1 Improving FDA/Industry Improving FDA/Industry Interactions: Interactions: Suggestions from Suggestions from FDA/CDER Statisticians FDA/CDER Statisticians Rafia Bhore, Ph.D. Rafia Bhore, Ph.D. Janice Derr, Ph.D. Janice Derr, Ph.D. FDA / CDER / Office of FDA / CDER / Office of Biostatistics Biostatistics The views expressed in this presentation are those of the speakers and not necessarily of the U.S. Food and Drug Administration (FDA).
22
Embed
1 Improving FDA/Industry Interactions: Suggestions from FDA/CDER Statisticians Rafia Bhore, Ph.D. Janice Derr, Ph.D. FDA / CDER / Office of Biostatistics.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Improving FDA/Industry Improving FDA/Industry Interactions: Interactions: Suggestions from Suggestions from
FDA/CDER StatisticiansFDA/CDER Statisticians
Rafia Bhore, Ph.D.Rafia Bhore, Ph.D.
Janice Derr, Ph.D.Janice Derr, Ph.D.
FDA / CDER / Office of BiostatisticsFDA / CDER / Office of Biostatistics
The views expressed in this presentation are those of the speakers and not necessarily of the U.S. Food and Drug Administration (FDA).
2
CDER Office of BiostatisticsCDER Office of Biostatistics
Follow-up (meeting minutes, further discussion)Follow-up (meeting minutes, further discussion)
6
Suggestions from CDER Suggestions from CDER StatisticiansStatisticians
(Good Meeting Practices)(Good Meeting Practices) Use a meeting with FDA as an opportunity to Use a meeting with FDA as an opportunity to
send in questions about statistical issues send in questions about statistical issues
Ask good questions that will give you useful Ask good questions that will give you useful answers answers
Provide sufficient detail to help us give useful Provide sufficient detail to help us give useful statistical review commentsstatistical review comments
Use the channels of communication to get a Use the channels of communication to get a response from FDA statisticians about response from FDA statisticians about statistical issuesstatistical issues
7
Investigational New Drug Investigational New Drug Application Application (IND) Stage(IND) Stage
Special Protocol AssessmentSpecial Protocol Assessment
Special Protocol AssessmentSpecial Protocol Assessment
Sponsors can submit certain types of Sponsors can submit certain types of protocols with specific questions prior to start protocols with specific questions prior to start of study (Guidance recommends 90 days).of study (Guidance recommends 90 days).
FDA determines if SPA process applies to the FDA determines if SPA process applies to the request, and if so, responds to questions request, and if so, responds to questions within 45 days (PDUFA goal).within 45 days (PDUFA goal).
Protocol agreements under SPA are part of Protocol agreements under SPA are part of the administrative record. Regulations the administrative record. Regulations describe the circumstances under which the describe the circumstances under which the agreements can be changed. agreements can be changed.
SPA Guidance 2002SPA Guidance 2002
9
Suggestions from CDER Suggestions from CDER StatisticiansStatisticians
(Special Protocol Assessment)(Special Protocol Assessment) Ask good questions that will give you useful Ask good questions that will give you useful
answers answers
Provide sufficient detail to help us give useful Provide sufficient detail to help us give useful statistical review commentsstatistical review comments
10
Statistical Analysis Plan (SAP)Statistical Analysis Plan (SAP) Prospective plan of statistical methods not Prospective plan of statistical methods not
detailed in the Protocoldetailed in the Protocol
Protocol details design considerations vs. Protocol details design considerations vs. SAP details analysis considerationsSAP details analysis considerations
Design: Endpoints, type of control, planned Design: Endpoints, type of control, planned comparisons, multiple testing, interim analysescomparisons, multiple testing, interim analyses
Analysis: Statistical models, handling of missing Analysis: Statistical models, handling of missing data, nature of censoring, analysis populations, data, nature of censoring, analysis populations, repeated measurements over time, study repeated measurements over time, study windows, etc.windows, etc.
11
Suggestions from CDER Suggestions from CDER StatisticiansStatisticians
(Statistical Analysis Plan)(Statistical Analysis Plan) Statistical Analysis Plan (SAP) should be Statistical Analysis Plan (SAP) should be
detailed and prospectively writtendetailed and prospectively written
Prospectively submit to FDA for Phase 3 Prospectively submit to FDA for Phase 3 studies and Phase 2 supportive studiesstudies and Phase 2 supportive studies Open-label studies submit before study Open-label studies submit before study
beginsbegins
Blinded studies submit prior to last patient Blinded studies submit prior to last patient enrolled or first interim analysis (whichever enrolled or first interim analysis (whichever comes first)comes first)
12
Suggestions from CDER Suggestions from CDER StatisticiansStatisticians
(Statistical Analysis Plan contd.)(Statistical Analysis Plan contd.) Identify critical issues at protocol Identify critical issues at protocol
design stage or at least Statistical design stage or at least Statistical Analysis Plan writingAnalysis Plan writing
Commercial Sponsors should Commercial Sponsors should encourage co-operative trialists to encourage co-operative trialists to write a Statistical Analysis Planwrite a Statistical Analysis Plan
13
New Drug Application New Drug Application (NDA) Stage(NDA) Stage
Integrated Summary of EfficacyIntegrated Summary of Efficacy
LabelingLabeling
14
Integrated Summary of Efficacy Integrated Summary of Efficacy (Suggestions from CDER (Suggestions from CDER
Statisticians)Statisticians) Important component of New Drug Application Important component of New Drug Application
ReviewReview
Provide clinically meaningful and logically tight Provide clinically meaningful and logically tight argument whether drug has necessary evidence for argument whether drug has necessary evidence for efficacy claimefficacy claim
Provide side by side comparison of studiesProvide side by side comparison of studies
NOT necessarily pooled or meta-analysis of efficacyNOT necessarily pooled or meta-analysis of efficacy
Discuss pooling study results with FDADiscuss pooling study results with FDA
15
Integrated Summary of Efficacy Integrated Summary of Efficacy Example: Example:
CCanan YoYoUU PPRRove ove EEfficacy and fficacy and SSafetyafety of curevirof curevir (CURES)(CURES)
Treatment Success Rate by Study and for Pooled Studies
41%
15%
29%
22%18% 20%
0%5%
10%15%20%25%30%35%40%45%
CURES 1 study CURES 2 study CURES
TestPlacebo
p
p-value < .0001
p-value < .0001
p-value = .380
16
Statistical Input on Labeling Statistical Input on Labeling TextText
FDA Statisticians review labeling text:FDA Statisticians review labeling text:
Statistical support for study conclusions, Statistical support for study conclusions, claims and indicationsclaims and indications
Description of study results, summary Description of study results, summary statistics and inferential languagestatistics and inferential language
Information in tables and figuresInformation in tables and figures
Labeling Labeling Example Example
#1: #1: Statistical Statistical
Input Input ProvidedProvided
CLINICAL STUDIES
…
The NAGLAZYME-treated group showed greater mean increases in the distance walked in 12 minutes (12-minute walk test, 12-MWT) and in the rate of stair climbing in a 3-minute stair climb, compared to the placebo group (Table 2).
18
Labeling Example #2: Labeling Example #2: Statistical Input NeededStatistical Input Needed
Proposed textProposed text: “The combination of : “The combination of AA and and BB is is effective in lowering LDL-C levels beyond effective in lowering LDL-C levels beyond that achieved by either agent alone.”that achieved by either agent alone.”
Statistical issueStatistical issue: The study was not designed : The study was not designed to support this conclusion. The study had to support this conclusion. The study had two arms, two arms, (A+B)(A+B) combination product, and combination product, and AA monotherapy. monotherapy.
19
Labeling Example #3: Labeling Example #3: Statistical Input NeededStatistical Input Needed
Proposed tableProposed table: The symbol “: The symbol “**” was used for ” was used for p<0.05, and “p<0.05, and “****” was used to indicate ” was used to indicate nono statistically significant difference between statistically significant difference between the active treatment arm and the placebo the active treatment arm and the placebo arm.arm.
Statistical issueStatistical issue: This is not a typical way to : This is not a typical way to depict this outcome and may be confusing to depict this outcome and may be confusing to some readers. some readers.
20
Suggestions from CDER Suggestions from CDER StatisticiansStatisticians(Labeling)(Labeling)
Provide your statistical perspective in Provide your statistical perspective in the development of labeling text.the development of labeling text.