Active Control Studies

Download Report

Transcript Active Control Studies

Equivalence, Similarity, and Non-inferiority
Clinical Trials in Neurotherapeutics
ASENT 10th Annual Meeting
March 2008
Marc K. Walton, M.D., Ph.D.
Senior Medical Policy Advisor
Office of Policy
Office of the Commissioner, FDA
The views expressed are those of the author, and do not represent an official
FDA position
Equivalence / Similarity / Noninferiority
 Distinctions of study circumstances and study goal
 Need to be clearly identified
 Two treatments
 Both known efficacious
 Want to know comparative efficacy
 One treatment with unknown efficacy
 Second treatment known efficacious
 Want to prove first is also efficacious
Comparative Efficacy
 Can the two treatments be shown Equivalent?
 Are two agents are essentially the same?
 Implies same amount of efficacy for each treatment to
within some level of disinterest
 Neither drug is better than the other
Two sided interest
 Not a regulatory requirement for marketing
 Not common regulatory interest (ie, sponsor interest,
possibly with desire for regulatory affirmation)
 “The Same”?
 With respect to characteristics planned for rigorous
evaluation
Proof of Efficacy for an Unproven Treatment
 Is the new treatment efficacious?
 In the same manner as the established treatment
 Often a goal for regulatory decision
 Based on evidence that new drug has some efficacy
 No explicit requirement for same efficacy of other
drugs, or any predefined fraction of efficacy of other
drugs for same disorder.
 One sided comparison of interest
 Precise comparative efficacy is not a goal
 Non-inferiority: A misnomer
 Goal to is to show “sufficient efficacy”
 Accomplished by showing “not unacceptably inferior”
to control
 Does not imply truly not inferior
Similarity
 Perhaps not equivalent, but close enough
 ‘Close enough’?
 In the eye of beholder
 May imply a two sided interest as well
 Individual judge dependent
 Difficult to define/describe
 Principles of Equivalence / Noninferiority studies apply
 More laxity in interpretation granted on an individual
person basis
Equivalency & Noninferiority: Active Control Studies
 Why / When do Active Control studies?
 Comparison of effects of two agents
May not need to exclude Placebo


Issues of assay sensitivity (later) and validity of
interpretation much easier if placebo included
Assessment of effects of one agent when placebo
control not permissible
Earlier lecture touched on circumstances
Ethics of withholding a known effective treatment
with life-saving or irreversible life-altering effect
Can be dose-ranging study with single agent


Ethical need to avoid non-effective doses
May not be a practicable approach
 Remainder of talk assumes active control study of two drugs
Active Control Studies
 Study aspects enabling data to be validly interpreted
 Validity of Design
 Integrity of Conduct
Adherence to protocol
 Same overall issues with standard placebo-control
studies
 Additional complexities in Active Control studies
Increased Complexities
 Assay sensitivity of the design
 Interpretation of the quantitative result
 Determination of margin of acceptable difference
 Margin
 Built on combination of
Historical knowledge of comparator’s efficacy
Clinical judgment of what is an ‘acceptable’
difference
Assay Sensitivity
 Can study detect a difference if one exists?
 Unrecognized failure of assay sensitivity leads to:
 Type II error for superiority study
 Type I error for non-inferiority study
 Factors which affect sensitivity can seem to impair or
advance study organizer’s goals oppositely
 Promotes ability to distinguish between
 Evidence of Absence
 Absence of Evidence
 Laboratory assays often include positive and negative
controls; rarely can be done in clinical trials
Interpretation of Study Results
 Results analyzed as comparison of two groups
 Need numeric criterion to form interpretation
 Placebo control & other superiority studies have same
requirement, but easy to define
 Show between-group difference > 0
 New drug superior to placebo / other control
 Superiority of new treatment over old treatment difficult
to achieve with active agents
 Thus non-inferiority approach attractive
 Need to have quantitative comparison criterion to
interpret study result, allowing for potentially the same
efficacy
Margin of Acceptability
 How much less efficacious can the new drug be than old
drug, and still deem it acceptable to use (or equivalent)?
 Two components to consider
 First – what is the efficacy the old drug provides?
Statistical analysis of existing information
Can not allow new drug to be worse than old drug
by that amount, or it provides no efficacy
 Second – how much of that quantitative amount of
efficacy is it permissible to give up?
Clinical judgment
Usually easier to assess after know first
component
Express as an absolute amount or relative amount
Margin of Acceptability
 Two components – M1, M2
 Separate basis, separate purpose
 Margin of acceptable inferiority cannot be larger than:
 Statistical component – M1
 Clinical component – M2
 Overall margin (M)
 If M2 << M1: M = M2
 M2 > M1: Either no need for AC study (use placebo
control study) or non-inferiority not feasible
 M2 <(modest) M1 leads to complexities and anxieties
Active Control Agent Existing Knowledge
 Solid knowledge of efficacy of active control agent is
essential
 From prior studies – historical placebo control studies
 Preferably from multiple different prior studies
 Historical studies each conducted in a specific manner
 Regimen of use, population of use
 Concomitant care, alternative therapies
 Existing knowledge applicable in circumstances of
prior experience
May not be reliable in other circumstances
 Quality of historical studies needs to be considered
 Adherence to protocol
 Quality of data collection for outcome now of interest
(may not have been primary outcome in study design)
M1 – Statistical Margin Component
 What is the treatment effect of the comparator agent?
 Typical?
 Reasonably likely present?
 Highly likely present?
 There will be no ability to actually confirm in new AC
trial
 Generally derived from some form of meta-analysis
 With allowance for uncertainty in historical
quantitative estimates (i.e., variance of each effect
estimate; variance of meta-estimate)
 The more precisely treatment effect was estimated in
historical trials and the more trials there were, the
more precise the M1 value may be determined
M1 – Estimate from Historical Evidence
 Objectively done meta-analysis is important
 What studies to include
 What portions to include (e.g., patient subsets)
 Selection to improve strength of meta-analysis and
applicability to planned new study use of drug
 How relevant is historical M1 estimate to a new study:
 Relates to issues outside of purely mathematical
variance – Not in meta-analysis
 Population unchanged?
Planned eligibility criteria
Unplanned shift in available population related to
changes in medical practice (e.g., development of
alternative treatments) in intervening time


Active Control’s treatment regimen unchanged?
Concomitant care impact changed?
Uncertainties Affecting Interpretation of New Study
 Imprecision of control agent’s effect in each of the historical
studies
Partially estimated by statistical variance of each study
 Variability of treatment effect size between historical studies
Partially estimated by meta-analysis methods
 Uncertainty of quantitative extrapolation of historical treatment
effect to treatment effect of control agent in new study
The Constancy Assumption – relevance of estimate
Not statistically calculable
Can create an (arbitrary) adjustment inserted into
mathematical procedures
 Uncertainty of comparison between control and new agent in
the new study
Estimated with statistical variance in new study
AS ALWAYS, Statistical approach adequate only if
non-statistical sources of bias ignorable
Circumstances for Confidence in M1 Estimate
 Good quality of prior placebo control studies – A&WC
 Multiple prior placebo controlled trials with comparator
 Good grounds for combining data in a meta-analysis (all
randomized, combinable doses/regimen; outcome
measure defined, assessed the same way)
 Consistent results across all the prior studies
 Prior studies done over extended period of time, but
completed not long ago
 Allows for some changes in populations, concomitant
care, etc, showing this is not critical
 Prior studies done with some variations in design
 Implies drug treatment effect not highly sensitive to
design
 Better with large treatment-associated efficacy effect
Historical Knowledge for Non-inferiority Purposes
 May be source of difficulty
 Active control design being used because placebo
control cannot
 Due to nature of benefit of established drug
 May be few (one?) placebo controlled studies with a lifesaving drug before it becomes an established, necessary
standard of care
 Limited ability to form precise estimate of treatment
effect
M2 – Clinical Component
 How much efficacy is clinically acceptable to give up in
the new agent and still allow new treatment to be
medically acceptable to use in place of an already
proven active agent
 Clinical judgment, given importance of endpoint, nature
of disorder, amount of efficacy of the active control agent
 Absolute (% of patients, points on scale) or relative to
control’s effect (% of control’s efficacy)
 M2 does not serve to make a hazy M1-statistical
analysis more reassuring by making it more stringent;
M2 is from solely clinical meaning point of view
 ½ often used
 No automatic reason why this is the correct M2 in
different settings
 Each case should be considered independently
Non-inferiority Analysis: Conceptual Methods
 Sequential CI method
 Dual (Double) CI method
 Fixed (Pre-determined) Margin method
 Synthesis method
 Putative Placebo method
 Strengths and Weaknesses to each
 How confident can ‘we’ be in effect estimates from
historical data?
 How many AC studies will we have to consider?
 Given nature of benefit, how much risk of being wrong
in favor of new drug is acceptable?
 How much risk of being wrong by discrediting new
drug is acceptable?
Conceptual Methods
 Sequential CI method
M1 estimate of active comparator’s effect from metaanalysis confidence interval
 Apply M2 limit to form overall margin (M)
 New agent’s comparative effect estimated with confidence
interval from new study
 Show C.I. does not exceed overall margin M

 Synthesis method

Combines historical placebo-comparator data and new
study comparator-test agent data into single calculation.
 Result is numeric value indicating putative placebocontrolled efficacy estimate of test agent.
 Chief difference is conceptual approach of how to address
inter-study variability and strength of constancy
Quality of Study Data & Analysis
 Constancy assumption needs good study quality
 Effects of study flaws
 Anti-conservative; opposite of superiority study
 Flaws
 Adherence to study protocol
 Cross-overs
 Completeness of data
Dropouts, missing data
 Analysis population
 ITT keeps errors in but is statistically pure
 Per Protocol keeps errors out, but impure
 Probably best to do both and show that all ways of
looking at dataset are consistent
Interpretation of Results
 Examples of CIs and assessment
 Apparent paradox of non-inferiority success with an inferior drug
a
b
c
d
e
f
Closing Comments
 Size of non-inferiority studies
 Often large to achieve sufficient precision in estimate
of New-Active effect difference so that CI of
comparison falls within the limited range defined by
margin
 When M is small, study is large; when M is large,
study can be moderate in size
 Biggest factor can be efficacy of active comparator
 Equivalence studies
 If this implies greater closeness of effect than “notunacceptably-inferior”, then M is smaller based on M2
component
 Consequently sample size of study will increase
Closing Comments
 Active Control-Noninferiority studies of successive new
agents ( risk of ‘bio-creep’)
 M based on quantitative data applicable to the
comparator being used. Risk of using New1, based
on single AC study vs. Std, as the AC agent for New2
assessment, and so on
 New1 might be inferior to original comparator by a
small amount (in itself negligible)
 Negligible inferiority added twice (Std-New1; New1New2) may be not negligible ; frequently only 1-2
study with New1 so that meta-analysis of
New1StdPbo gives larger CI, and less assured
efficacy
 If ignored, New3 may have little to no efficacy yet
pass the erroneously applied non-inferiority margin
Closing Comments
 Planning and conducting a non-inferiority study does not
prohibit achieving a conclusion of superiority if such is
the case. Analytic plan of study can allow for this when
done in the proper manner.
 Safety issues also always need considering

Often not as enticing as working on efficacy aspect
 Reading and Reporting of Active Control Studies
 Planning and analysis aspects unique to AC studies
 Great difficulty in assessing study result of these
aspects not well described in publication