Risk Assessment to Support Decision...
Transcript of Risk Assessment to Support Decision...
Risk Assessment to Support Decision Making
Katherine von Stackelberg, ScD
Center for Climate, Health and the Global Environment
Harvard Center for Risk Analysis
NEK Associates LTD
1.617.998.1037
RISK ASSESSMENT TO SUPPORT DECISION MAKING
Risk = p(Adverse Effect|Exposure, Effects)
Exposure Effects (hazard)
Probability of adverse ecological effect Risk
Fate, transport, transformationmovement in the environment
Laboratory studies
Environmental source of stressors
Fieldstudies
Probabilistic Approaches in Risk Assessment
Range of Values
Freq
ue
ncy
Most are at this value: highest frequency
Source: Using Probabilistic
Methods to Enhance the Role of
Risk Analysis in Decision Making -
Managers' Summary, 2009
Prepared by the EPA Risk
Assessment Forum Working
Group
Concentration-Response Function
Probabilistic Ecological Risk Assessment
% reduction in fecundity
0 50 1000
% of population
1.0
0.5
Probability of
exceedance of a
threshold
Probability of
exceeding an
increasing magnitude
of effect
Conceptual Model of Bt Exposure
Source: Carstens et al. 2012 Transgenic Research 21:813-842
Example: Invasive Species
Source: Kolar et al. 2002 Science 298(8):1233
Example of an Invasive Species Model
Source: http://www.indiana.edu/~preserve/InvasiveSpread/model.html
Output: visualization of probability of invasion
Need to Understand Conditions Under Which Impacts Occur
Can be downloaded from www.biorxiv.org
Examples of outcomes
Offtarget effectsResistance over timeEfficacy – target effects
• The Biosafety Clearing-House (BCH) is a mechanism set up by the Cartagena Protocol on Biosafety to facilitate the exchange of information on Living Modified Organisms (LMOs) and assist the Parties to better comply with their obligations under the Protocol. Global access to a variety of scientific, technical, environmental, legal and capacity building information is provided in the six official languages of the UN.
• https://bch.cbd.int/onlineconferences/ahteg_ra.shtml
• An Ad Hoc Technical Expert Group (AHTEG) on Risk Assessment and Risk Management has a mandate to develop risk assessment guidelines
• Review Article: What Risk Assessments of Genetically Modified Organisms Can Learn from Institutional Analyses of Public Health Risks
• S. Ravi Rajan and Deborah K. Letourneau, Journal of Biomedicine and Biotechnology, Vol 2012, Article ID 203093, http://dx.doi.org/10.1155/2012/203093
Some Examples of Risk from the Literature
Ecosystem Services: Context for Ecological Risk
Source: Munns et al. 2015 Integrated Environmental Assessment and Management, DOI: 10.1002/ieam.1707
Decision Analytic Approaches for Evaluating Alternatives
Risk is not just an academic exercise. There must be a decision context.
Possible alternatives
Criteria against which to compare
alternatives
Example Metrics for Forest Ecosystem Health
Metric Measurable Outcome
Primary productivity Wood growth + litterfall or eddy covariance CO2 flux
Nutrient acquisition Foliar chemistry
Mortality Mortality rate or standing dead trees
Nutrient retention Leaching rate
Water use/evapotranspiration (Precip. – streamflow) or eddy covariance H2O flux
Resilience to moderate stress (pest, disease) Mortality after stress event
Physical structure, age structure, and plant species composition
Forest inventory / canopy assessment
Food web structure Monitoring animal populations
Soil quality Soil chemistry
Biodiversity Abundance
Demand (products, recreation) Market or nonmarket value
Adapted from Gary Lovett’s presentation http://nas-sites.org/dels/files/2018/01/Gary-Lovett-Presentation-Updated.pdf
Take Away 1
• All non-trivial decisions involve the potential for both good and bad outcomes
– Risk applies to more than just “bad” outcomes
• Decisions are informed by risk as a probability and as a perception
• Risk as a probability
– If we make a decision today, what are the probabilities of detrimental and beneficialoutcomes in the future, and how confident are we in our estimates of these probabilities?
Take Away 2
• Risk as perception
– Heuristics; Individual; Voluntary; Familiar (Known)
– “Safe”
• Your audience is regulatory decision and policy makers, their constituencies, and associated stakeholders• Their acceptance of “risk” will likely be driven more by
perception (“safe”) than by any scientific determination (i.e., calculated probabilities)
• Risk of bad outcomes to humans will almost always out-weigh those to ecological resources
Take Away 3
• New ecological assessment tools are available:– Bayesian belief networks
– Weight-of-evidence analysis
– Ecosystem services
– Decision analytics
• But: what constitutes a gene drive hazard/adverse effect?
• Things could go easier if you show that:– Your technology is self-limiting (can’t “escape”)
– Indirect consequences won’t happen• e.g., loss of the target population won’t affect non-target populations
or communities, etc.