RISK ASSESSMENT AND MANAGEMENT

Fanny K. Ennever

Fall Term, 2003

Risk Assessment

Risk assessment  the process of correlating the amount of exposure with the amount of harm.

The question:

How much of a chemical is OK?

Steps in Risk Assessment

Hazard identification

Exposure assessment (DOSE)

Quantitative toxicological assessment (DOSE-RESPONSE)

Risk characterization

Risk Management

Decisions on whether to act and how

Uses the numbers from risk assessment

Considers cost of alternatives

Is influenced by risk perception

1. Hazard Identification

Toxicological concepts:

Any substance is toxic if dose is high enough, but only some chemicals can cause cancer

Non-cancer toxicity: Protecting against the most sensitive effect protects against all effects: “threshold”

Cancer: Any dose of a carcinogen carries some risk, but the smaller the dose, the smaller the risk

Key question for hazard identification:
Is it a carcinogen or not?

Current methods:

Epidemiology

Animal testing

In vitro (bacterial and mammalian cell) testing

Structure-activity relationships

Scope of the identification problem

Synthetic chemicals cause only

1-5% of all human cancers

>1 million chemical substances are known

~3 thousand produced in high volumes

Full information available for 7%

No information available for 43%

Tests (mutagenicity but not carcinogenicity) cost $200,000 per chemical

Questions in Hazard Identification

Is human cancer predicted well enough by

animal cancer tests?

mutagenicity?

Are we controlling the right chemicals?

2. Exposure Assessment

Who’s Exposure?

Numerical estimate of exposure

Must know frequency and duration of contact

Depends on physiology and activities

Uncertainty

Report both central tendency and upper bound values

Example – methylene chloride (MC) in soil

Adult central tendency soil ingestion:

3. Quantitative Toxicological
Assessment

Non-Cancer Toxicity (has a threshold)

Determine which species, durations, and endpoints have been studied

ß

Identify the most sensitive effect

No Observed Adverse Effect Level: NOAEL (or Lowest Observed Adverse Effect Level: LOAEL)

ß

Non-Cancer Toxicity
(continued)

NOAEL (or LOAEL)

ß

Use uncertainty factors to account for within-human variability (¸10), animal-to-human

variability (¸10), threshold (¸10), durations (¸10), and completeness of data (¸10)

ß

“Safe” dose = RfD

Derivation of Reference Dose (RfD)

Example – methylene chloride RfD

NOAEL: 5.85 (male mice) and 6.47 (female mice) mg/kg/day, liver toxicity

Uncertainty factors: 10 for within-human variability and 10 for animal-to-human variability

RfD = 6 × 10^-2 mg/kg/day

Only if the chemical is a
“carcinogen”

Cancer toxicity (no threshold)

Identify the most sensitive tumor

ß

Extrapolate risk to low doses

ß

An estimate of carcinogenic potency

Uncertainties in low-dose extrapolation methods

Linerized Multistage Model

P(d) = 1 – exp[-(q₀ + q₁d + q₂d² + …+ q_kd^k)]

q₁ – coefficient of linear term

q₁* – upper 95% confidence limit of q₁

– also called Slope Factor (SF)

– used by EPA for carcinogenic potency

Methylene Chloride Slope Factor

4. Risk Characterization

Calculate risk by comparing calculated dose to dose-response

Noncancer: No risk if dose is less than safe dose (RfD)

Cancer: Risk = dose ´ Slope Factor

Example – methylene chloride

Dose = 2.3 × 10^-3 mg MC/kg/day

RfD = 6 × 10^-2 mg/kg/day

→ Dose is less than RfD so no noncancer risk

SF = 7.5 × 10^-3 per (mg/kg/day)

Risk = 1.7 × 10^-5

Risk assessment is done!

Now What?

Rules of thumb

If risk is less than 10^-⁶ rarely take action

If risk is greater than 10^-⁴ usually take action

Cost-Benefit Analysis

Risk analysis:

How many premature deaths would action X prevent?

Cost analysis:

How much would action X cost?

Benefit analysis:

How much is preventing each premature death worth?

Approaches to benefit analysis

Human capital

Willingness-to-pay

Survey

Occupational behavior

Consumer behavior

Credible range from above:

$2.1 million to $11 million (1995 dollars)

$2.5 million to $13 million (2003 dollars)

Risk Perception

Dread Factor

Perceived as More Risky Perceived as Less Risky

Uncontrollable Controllable

Involuntary Voluntary

Inequitable Equitable

Dread result Commonplace result

Global consequences Localized consequences

Risk to future generations Risk to existing people

Risk Perception (continued)

Familiarity Factor

Perceived as More Risky Perceived as Less Risky

New risk Old risk

Not observable Observable

Delayed effect Immediate effect

No scientific consensus Scientific consensus

EPA's Seven Cardinal Rules of Risk Communication

CR 1 — Accept and involve the public as a legitimate partner.

CR 2 — Plan carefully and evaluate your performance.

CR 3 — Listen to the public's concerns and feelings.

CR 4 — Be honest, open and frank.

CR 5 — Coordinate and collaborate with other credible sources.

CR 6 — Meet the needs of the media.

CR 7 — Speak clearly and with compassion, kindness and respect.

Guide to Ineffective Risk Communication

1. Avoid eye contact, keep your arms and legs crossed, and act nervous and/or bored

2. Use jargon and mountains of technical details

3. Emphasize the benefits of industry and the cost of cleanup

Guide to Ineffective Risk Communication (continued)

4. Blame others for mistakes and confusion

5. Make unrealistic promises

6. Be sarcastic when people express concerns or don't understand you

7. Give long, prepared, technical speeches when someone asks a question

Guide to Ineffective Risk Communication (continued)

8. Get angry; attack opponents

9. Refuse to answer personal questions

10. Minimize risks and make inappropriate comparisons

Bottom Line

Risk assessment can’t give the “right” answer

More modest goal:

Assessments are

Consistent

Transparent


	Risk assessment  the process of correlating the amount of exposure with the amount of harm.

	The question:
	How much of a chemical is OK?


		Hazard identification
		Exposure assessment (DOSE)
		Quantitative toxicological assessment (DOSE-RESPONSE)
		Risk characterization


		Decisions on whether to act and how
		Uses the numbers from risk assessment
		Considers cost of alternatives
		Is influenced by risk perception


	Toxicological concepts:
	Any substance is toxic if dose is high enough, but only some chemicals can cause cancer
	Non-cancer toxicity: Protecting against the most sensitive effect protects against all effects: “threshold”
	Cancer: Any dose of a carcinogen carries some risk, but the smaller the dose, the smaller the risk


	Current methods:
	Epidemiology
	Animal testing
	In vitro (bacterial and mammalian cell) testing
	Structure-activity relationships


	Synthetic chemicals cause only
	1-5% of all human cancers
	>1 million chemical substances are known
	~3 thousand produced in high volumes
		Full information available for 7%
		No information available for 43%
	Tests (mutagenicity but not carcinogenicity) cost $200,000 per chemical


	Is human cancer predicted well enough by
		animal cancer tests?
		mutagenicity?
	Are we controlling the right chemicals?


	Numerical estimate of exposure
	Must know frequency and duration of contact
	Depends on physiology and activities

	Uncertainty
	Report both central tendency and upper bound values


	Non-Cancer Toxicity (has a threshold)

	Determine which species, durations, and endpoints have been studied
	ß
	Identify the most sensitive effect
	No Observed Adverse Effect Level: NOAEL (or Lowest Observed Adverse Effect Level: LOAEL)
	ß


	NOAEL (or LOAEL)
	ß
	Use uncertainty factors to account for within-human variability (¸10), animal-to-human
	variability (¸10), threshold (¸10), durations (¸10), and completeness of data (¸10)
	ß
	“Safe” dose = RfD


	NOAEL: 5.85 (male mice) and 6.47 (female mice) mg/kg/day, liver toxicity
	Uncertainty factors: 10 for within-human variability and 10 for animal-to-human variability
	RfD = 6 × 10^-2 mg/kg/day


	Cancer toxicity (no threshold)

	Identify the most sensitive tumor
	ß
	Extrapolate risk to low doses
	ß
	An estimate of carcinogenic potency


	P(d) = 1 – exp[-(q₀ + q₁d + q₂d² + …+ q_kd^k)]

	q₁ – coefficient of linear term

	q₁* – upper 95% confidence limit of q₁

	– also called Slope Factor (SF)

	– used by EPA for carcinogenic potency


	Calculate risk by comparing calculated dose to dose-response

	Noncancer: No risk if dose is less than safe dose (RfD)

	Cancer: Risk = dose ´ Slope Factor


	Dose = 2.3 × 10^-3 mg MC/kg/day
	RfD = 6 × 10^-2 mg/kg/day
	→ Dose is less than RfD so no noncancer risk
	SF = 7.5 × 10^-3 per (mg/kg/day)
	Risk = 1.7 × 10^-5


	If risk is less than 10^-⁶ rarely take action

	If risk is greater than 10^-⁴ usually take action


	Risk analysis:
		How many premature deaths would action X prevent?
	Cost analysis:
		How much would action X cost?
	Benefit analysis:
		How much is preventing each premature death worth?


	Human capital
	Willingness-to-pay
		Survey
		Occupational behavior
		Consumer behavior
	Credible range from above:
		$2.1 million to $11 million (1995 dollars)
		$2.5 million to $13 million (2003 dollars)


	Dread Factor
	Perceived as More Risky Perceived as Less Risky
	Uncontrollable Controllable
	Involuntary Voluntary
	Inequitable Equitable
	Dread result Commonplace result
	Global consequences Localized consequences
	Risk to future generations Risk to existing people


	Familiarity Factor
	Perceived as More Risky Perceived as Less Risky
	New risk Old risk
	Not observable Observable
	Delayed effect Immediate effect
	No scientific consensus Scientific consensus


	CR 1 — Accept and involve the public as a legitimate partner.
	CR 2 — Plan carefully and evaluate your performance.
	CR 3 — Listen to the public's concerns and feelings.
	CR 4 — Be honest, open and frank.
	CR 5 — Coordinate and collaborate with other credible sources.
	CR 6 — Meet the needs of the media.
	CR 7 — Speak clearly and with compassion, kindness and respect.


	1. Avoid eye contact, keep your arms and legs crossed, and act nervous and/or bored
	2. Use jargon and mountains of technical details
	3. Emphasize the benefits of industry and the cost of cleanup


	4. Blame others for mistakes and confusion
	5. Make unrealistic promises
	6. Be sarcastic when people express concerns or don't understand you
	7. Give long, prepared, technical speeches when someone asks a question



	8. Get angry; attack opponents
	9. Refuse to answer personal questions
	10. Minimize risks and make inappropriate comparisons


Risk assessment can’t give the “right” answer

More modest goal:
	Assessments are
		Consistent
		Transparent