Transcript A regulator`s view of end of Phase II decision making and

Safeguarding public health
A regulator’s view of end of Phase II
decision making and phase II studies
Dr David Wright
Senior Statistical Assessor and Scientific Advice
Coordinator
MHRA
©
Contents
• Sponsor’s decision making at End of Phase II- a view from
an interested methodologist
• Are the right Phase II studies being done?
Slide 2
©
Sponsor’s decision making at End of Phase II
• Regulators rarely see company decision making strategies
• Experience based on hundreds of development programmes
(albeit a biased subset; better programmes, better medicines)
and scientific advice requests. Also, formal and Informal
discussions with senior colleagues in clinical development
teams in pharma.
• Drug development is risky. The risk is primarily the
sponsor’s to bear (as is the profit)
• The question is whether risks can be better elucidated,
better quantified to aid expert decision making.
Slide 3
©
Decreasing efficiency of drug development
Industry R&D
Expense
($ Billions)
Annual NME
Approvals
$50
200
$45
180
$40
$35
R&D Investment
NME & Biologics
Approvals
160
140
120
$25
100
$20
80
$15
60
$10
40
$5
20
$0
0
19
8
19 0
8
19 1
8
19 2
8
19 3
8
19 4
8
19 5
8
19 6
8
19 7
8
19 8
8
19 9
9
19 0
9
19 1
9
19 2
9
19 3
9
19 4
9
19 5
9
19 6
9
19 7
9
19 8
9
20 9
0
20 0
0
20 1
0
20 2
0
20 3
0
20 4
0
20 5
0
20 6
07
$30
Source: PhRMA, FDA, Lehman Brothers; [Dr. Robert Ruffolo]
Slide 4
©
Decreasing efficiency of drug development
• Multiple possible causes …
• High Phase III failure rate, reasons include:
- Inadequate efficacy
· doesn’t beat placebo
· not competitive in marketplace
- Emerging toxicity
- Inadequate B-R
• Lots of Phase II questions – demonstrate activity, quantify
dose-response, what dose, will we beat placebo, what effect
size, what toxicity profile, how do efficacy and safety
change in broader patient population, …? Ultimately go /
no-go decision.
Slide 5
• Impossible to address in the same design
©
Sponsor’s decision making at End of Phase II
•
Hypothesis 1: Exploratory development programmes are shrinking
and are not fit for purpose. Increasing exploratory development
(‘learn’) will reduce failure rate in confirmatory trials.
•
Hypothesis 2: Information needs to be collected efficiently. Novel
clinical trial designs add to the ‘toolkit’ and will bring benefits in some
indications / development programmes. There is the possibility for
adverse effects of novel approaches in other experimental situations.
There are many improvements that could be made to the current
paradigm without being ‘novel’.
•
Two questions for Phase II:
1. With what certainty (how much data) do we wish to address each
question of interest?
2. How can we collect the data efficiently?
•
Focus of many current discussions seem to be only on hypothesis 2
Slide 6
©
Sponsor’s decision making at End of Phase II
• The tools exist to help quantify the risks being taken to best
inform expert (clinical, statistical, regulatory, health-technology
etc.) judgement
- Benda N, Branson M, Maurer W, Friede T (2009) Clinical
Scenario Evaluation: A framework for the evaluation of
competing development strategies. Drug Development 4: 8488.
• Could help to increase likelihood of success, identify early
kills and conduct development more efficiently across a
portfolio
• Could help to move from high risk, high reward to lower risk,
high reward?
Slide 7
©
Sponsor’s decision making at End of Phase II
• Also increased use of modelling and simulation
- Help quantify what is known
- Use to guide planning
- Use to replace clinical trial data
Slide 8
©
Personal reflections on these hypotheses
• “Current paradigm unsustainable – must be more efficient”
- Efficient OED – “do same for less”
- NOT “do less”
• Search for ‘efficiency’ concentrates on Phase II and Phase III. Is
reducing amount of late Phase II data ‘efficient’ or simply ‘less’? In
particular will this reduce Phase III failure rates?
• Some arguments in favour of late phase exploratory data being more
extensive than in the past …
- Increased cost of Phase III trials
- More competition (nth statin, mth triptan)
- More hurdles (e.g. cost effectiveness, risk-averse / questioning
society)
Slide 9
©
Personal reflections on these hypotheses
• Simulation work has shown that ‘optimal’ dose is rarely
identified by a standard phase II trial.
- Innovative Approaches for Designing and Analyzing
Adaptive Dose-Ranging Trials Bjorn Bornkamp et al,
Journal of Biopharmaceutical Statistics, Volume 17, Issue
6 November 2007 , pages 965 - 995
• Clinical team metrics include Phase III FPE
Slide 10
©
Are the right Phase II studies being done?
“The changes in effect of fixed dose studies of recently
approved antidepressants employing low and high doses
showed that the effect tends to be smaller at higher doses
compared with lower doses:
Demonstrating that effect size is not a good predictor of
effectiveness of higher antidepressant doses.”
Slide 11
©
What did they mean?
Probably referring to data like this:
Fixed dose trials
Change from baseline in HAM-D:
Placebo: -10.1
10mg: -8.9
20mg: -12.4
30mg: -11.5
40mg: -11.5
Slide 12
©
SSRI expert working group
Report of the CSM expert working group on the safety of
selective serotonin reuptake inhibitor antidepressants
EWG consists of experts and lay members from across the
UK
Chair: Professor Ian Weller – University College London
Group included consultant psychiatrists for both child &
adolescent and adult patients, professor’s of statistics,
epidemiology, psychiatry, patient representatives.
Expert testimony given by patient user groups
Slide 13
©
SSRI expert working group
Driven by concerns regarding increased “suicidality” in
children and adolescents taking SSRIs
Went on to become a wider review of the drug class
Including pharmacovigilance, clinical pharmocology, children
and adolescents, suicidal behaviour, withdrawal reactions,
dose response – Chapter 9
Group started May 2003 – report issued December 2004
Slide 14
©
Dose finding
There was a concern, in the light of the potential AEs that were being
investigated, whether the dosing was appropriate
If an AE profile is “clean” maybe there is less concern about precise
dosing levels
Usage databases suggested that many patients were being started on
higher than the recommended dose
Licence holders were asked to provide the data which supported dosing
statements for fresh review by the EWG
Slide 15
©
Typical SSRI licence
Most SSRI dosing recommendations of the form:
Recommended dose x mg
Patients responding adequately to x mg may increase their dose in
increments of y mg to a maximum of z mg
Starting doses established as efficacious
What is the basis for the recommendation to try higher doses if the lower
doses fail?
Slide 16
©
What did we typically see?
Supporting data was usually a mixture of:
Fixed dose studies
Flexible dose studies
Slide 17
©
Fixed dose studies
Patients randomised to one of a range of doses (or placebo)
Problem:
In many early trials patients were not titrated to higher doses, but started
on them straight away
This was corrected in later years
Some companies adjusted by including only patients who lasted for a
certain duration on treatment
Slide 18
©
Fixed dose studies
Often the trend favoured lower doses over higher doses
This doesn’t prove that failing patients will not respond to an increase in
dose
But certainly doesn’t prove that they will!
Not studying the situation that we’re interested in
Conclusion: Do allow some comparison between doses – a positive trend
would provide some support for up-titration
Slide 19
©
Flexible dose studies
Patients have their daily dose increased or decreased based upon
response
Placebo control (also “increased” and “decreased”)
Mimics the way treatment used in practice
Slide 20
©
Flexible dose studies
Useful:
Ideal for testing whether a proposed strategy is efficacious
Not so useful:
Cannot tell us whether dose increases are beneficial
Slide 21
©
Flexible dose studies
Generally patients are titrated to the higher doses in these trials:
But that does not been it was necessary to use them
Do not know what would have happened if dose had NOT been increased
Slide 22
©
Flexible dose studies
Patients generally improve with time (even on placebo)
Effect of dose increases confounded with time
Also dose increase is “open” – patients know they have increased and this
may influence response – attempt to analyse before and after change
Although there is a control arm there is no control for the dose increases
Slide 23
©
Flexible dose studies
Useful for marketing – tell us whether higher doses would be used if
available!
Conclusion: Provide no information to support dose increases
Slide 24
©
Randomised non-responder trials
All patients start on “low” dose
Patients not responding randomly assigned to increase dose or stay on
current dose
Dose increase blinded and controlled
Slide 25
©
Randomised non-responder trials
Rarely seen
Ideal design to assess the problem
Experience
One drug: No difference (trend favoured no change)
Another (in paediatric trial): large positive trend
Slide 26
©
Results
“For the majority of SSRIs in the treatment of depressive illness, clinical
trial data do not show an additional benefit from increasing the dose of an
SSRI above the recommended daily dose”
Expert working group report summarised where they felt
recommendations were justified and where they weren’t
Text added to relevant SmPCs stating that benefits of up-titration not
demonstrated in clinical trials
Slide 27
©
Evidence of what happens
Usage data – patients started on higher doses
Quick titrations
Personal experience – prescribers very keen on quick up-titration (even in
presence of adverse events)
Evidence base not good that this is correct strategy
Slide 28
©
Why?
Prescribers personal experiences
“Treating the physicians” – sometimes referred to in cancer as well
Problem – important adverse events difficult to distinguish from lack of
efficacy e.g. suicidality, agitation
This makes up-titration dangerous in some cases
Will EWG report influence things over time?
Slide 29
©
Differences between treatments
Some treatments showed positive trend – others did not
Implausible?
Could be at different part of dose response curve
Products are difference (or they would all have same name)
Don’t definitively understand the mechanisms of how depression occurs –
or why SSRIs work
Slide 30
©
Conclusion
Null-hypothesis disproves the data!
Very informative prior!
Some up-titration decisions based upon an assumption that this is correct
But maybe titration is appropriate – with the data we have we cannot know
for certain
But it would be nice to see some randomised non-responder trials to find
out!
Slide 31
©
Conclusion
Voltaire (in 1759) advocated the induction of ideas from concrete
evidence
Pangloss wilfully ignores any evidence that contradicts his initial opinion
Dosing of SSRIs may be a little Panglossian corner of the evidence based
medicine world
Slide 32
©
Conclusions
End of Phase II decision making and design of Phase II studies:
• Increasing scope of exploratory development programmes should
decrease phase III failure rate. Potential ‘pay-off’ for requirements in
phase III.
• There is a role for increased use of quantification in company decision
making
• Choose design of Phase II studies carefully.
Slide 33
©