Beneris (Benefit-risk assessment for food: an iterative value-of-information approach)?

1
Beneris (Benefit-risk assessment for food an
iterative value-of-information approach)?

• Jouni Tuomisto
• KTL, Finland

2
Partners
Participants
3
Objectives (selection)?

• A framework for handling complicated benefit-risk
  situations
• Benefit-risk analysis methods
• Bayesian belief networks (BBN)?
• Methods for dose-response assessment, combining
  epidemiological and toxicological data
• A result database for information relevant for
  benefit-risk assessments
• Food risks and benefits
• To estimate nutrient intakes and food consumption
  in various subgroups
• To identify food consumption patterns and food
  choices that determine the intake
• Dissemination
• To integrate results into updated benefit-risk
  assessments, and evaluate the remaining
  uncertainties and their importance for
  decision-making.
• To develop an internet interface for publishing
  risk assessment results.
• To develop methods to collect feedback from
  end-users about benefit-risk analyses.

4
Timeline

• Project started April 1, 2006
• Heande website opened September, 2006
• Open Risk Assessment report September, 2007
• Mid-term meeting November 7-9, 2007
• Result database opened January, 2008
• Full case study, fish fall 2008
• Full case study, vegatable spring 2009
• Final project meeting June, 2009
• Project ends September 30, 2009

5
The ORA report
6
Results and deliverables achieved

• Bayesian belief network (BBN) on fish prepared
• Work on parameter values under way
• A methodology report
• Tuomisto and Pohjola Open Risk Assessment, 2007.
• A website for making open assessments
  http//heande.pyrkilo.fi
• A test database for the data repositoryhttp//www
  .pyrkilo.fi/resultdb

7
Methods and Approaches

• Open assessment
• A general assessment method that enables
  unrestricted participation (i.e. mass
  collaboration) at all phases of the assessment
  process
• Applies a defined information structure causal
  diagrams with variables
• Formal argumentation is used to resolve disputes
• Bayesian belief networks as the decision support
  system

8
(No Transcript)
9
Potential synergies and/or overlaps with BRAFO

• Synergies
• The same aim method development for benefit-risk
  analysis
• The methods and practices already developed in
  Beneris are available for Brafo to use
• The Heande website is available for doing case
  studies
• Overlaps
• Dissemination aims at the same audience?

10
SWOT analysis
Strengths Tight connections to methodology development in several EU-funded projects Intarese, Heimtsa, Hiwate, Qalibra Work done on several levels theoretical foundation, methods and application, computer tools Weaknesses Lack of resources in coordination The combining of different parts together is running late (e.g. nutrient data to Bayesian belief networks) Too much focus on theory, too little on food benefit-risk
Opportunities Methods widely applicable food, environmental health, other domains... Mass collaboration a quicker and more efficient way of making assessments Result database as an open information source Threats Failure in convergence of project approaches Mass collaboration not accepted as a way to work
11
Case studies

• Fish benefits of nutrients and risks of
  pollutants in fish
• Dioxin, PCB, methyl mercury
• Omega-3 fatty acids, selenium, iodine
• Cardiovascular and cancer mortality, IQ loss,
  developmental defects (teeth)?
• Vegetables impacts of vegetable-rich and
  vegetable-poor diets in children
• The detailed scoping under way

12
BBN fish case study
13
Quality of data

• Large national studies on food intake of children
  and adults from Finland, Denmark, Ireland, Spain
• National studies on pollutant concentrations in
  various foodstuff
• Exposure-response functions for the endpoints
  mainly from scientific literature, and also from
  in-house expert judgement

14
Comments on Benefit-risk assessment tiered
approach

• Jouni T. Tuomisto
• National Public Health Institute (KTL), Finland

15
Important points

• Question must be clear and for a need!
• Iterative approach
• Transparency
• Need for procedural decisions acknowledged
• Utilises approaches developed in other areas
  DALYs, QALYs

16
Comments on procedural decisions

• Who actually decides what is needed or
  sufficient?
• About the main questions asked.
• About the outcomes considered.
• About when the preference between scenarios is
  clear enough.
• What is the basis for these decisions? Are the
  criteria explicated in an assessment?
• ?Truth should be used as the ultimate criterion

17
Comments on Margin of Exposure (MoE)

• DALYs of QALYs can be used in measuring both
  risks and benefits
• The use of MoE is ambiguous and should be
  discouraged.
• 10 impact on a risk and on a benefit are NOT
  comparable in any meaningful way.

18
Comments on the process

• Problem definition is an iterative process. ?
  If the main question changes, when is the
  assessment no longer the original assessment?
• If the approach is a general approach, it should
  work fine with risk assessments (no benefits) and
  benefit assessments (no risks) as well.