Sudaan - koulutus KTL/TTO 2004. 2 Research Triangle Institute.
National Public Health Institute, Finland 1 Open risk assessment Lecture 1: Introduction Jouni...
-
Upload
tyrone-clarke -
Category
Documents
-
view
217 -
download
2
Transcript of National Public Health Institute, Finland 1 Open risk assessment Lecture 1: Introduction Jouni...
National Public Health Institute, Finlandww
w.kt
l.fi
1
Open risk assessment Lecture 1: Introduction
Jouni TuomistoKTL, Finland
National Public Health Institute, Finlandww
w.kt
l.fi
2
Guidance for the workshop• Forget everything you knew about risk
assessment (RA)• you won't need it because our focus is different• During this week, we will describe a new approach
to risk assessment. • Ask briefly – use hand signs
– Write questions down– Thorough discussions should happen on Heande
• All questions answered by the following day
• Don't panic!
National Public Health Institute, Finlandww
w.kt
l.fi
3
Hand signs• A question or comment about…• Beyond my understanding• I agree• I disagree• Move forward
National Public Health Institute, Finlandww
w.kt
l.fi
4
Outline• What is wrong with the current risk assessment?• Why risk assessment is needed in the future?• What is needed from the new risk assessment?• Can it work?• What are the highlights of the workshop?
National Public Health Institute, Finlandww
w.kt
l.fi
5
What is wrong with the current risk assessment?
• Limited area of application• Lack of flexibility and breadth• Inefficiency and slowliness of the process• Deliberate biases towards "safety"• Communication problems• Lack of acceptability among stakeholders
National Public Health Institute, Finlandww
w.kt
l.fi
6
Limited area of application• Only a few chemical groups require pre-market RA:
– Pesticides, drugs, food additives– This will improve with Reach but not disappear
• RA not triggered for many important "natural" exposures:– Traditional foods and food items vs. GMO– Environmental exposures: moldy buildings vs. PM
• Often limited to situations where the release links to someone's economic interest
• Who can and should trigger a RA?
National Public Health Institute, Finlandww
w.kt
l.fi
7
Lack of flexibility and breadth• Each discipline has developed an own framework
– Scientific opinions on food issues by EFSA– Chemical risk assessment for pesticides– Safety assessment for drugs– Life cycle assessment for consumer products– Environmental impact assessments for major construction sites– "Not tested with animals" for cosmetics
• Is this just cultural diversity or a problem of administration and a health hazard?
National Public Health Institute, Finlandww
w.kt
l.fi
8
Importance of boundaries• Risk-benefit analysis of farmed salmon (Tuomisto et al, Science 2004)
BAU
Restrict fish use
30900
23400
-206
-154
If risk managers assume responsibility of total health effect of salmon consumption
If risk managers care only for cancer due to pollutants
30000250000
BAU
Restrict fish use
30900
23400
-206
-154
If risk managers assume responsibility of total health effect of salmon consumption
If risk managers care only for cancer due to pollutants
30000250000
National Public Health Institute, Finlandww
w.kt
l.fi
9
Inefficiency and slowliness of the process• Inefficiency: it takes a lot of person-months to complete
– A lot of expensive expert work– The risk assessments done are not available for others in a
useful format• Slowliness: it takes a lot of calendar months to complete
– The process has data collection, systematic literature searches, public hearings, reviews, scientific advisory panels…
– The dioxin RA by the U.S.EPA: • a draft was published 1996• a second draft was published 2000• …we are still waiting for the final version
• With the same money, there could be more better RAs
National Public Health Institute, Finlandww
w.kt
l.fi
10
Major chemical reviews in IRIS
1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 20070
100
200
300
400
500
600
Chemicals in IRIS website
The most recent major update
Date of update
Num
ber
of c
hem
ical
s
National Public Health Institute, Finlandww
w.kt
l.fi
11
Deliberate biases towards "safety"• Approaches to minimize the false negative error
– Reference dose=NOAEL/UFa/UFi
– BMDL: lower CI of the benchmark dose– LMS (q1
*): linearized multistage• Poorly known chemicals are perceived worse than
well known major hazards• The problems tend to fall out of YOUR mandate
(to others to solve (or ignore))
National Public Health Institute, Finlandww
w.kt
l.fi
12
On whose side does the problem fall?• Risk-benefit analysis of farmed salmon (Tuomisto et al, Science 2004)
BAU
Restrict fish use
30900
23400
-206
-154
If risk managers assume responsibility of total health effect of salmon consumption
If risk managers care only for cancer due to pollutants
30000250000
BAU
Restrict fish use
30900
23400
-206
-154
If risk managers assume responsibility of total health effect of salmon consumption
If risk managers care only for cancer due to pollutants
30000250000
National Public Health Institute, Finlandww
w.kt
l.fi
13
Communication problems• Science-policy: "Decision-makers want clear numbers,
not distributions"• Policy-science: "The Commission wants to promote
distributions; this is an educational issue."• Science-stakeholders: The assessments are not easily
available in the format meaningful for the stakeholders
National Public Health Institute, Finlandww
w.kt
l.fi
14
Lack of acceptability among stakeholders• Stakeholders may not accept
– the premises (e.g. giving monetary values to health)– the group that did the assessment (the energy
company about nuclear waste disposal)– the involvement of the public (we were not heard in
our own cause)– the handling of the contributions (our comments did
not have any impact)
National Public Health Institute, Finlandww
w.kt
l.fi
15
Why is risk assessment needed in the future?• The real problems in the future are NOT those that the
current risk assessment was developed for:– Drugs– Pesticides– Food additives
• It is needed because it would be nice to do something useful for the real risks of the future. The risks that are so complex that no single expert is an expert in all parts of the issue. These risks are such as...
National Public Health Institute, Finlandww
w.kt
l.fi
16
Climate change• Uppsala glacier, Patagonia, Argentina: above, in
1928 and below, in 2004.
National Public Health Institute, Finlandww
w.kt
l.fi
17
Fine particle air pollution
National Public Health Institute, Finlandww
w.kt
l.fi
18
Energy efficiency
National Public Health Institute, Finlandww
w.kt
l.fi
19
Urban living environment
National Public Health Institute, Finlandww
w.kt
l.fi
20
Drinking water amount and quality
National Public Health Institute, Finlandww
w.kt
l.fi
21
Biodiversity
National Public Health Institute, Finlandww
w.kt
l.fi
22
Population growth
National Public Health Institute, Finlandww
w.kt
l.fi
23
Global environmental taxes
National Public Health Institute, Finlandww
w.kt
l.fi
24
What is needed from the new assessment?
Stakeholders must have a say on everything in advanceValue judgements included in the assessment
Lack of acceptability among stakeholders
Everything available for clarification questionsCommunication problems
Best estimates (incl uncertainty) usedDeliberate biases towards "safety"
Info structured & directly reusableDelegation, non-experts includedRoutines automated
Inefficiency and slowliness of the process
Fully scalable to very simple and very complex questions
Lack of flexibility and breadth
Adoptable by any area of administration or policy-making
Limited area of application
National Public Health Institute, Finlandww
w.kt
l.fi
25
Paradigm shift• Thomas Kuhn (1922-1996)• Science progresses in a regular
way until too many faults are identified in the current paradigm. Then, there is a period of extraordinary science, which leads into a shift of paradigm
National Public Health Institute, Finlandww
w.kt
l.fi
26
Pyrkilo• Jouni Tuomisto (1997): It is possible to develop
such a system, pyrkilo, that transforms ideas, information, and people's opinions into a description that tends to converge towards scientific validity.
• After several years of work, the Heande website was opened Sept 2006.
National Public Health Institute, Finlandww
w.kt
l.fi
27
Open assessment• The objectives:
– Find solutions to ALL the challenges at the same time– Systematize and "industrialize" the risk assessment– Maintain high scientific quality
• The current situation: there are suggestions available to all challenges listed within the open assessment– Many of the suggestions have not been tested in practice– Not everything will probably work
• However, there is already a critical mass of solutions available so that full-scale testing can be started
• Further problems should be solved as they appear
National Public Health Institute, Finlandww
w.kt
l.fi
28
What is the acceptability of the idea of open assessment?
• Poll (informal, based on observations of several audiences): – 30 % think it is a stupid idea– 50 % think it cannot work– 15 % find it interesting, but…– 5 % are fond of the idea
• YOU are the 5% of the poll
National Public Health Institute, Finlandww
w.kt
l.fi
29
Can open assessment work?• I am convinced it can work• I am convinced the remaining problems can be
solved• However, this does not mean that it WILL
succeed, at least in our time…
National Public Health Institute, Finlandww
w.kt
l.fi
30
Leonardo's parachute ca. 1500– first applications in 20th century
National Public Health Institute, Finlandww
w.kt
l.fi
31
Bayes' theorem ca. 1750• Reverend Thomas Bayes published the Bayes'
theorem in ca. 1750
– first real applications in 1960's
National Public Health Institute, Finlandww
w.kt
l.fi
32
What is the potential for mass collaboration in environmental health risks?
• 6,000,000,000 people in the world• 1,000,000,000 of them have access to Internet• 10,000,000 of them are seriously thinking about
environmental and/or health problems: ”What could I do?”
• 1,000,000 of them can speak English• 100,000 of them have a good background for the
work (e.g. university degree)• 10,000 of them are willing to spend 1 h/week on
this• 250 person-weeks/week work force available
National Public Health Institute, Finlandww
w.kt
l.fi
33
How would the world look like with full-scale open assessments?
• The turnover of scientific information speeds up• Scientific information is easily available in a readily
useful form always, from anywhere, and in your own language.
• Time is spent on solving problems, not on talking about solving them.
• People start to respond to the politicians suggesting something stupid: ”Did you not even check Heande?”
National Public Health Institute, Finlandww
w.kt
l.fi
34
Open assessment• The research question for the (pyrkilo) method:
– "How can scientific information and value judgements be organised for societal decision-making in such a way that open participation is possible?"
• Full range of development– a new ontological foundation– strictly object-oriented approach– a new structure for information objects– traditional RA methods for processing information, but
organised in a more systematic way– tools that enable open collaboration– data sources that are directly available and applicable
National Public Health Institute, Finlandww
w.kt
l.fi
35
The ORA report
National Public Health Institute, Finlandww
w.kt
l.fi
36
Falsification• Karl Popper (1902-1994)• Science consists of statements
(theories) that can be falsified• Science is an evolutionary
process where poor theories are falsified
• The current knowledge consists of those theories that have not (yet) been falsified
National Public Health Institute, Finlandww
w.kt
l.fi
37
Bayes' theorem• Thomas Bayes (1702-1761)• A posterior probability given new data can be
calculated from a prior and the likelihood of the data
National Public Health Institute, Finlandww
w.kt
l.fi
38
•Decision theory• Howard Raiffa• Decision analysis is a rational
method for making decisions. In addition, the use of subjective (Bayesian) probabilities in decision analysis should be promoted.
National Public Health Institute, Finlandww
w.kt
l.fi
39
Quality of an estimate• Roger Cooke• The quality of a quantitative
estimate (probability distribution) can be evaluated against a golden standard using informativeness and calibration
National Public Health Institute, Finlandww
w.kt
l.fi
40
Vines in Bayesian belief network• Roger Cooke• BBNs describe the reality by using
conditional probabilities• These probability distributions can
have any form and they can still be solved analytically, if vines are used
National Public Health Institute, Finlandww
w.kt
l.fi
41
Argumentation• Frans van Eemeren• Disputes can be solved by using
formal argumentation that consists of attacks and defends of a specified statement
National Public Health Institute, Finlandww
w.kt
l.fi
42
PSSP• Veikko Pohjola• A system can effectively be
described using two kinds of objects: processes and products that are produced by these processes. Each object has attributes purpose, structure, state, and performance.
National Public Health Institute, Finlandww
w.kt
l.fi
43
Wisdom of crowds• James Surowiecki• A group of people is likely to
outperform an individual expert, if they can use individual knowledge, act independently and in a decentralized way, and their opinions are effectively aggregated
National Public Health Institute, Finlandww
w.kt
l.fi
44
Web encyclopedias• Jimbo Wales• Encyclopedia that anyone can
edit: It is possible to motivate a very large group in collecting information and write articles about important issues.
National Public Health Institute, Finlandww
w.kt
l.fi
45
Mass collaboration• Don Tapscott, Anthony Williams• A large group of unorganised
people are able to produce complex artefacts, if the product is information or culture, the work can be chopped into bite-size pieces, and the pieces can be effectively synthesised.