Transcript Emerging Role of Pharmaco

Focusing on the key challenges
Decision-making &
drug development
Peter Hertzman
Paul Miller
Rationale
1. From societal perspective the case for
bayesian analysis (BA) to inform adoption
decisions for new technologies is strong
2. From an individual firm’s perspective there
may be a lot more (other) reasons to use
bayesian analysis
3. (2) may be the best foundation for (1)
•
•
benefits (internal to the firm) during drug development
will provide incentives to invest (promote) in BA
this will impact societal HTA
Q1. What are the objectives of
pharmacoeconomics?
st
1
Objective of
PharmacoEconomics
PRICE
Sales
Revenue
VOLUME
2nd Objective of PharmacoEconomics
PRICING APPROVAL
FORMULARY LISTING
Approval
REIMBURSEMENT
(PRODUCT LICENSING)
3rd Objective of PharmacoEconomics
R&D
Costs
Assist internal
decision-making
and resource
allocation during
drug development
R&D Costs
Approval
Sales
Revenue
Q2. Which tools do we use?
The Economist
The Phase III Trial
Q3. Which tools could we use?
CTS
Bayesian Analysis
QoL Assessment
Contingent Valuation
Threshold Analysis
PROs
Longitudinal Databases
Conjoint Analysis
Economic Modelling
Value of Information
So, how can we deliver more?
1. Exploit methodological advances in
economic evaluation and decision
theory
2. Integrate these into a broader range of
activities
3. Review the timing of these activities in
the product lifecycle
Synthesis of
Evidence
Modelling
Gathering
information
Quantifying
Uncertainty
Predicting
Optimizing
decisions
Benefits
Need for Bayesian Analysis
(Regulator’s perspective)
•
•
•
To synthesise all available evidence in an
explicitly quantitative analysis
To quantify uncertainty
To understand the marginal value of more
information
–
Weigh contribution of more information
(certainty) vs. Opportunity costs of delayed
adoption
•
•
•
Adopt in awareness of level of uncertainty; or
Adopt, retricted to more certain domains (populations)
Reject, value of more information > cost to society
Need for Bayesian Analysis
(Pharma perspective)
• Where ’regulator’ requires it! Eg.UK NICE
– Still not viewed as a real barrier to market access
• What proportion of global sales will be affected?
– “only one (small) market” argument
• Does Pharma need to change the way it works?
• Weak incentive: “the stick” is not perceived as big enough!
• Carrots may be more effective!
– scope for bayesian analysis in drug development
process is large
Drug Development Process
1. Test 5,000 -10,000 compounds, to identify
candidates for further development
2. Send approx. 250 for pre-clinical testing
3. Enter approx. 5 into:
– Phase 1 trials (<100 healthy volunteers, to determine
safety and dosage). If successful:
– Phase 2 trials (<300 volunteers, to test for efficacy and
side effects). If successful:
– Phase 3 trials (> 1,000 volunteers, to monitor longer-term
use and adverse reactions). If successful:
4. Approval of the new drug: license
– 10+ years after identification for development
– Cost incurred per NCE = $ 600 million
5. Pricing & Reimbursement discussions
Observations
• Process is long, costly and risky!
• Highly regulated industry:
– Gather drug profile information for
regulatory authorities to make decisions
(license, price, reimburse)
– Some information then used for
promotional claims to persuade customers
to make decisions (also regulated)
Two fundamental questions
1. Which projects do we invest in?
2. How do we maximise the efficiency of the
projects we do choose?
(i.e allocative and technical efficiency issues)
uncertainty
Decisions
Select candidate drugs to develop
Clinical indication?
Clinical trial design?
P&R strategy?
NICE REVIEW
I
Stop/go?
II
III
time
Pilot outcomes & resource use questionnaire
LAUNCH
Preclinical
Phase I
Phase II
Phase III
Phase IV
Collect cost & outcome data
Populate economic models
Inform external decision-makers
Ongoing evaluation in ‘real world’
Pilot outcomes & resource use questionnaire
LAUNCH
Preclinical
Phase I
Scenario Modelling:
estimate c/e ranges
estimate budget impact
determine price bands
Inform internal decisions:
1. Project management
2. Portfolio management
Phase II
Phase III
Phase IV
Collect cost & outcome data
Populate economic models
Inform external decision-makers
Early = data vaccum?
• Not necessarily…
• Use what we do know!
– PK & PD data, CTS, predict drug profile
– Disease knowledge & Epidemiology
– Market knowledge
– Competitor knowledge
– Regulatory requirements
Project Management
• Clinical development programme design:
optimising decisions, eg.
•
•
•
•
•
•
Number and timing of decision points
Speed of development
Order of trials
Dose
Sample size
Sample selection
• About efficiency in trial design
• Optimise what? Intermediate or final endpoints?
Portfolio Management
References:
• Burman CF, Senn S. Examples of option values in drug
development. Pharmaceut Statist. 2003;2:113-125
• Poland B, Wada R. Combining drug-disease and economic
modelling to inform drug development decisions. Drug
Discovery Today 2001: 6(22):1165-1170.
• Shih Y-C T. Bayesian approach in pharmacoeconomics:
relevance to decision-makers. Expert Rev.
Pharmacoeconomics Outcomes Res. 2003; 3(3): 237-250.