Transcript template
Preparing for a career
in regulatory biostatistics What does it take ?
Robert T. O’Neill Ph.D.
Director, Office of Biostatistics
Office of Translational Sciences, CDER
For presentation at the 17th annual International Chinese Statistical
Association Applied Statistics Symposium ; June 4-7th, 2008
Outline of Issues
What is regulatory statistics
Where do you learn
Opportunities
Training the next generation
The science of regulatory
statistics
The field of regulatory statistics evolved from the need to
apply statistical principles and practices to implement
regulations developed to promote and protect the public
health by facilitating the development of effective and safe
medical products. The modern era began about 1970
Many statisticians have begun careers in FDA and moved to
industry or academic positions where they built upon their
regulatory statistics background
There are about 105 biostatisticians in CDER, and about that
many more in all of FDA, established programs in CBER,
CDRH, CFSAN and CVM and NCTR
http://www.fda.gov/cder/Offices/Biostatistics/
What is the role of
biostatistician in a regulatory
agency and how did it evolve
The need for biostatisticians was created by the
regulations and standards for efficacy and safety
What shaped this role
Culmination of accumulating experience
The development and evolution of the discipline
of regulatory statistics
International roles, influences and events
Major Events Impacting
Determination of Evidence
The 1962 Kefauver-Harris Amendments: the foundation for
experimental evidence as the basis for drug approvals
The 1970 definition of ‘Adequate and well controlled
investigations’: the foundation for statistical principles: the
concept of hypothesis testing and estimation, randomization,
blinding
The 1986 NDA Rewrite: the foundation for documentation
of evidence, including statistical evidence and introduction
of the integrated efficacy and safety section -
Major Events Impacting
Determination of Evidence (cont.)
The 1988 Guideline for the Format and Content of the
Clinical and Statistical Sections of an application
1992; Subpart H - Accelerated Approval of New Drugs for
Serious or Life-threatening Illnesses - surrogate endpoints
(AIDS crisis)
The 1997 Food and Drug Modernization Act (FDAMA); a
modification of the substantial evidence criteria
The 1998 ICH Statistical Principles for Clinical Trials: the
foundation for global understanding , harmonization and
implementation of statistical principles
Roles in the career of a
regulatory biostatistician
Statistical review team member
Entry level and senior reviewer
Expert statistician
Applied researcher
Leadership, manager , policy development
Office Director, Deputy
Division Director, Deputy
Associate Director
Other organizational roles
Bioinformatics, training, compliance, epidemiology
What does a regulatory
statistician do ?
Evaluate , critique large numbers of clinical studies, with access to
patient level data - data base management skills
Make recommendations inference and evidence
Prepare and deliver public advisory committee presentations that are
video taped, webcast
Write reports that may be available publically through FOI
Develop and Negotiate statistical and clinical guidances
Domestic and international (ICH)
Dispute resolutions
Administrative hearings (rarely)
Influence, educate colleagues
Interactions with multiples audiences, including industry statisticians,
consultants, and academics
An advisory committee
biostatistician - a special
government employee (SGE)
Usually an academic with minimum conflicts of
interest
Difficult job
Requires unique skill mix
Goes beyond understanding the science and the
statistics
Multi-disciplinary committee
Voting is often the decision making choice
Training and Experience as a
critical part of the career path in
regulatory biostatistics
A reviewer of clinical and pre-clinical data within the
context of scientific and regulatory standards of evidence
A policy maker
A decision maker
A negotiator
An educator
A speaker
A writer
Subject matter expertise
Impact of regulations and health care on mission of FDA
Drug development
Pre-Clinical and clinical study design and analysis methodology
Exploration / Confirmation
New proposals: adaptive, enrichment
Surveillance and life cycle risk assessment
Epidemiology, observational study methods, causal inference,
propensity score methods, multiple events
Data mining strategies
Meta-analytic methods - individual and study level covariates - not
traditional literature based meta-analysis
Data base management, programming, scientific computation
Modern process control and quality by design methods for manufacturing- non
invasive testing, development of standards and setting specifications (wastage
and safety)
Statistical areas
Clinical trials – all aspects (read ICH E9)
Experimental designs for clinical trials
Repeated measures,Time to event, K period
designs, Adaptive methods
Methods development, application
All areas dealing with FDA guidances
Pre-clinical animal study designs
Chemistry , manufacturing, contols, specification
setting and monitoring
Simulation practices - modern protocol planning and
scenario planning - not of a just a single study but a
series of studies or a development program
Statistical areas
Epidemiology
Observational data methods Cohort and case control studies
Large data base and outcomes based study methods
Meta-analytic methods or combining
Multiplicity - outcomes, subgroups - very important for inferential claims
Prediction, prognosis, differential benefit and harm - important for
understanding the difference between individual and group prediction and
‘personalized medicine’
Sampling, surveys
Exploratory – bayesian , frequentist and likelihood strategies
Some Training that FDA
provides to our staff
Statistical courses/ seminars/ workshop
Subject matter courses/ seminars/
Genomics, nanotechnology, drug
safety evaluation,
Speaking, enunciation, toastmasters,
negotiating
Designing Quality into Clinical Trials
Biostatistics and Statistics
Broadening the field of
application
Quality by design - modern quality control - in process control and
specification setting and monitoring - designing quality into the design
space
The manufacturing process - heparin
The clinical trial
Sampling - clinical trial auditing and inspections - consumer surveys and
OTC label comprehension studies
Micro-array , SNP, genomics marker identification and validation studies
Study design and analysis to support choice of and validation of patient
reported outcomes in clinical trials
Scientific computing, large scale data base analysis and data mining, record
linkage studies, health records analysis
FDA Statisticians Collaborate
with and are professionally
involved in external activities
Professional society involvement
Organizing meetings externally - with
Phrma, ASA, DIA, SCT, ISCA, IBS
Training the next generation
Education and training vs
experience in the career path
Experience , case studies, breadth of involvement
Interest in continual learning vs. comfort zone
Academic contributions need to be aligned with
needs of society and reality of modern health care
FDA providing case study material to the
academic sector
Fellowships
We are looking
for a few good
statisticians