilize more sophisticated statistical tools, including Bayesian statistical methods. We hope that the readers of this special issue will benefit from this set of papers and that the papers will allow appropriate implementation of Bayesian methods in a way that will add value to the medical product development process. The DIA BSWG hopes that this special issue is a step in the right direction in order to achieve our mission of ensuring that Bayesian methods are well-understood, accepted, and broadly utilized for design, analysis, and interpretation in order to improve patient outcomes throughout the medical product development process and to improve industrial, regulatory, public health and economic decision making. This has been a rewarding experience for those involved, and we thank all of the authors for their commitment to this issue. We thank DIA for their support of the BSWG. Finally, we thank the many anonymous referees who provided insightful comments and suggestions which greatly improved the quality of the papers, and the journal for its support of this issue.

REFERENCES

  1. Top of page
  2. OVERVIEW OF SPECIAL ISSUE
  3. REGULATORY PERSPECTIVE
  4. DRUG INFORMATION ASSOCIATION (DIA) BAYESIAN SCIENTIFIC WORKING GROUP
  5. ADDITIONAL PAPERS
  6. CONCLUSION
  7. REFERENCES
  • 1
    United States Food and Drug Administration. Use of Bayesian Statistics in Medical Devices Clinical Trials – Guidance for Industry and FDA Staff. Center for Devices and Radiological Health, Food and Drug Administration, February 5, 2010. Available at: http://www.fda.gov/downloads/MedicalDevices/DeviceRegulationandGuidance/GuidanceDocuments/ucm071121.pdf [Accessed on July 17, 2013].
  • 2
    Campbell G. Bayesian statistics in medical devices: Innovation sparked by the FDA. Journal of Biopharmaceutical Statistics 2011; 21(5):871887.
  • 3
    Institute of Medicine of the National Academies. Ethical and Scientific Issues in Studying the Safety of Approved Drugs. National Academies Press: Washington, DC, 2012.