The Science Policy Bridge A Framework for Evaluating Complex Scientific Evidence on Environmental Factors in Disease Causation

In preparing a follow-up report to Late Lessons from Early Warnings: The Precautionary Principle 1896-2000 (EEA 2001), the EEA has been developing a framework to assist with the practical application of the precautionary principle via common approaches to evidence evaluation at the science—policy interface. It has also been developing a simple analytical model for approaching such complex, multicausal phenomena as endocrine-disrupting substances, mediated diseases, and childhood asthma (EEA 2003).

The draft EEA framework in Table 9.3 uses just three strengths of evidence: weak (10-33% estimated probability), moderate (33-66% estimated probability), and strong (more than 66% probability), which are the same as the "low likelihood," "medium likelihood," and "likely" categories of the IPCC (Table 9.4). The draft framework also invites users to judge whether the overall evidence has become stronger or weaker over a relevant period of time between major evaluations of the evidence or since, say, 1992.

Preventive and precautionary actions must usually be taken on the basis of much less than scientific certainty and well before an understanding of the mechanisms of action

Table 9.3. An EEA framework for evaluating complex and conflicting scientific evidence on environment and disease.

Statement of Hypothesis* Evaluation Factor Overall Strength of Evidence

Table 9.3. An EEA framework for evaluating complex and conflicting scientific evidence on environment and disease.

Statement of Hypothesis* Evaluation Factor Overall Strength of Evidence

Outcome (e.g.,

Relevant exposure Associationf Plausibilityf Causalityf

Mechanism of

Overall weight

Direction of

childhood asthma)

(e.g., indoor

action?f

of evidence

evidence

air pollution

(weak, moderate,

(stronger or

[NOJ)

or strong) $

weaker over last 5-10 years)

Note: This table is designed to be used to evaluate the scientific evidence for each hypothesis (at an appropriate level of detail, relevant to the expertise of those doing the evaluation) against the factors in the framework and judge the overall strength of evidence for the hypothesis and the direction in which the evidence is moving.

*From observations or theories.

tCommonly accepted among relevant scientists.

^Commonly accepted by relevant scientists as "weak," "moderate," or "strong," categories based on Hill (1965), IPCC (2001), and IPCS/WHO (2002).

Table 9.4. Different levels of proof for different purposes: Some examples and illustrations.

Quantitative Descriptor Qualitative Descriptor Illustrations

(probability bands based on IPCC 2001)*

Quantitative Descriptor Qualitative Descriptor Illustrations

(probability bands based on IPCC 2001)*

100% probability Very likely (90-99%)

• "Statistical significance"

• Part of strong scientific "causation" evidence

• "Beyond all reasonable

• Most criminal law; the Swedish chemical law, 1973

doubt"

(for evidence of "safety" from manufacturers)

90%

Likely (66-90%)

• "Reasonable certainty"

• "Sufficient scientific evidence"

• World Trade Organisation SPS Agreement, Art. 2.2, 1995, to justify a trade restriction

• International Agency for Research on Cancer (IARC) Category 1: "Probable Human Carcinogen"

£

Medium likelihood

• "Balance of evidence"

• Intergovernmental Panel on Climate Change 1995 and 2001

(33-66%)

• "Balance of probabilities"

• Much civil and some administrative law

• "Limited evidence"

• IARC Category 2 B: "Possible Human Carcinogen"

rt -D

• "Reasonable grounds for

• European Commission on the Precautionary Principle 2000

O JH

50%

concern"

• British Nuclear Fuels occupational radiation compensation

bo c

• "Strong possibility"

scheme, 1984 (20-50% probabilities triggering different

• "Scientific suspicion of risk"

awards up to 50%, which triggers full compensation)

'i

• Swedish chemical law, 1973, for evidence required for

14

regulators to take precautionary action on potential harm

V

from substances

Low likelihood

• "Some evidence of

• IARC criterion for selecting substances for evaluation

10%

(10-33%) Very unlikely

carcinogenicity"

• "Available pertinent information"

• Low risk

• WTO SPS Agreement, Art. 5.7, to justify a provisional trade restriction where "scientific information is insufficient"

• Household fire insurance

0% probability

(1-10%)

• "Negligible and insignificant"

• Food Quality Protection Act, 1996 (U.S.)

"^Simplified by removing the top (>99%) and bottom (<1%) levels. There is rarely precision on contested issues to allocate specific numerical probabilities, and the boundaries between categories are in practice fuzzy: The quantitative descriptors illustrate broad categories of evidence based on has been achieved. The appropriate level of proof varies in each case, depending on the likely nature and scale of the hazards and the availability and feasibility of alternatives.

After further discussion and improvements, the EEA believes that this framework for evaluating scientific evidence will be a helpful tool in the process of producing consistent overviews of the existing states of knowledge.

Was this article helpful?

0 0
Whole Health

Whole Health

Get All The Support And Guidance You Need To Be A Success At Better Total Health. This Book Is One Of The Most Valuable Resources In The World When It Comes To An Introduction To Your Overall Health For You And Your Loved Ones.

Get My Free Ebook


Post a comment