Methodological Guidelines for Economic Evaluation The International Experience

1
Methodological Guidelines for Economic
EvaluationThe International Experience

• Michael Drummond
• Centre for Health Economics
• University of York

2
Growing use of economic evaluation in
decision-making

• Several jurisdictions have introduced
  cost-effectiveness data as a requirement in
  decisions about the reimbursement of health
  technologies
• These include Australia, Canada, South Korea, New
  Zealand and several European countries
• In the USA more than 50 organisations, covering
  more than 100 million lives, are asking for
  economic data

3
The scope of requirements differs from place to
place

• All drugs (including existing ones)
• Sweden
• All new drugs (outside hospital)
• Australia, Canada, Scotland
• Drugs where a premium price is requested
• The Netherlands
• Drugs with a major impact
• England and Wales, Portugal

4
Economic evaluation co-exists with other policies

• Reference pricing, based on clustering or
  international comparisons
• Generic substitution
• Budgetary restrictions
• Incentives for cost-effective prescribing

5
Methodological Guidelines for Economic Evaluation

• Range and scope of guidelines
• Areas of controversy
• Views on the role of modelling

6
Range and Scope of Guidelines

• Two good sources
• - Hjelmgren et al. Value in Health 2001 4(3)
  225-250.
• http//www.ispor.org/PEguidelines.
• comp3.asp
• Around 25-30 guidelines in total
• Purposes of guidelines
• - For reimbursement submissions (mandatory and
  voluntary)
• - For general improvement of methods

7
Areas of Controversy

• Hjelmgren et al report agreement on 75 of
  methodological principles
• The main differences are in guidance on
• study perspective (e.g. inclusion of
  non-healthcare costs)
• economic outcomes (e.g. QALYs, WTP)
• handling uncertainty

8
Why do guidelines differ?

• Some important differences in national
  perspective (e.g. views on equity in healthcare)
• Methods develop over time
• Guidelines reflect the views of those who develop
  them, which may or may not reflect the wider
  consensus
• Some guidelines may contain methodological errors

9
Views on the Role of Modelling

• All guidelines recognise that there is a role for
  modelling, with varying degrees of suspicion
• Some restrict it to situations where
  effectiveness data are not available or
  applicable
• The NICE guidelines see a broad role, where
• patients in trials do not match the typical
  patient
• intermediate outcome measures are used
• relevant comparators have not been used
• long term costs and benefits extend beyond the
  trial

10
THE NICE Reference Case

• Developed by an expert working group in early
  2004.
• To be adopted by NICE in early 2005.
• Applies to both companies making submissions and
  the independent review teams undertaking
  Technology Assessment Reviews for NICE.

11
Summary of Reference Case Requirements (1)
12
Summary of Reference Case Requirements (2)
13
Summary of Reference Case Requirements (3)
14
Current Challenges

• Methodological challenges.
• Practical and policy challenges.

15
Current Methodological Challenges

• Making indirect comparisons (in the absence of
  head-to-head clinical studies).
• Extrapolating beyond the duration of clinical
  trials.
• Analysing multinational, economic, clinical
  trials.
• Incorporating equity considerations.

16
Making Indirect Comparisons (in the absence of
Head-to-Head Clinical Studies)

• Unless the decision can wait until such studies
  are available, some modelling/synthesis is
  required.
• Some studies show a reasonable agreement between
  head-to-head studies and indirect comparisons, in
  cases where studies with a common (third)
  treatment are used (Song et al, 2003).
• More complex models can be used when studies with
  common treatment are not available (e.g.
  multi-parameter synthesis).

17
Extrapolating Beyond the Duration of Clinical
Trials

• A full economic evaluation requires long-term
  outcomes (e.g. life-years gained, QALYs gained).
• Normally the decision (on use of the technology)
  cannot wait until long-term data are available.
• Projections are needed for (i) maintenance of
  treatment effect (ii) rates of withdrawal from
  therapy (iii) implications of withdrawal.
• There is no unambiguously right way to make these
  projections.

18
Analysing Multinational Economic Clinical Trials

• In these trials data on resource use are
  collected as part of the trial protocol.
• Doubts exist whether the resource data can be
  pooled, as is the normal approach for the
  clinical data.
• Several analytic strategies have been proposed,
  including the use of multilevel models, but more
  validation is required.

19
Incorporating Equity Considerations

• In most countries equity considerations are as
  important as efficiency considerations.
• Criteria include the patients socio-economic
  status, the seriousness of the health condition
  and the amount of good health already enjoyed.
• Decisions of reimbursement agencies seem to
  reflect these considerations.

20
Incorporating Equity Considerations

• Some methods for valuing health states partly
  take these into account (e.g. the person
  trade-off approach).
• Some economists propose explicit weighting of
  QALYs (e.g. the fair innings approach).
• The main question is whether these considerations
  are best made a part of the economic analysis, or
  tackled separately.

21
Current Practical Challenges for Jurisdictions
Introducing the Fourth Hurdle

• Do we request evidence for all new drugs, or just
  some?
• How do we prioritize drugs for assessment?
• Does it make sense to assess several drugs in the
  same class together?
• How prescriptive, or flexible, should we be in
  specifying the data requirements?

22
Current Practical Challenges for Jurisdictions
Introducing the Fourth Hurdle (continued)

• Should we be willing to accept data from other
  jurisdictions? If so, which?
• Should we be willing to accept commercial-in-confi
  dence data submitted by companies?
• Should the reasons for reimbursement decisions be
  made public?
• Should there be an appeals process? If so, what
  should this consist of?

23
Current Practical Challenges for Jurisdictions
Introducing the Fourth Hurdle (continued)

• Should we consider a two-stage appraisal process?
• Should we consider risk-sharing deals with
  companies?

24
Accepting Data from Other Jurisdictions

• An important issue for jurisdictions with limited
  resources to undertake or assess economic
  evaluations.
• The key issue is whether economic studies are
  generalizable from one setting to another.

25
Factors Likely to Limit Generalizability of
Economic Studies

• Demography and epidemiology of disease.
• Clinical practice patterns.
• Relative price differences.
• Incentives to health professionals or
  institutions.
• Community valuations of health and health care.

26
Which Data from other Jurisdictions can be
Accepted?

• Effectiveness?
• Resource use?
• Unit costs/prices?
• Health state valuations?

27
Conclusions

• Several jurisdictions now request
  cost-effectiveness data in respect of drug
  pricing or reimbursement decisions.
• These decision-making processes have proved
  workable, although many problems/ issues have
  emerged.
• Other jurisdictions introducing the Fourth
  Hurdle can learn from others experiences.