1. Philip Bedient, Ph.D. Rice University
Risk Assessment
Philip Bedient, Ph.D.
Rice University

2. Introduction
The concepts of risk and hazard are inextricably intertwined
Hazard - implies a probability of adverse effects in a particular situation
Risk - measure of the probability
The use of the results of a risk assessment to make policy decisions is called risk management

3. A Scientific Point of View
Scientists and engineers use models to calculate an estimated risk, such as:
Tornadoes
Hurricanes
Floods
Droughts

4. Risk Perception Some risks are well quantified
Ex: Frequency and severity of auto accidents are well documented.
In contrast, other hazardous activities, such as those resulting from the use of tobacco and alcohol are more difficult to document

5. Risk Perception
Table 5-1 illustrates different perceptions of risk. Four different groups were asked to rate 30 activities & technologies according to the present risk of death from each.
Table 5-1 shows how each group ranked the risk of of ten out of the original 30 topics.

6. Table 5-1. Ordering of Perceived Risk for 30 Activities and Technologies
Source of Risk
Group 1
LOWV
College students
Active Club Members
Experts
Nuclear Power 1 8 20
Motor vehicles 2 5 3
Handguns 4
Smoking
Motorcycles 6
Alcoholic Beverages 7
General (private) aviation 15 11 12
Police Work 9 17
Pesticides
Surgery 10

7. Putting risk perception in perspective, we can calculate the risk of death from some familiar causes.
Ex: In the U.S. in 2001, there were about 3.9 million deaths per year. Of these, about 541,532 were cancer-related.
The annual risk (assuming a 70-year life expectancy & ignoring age factors) is about:

8. For comparison, Table 5-2 summarizes the risk of dying from some causes of death.
Cause of Death
# of Deaths in Rep. Year
Individual Risk per Year
Black lung disease
1,135
8 x 10-3, or 1/125
Heart attack
724,859
2.7 x 10-3, or 1/370
Cancer
541,532
2.0 x 10-3, or 1/500
Coal mining accident
180
1.3 x 10-3, or 1/770
Fire fighting --
3 x 10-4, or 1/3,333
Motor vehicle
46,000
2.2 x 10-4, or 1/4,545
Truck driving
400
10-4, or 1/10,000
Falls
16,339
7.7 x 10-5, or 1/13,000
Football (averaged over participants)
4 x 10-5, or 1/25,000
Home accidents
25,000
1.2 x 10-5, or 1/83,333
Bicycling
1,000
10-5, or 1/100,000
Air Travel: 1 transcontinental trip/year
2 x 10-6, or 1/500,000

9. Risk Assessment
In 1989, the EPA adopted a formal process for conducting a baseline risk assessment
This process includes:
Data collection and evaluation
Toxicity assessment
Exposure assessment
Risk characterization

10. Data Collection and Evaluation
This part of the process includes:
Gathering background and site information
Preliminary identification of potential human and ecosystem exposure through sampling
To gather background information, it is necessary to find:
Possible contaminants
Concentrations of contaminants in key sources and media
Characteristics of environmental setting

11. Toxicity Assessment
Toxicity Assessment- the process of determining the relationship between the exposure to a contaminant and the increased likelihood of the occurrence or severity of adverse effects
Hazard Identification- determines whether exposure to a contaminant causes increased adverse effects for humans and to what level of severity
Dose- the mass of chemical received by the animal or exposed individual

12. Toxicity Assessment Cont.
Quantal- all-or-nothing responses
Ex.- mortality and tumor formation
Dose-response curve- statistical relationship of organism response to dose
This is usually expressed as a cumulative- frequency distribution

13. Example 5-1
An experiment was developed to ascertain whether a compound has a 5% probability of causing a tumor. The same dose of the compound was administered to 10 groups of 100 test animals. A control group of 100 animals was, with the exception of the test compound, exposed to the same environmental conditions for the same period of time. The following results were obtained:

14. Example 5-1 Results
Group
Number of Tumors A 6 F 9 B 4 G 5 C 10 H 1 D I E 2 J 7
*No tumors were detected in the controls(not likely in reality)

15. Example 5-1 Solution
The average number of excess tumors is 4.9%. These results tend to confirm that the probability of causing a tumor is 5%.
If, instead of using 1000 animals, only 100 animals were used, it is fairly evident from the data that, statistically speaking, some very anomalous results might be achieved. That is, we might find a risk from 1-10%.
Note that a 5% risk is very high in comparison with the EPA’s objective of achieving an environmental contaminant risk of to

16. Toxicity Assessment Terms
Linearized multistage model- a modification of the multistage model for toxicological assessment
This model states that we can extrapolate from high doses to low doses with a straight line
Slope factor- expressed in units of risk per unit dose
Integrated Risk Information- EPA toxicological data base that provides background information on potential carcinogens

17. Slope Factors for Potential Carcinogens
Chemical
CPS0
( kg*day*mg-1)
CPSi
(kg*day*mg -1)
Arsenic
1.5
15.1
Benzene
0.029
Benzo(a)pyrene
7.3
N/A
Cadmium
6.3
Carbon Tetrachloride
0.13
0.0525
Chloroform
0.0061
0.08
Chromium (VI)
42.0
DDT
0.34
1,1-Dichloroethylene
0.6
0.175
Dieldrin
16.0
16.1

18. Slope Factors for Potential Carcinogens Cont.
Chemical
CPS0
( kg*day*mg-1)
CPSi
(kg*day*mg -1)
Heptachlor
4.5
4.55
Hecachloroethane
0.014
Methylene Chloride
0.0075
Polychlorinated Biphenyls
7.7
N/A
2,3,7,8-TCDDb
1.5 * 105
1.16 * 105
Tetrachloroethyleneb
0.052
0.002
Trichloroethylenec w
0.006
Vinyl chlorideb
1.9

19. Limitations of Animal Studies
Most of the effects on people can be produced in species
Exceptions to this include:
Toxicities dependent on immunogenic mechanisms
Subtle toxicity is difficult to transfer from lab animals to people because of the effects of ancillary factors

20. Limitations of Epidemiological Studies
There are four difficulties presented in this type of study:
Large populations are required to detect a low frequency of occurrence of a toxicological effect
A long or highly variable latency period may be needed between the exposure to the toxicant and a measurable effect
Competing causes of the observed toxicological response make it difficult to attribute a direct cause and effect
Epidemiological studies are often based on data collected in specific political boundaries that do not necessarily coincide with environmental boundaries such as those defined by an aquifer or the prevailing wind patters

21. Exposure Assessment
Total exposure assessment- evaluation of all major sources of exposure
Elimination of a pathway of entry can be justified if:
The exposure from a particular pathway is less than that of exposure through another pathway involving the same media at the same exposure point
The magnitude of exposure from the pathway is low
The probability of exposure is low and incidental risk is not high
Reasonable maximum exposure- the highest exposure that is reasonably expected to occur and is intended to be a conservative estimate within the range of possible exposures

22. Intake Equation BW AT CDI- chronic daily intake
CDI= C [ (CR) (EFD) ] (1)
BW AT
CDI- chronic daily intake
C- chemical concentration, contacted over the exposure period
CR- contact rate, the amount of contaminated medium contacted per unit time
EFD- exposure frequency and duration
BW- body weight
AT- averaging time

23. Example 5-2
Problem: Estimate the chronic daily intake CDI of benzene from exposure to a city water supply that contains a benzene concentration equal to the drinking water standard. The allowable drinking water concentration (maximum contaminant level, MCL) is mg*L-1
Solution: From Table 5-7, we note that five routes of exposure are possible from the drinking water medium: 1.) ingestion,dermal contact while 2.) Showering and 3.) swimming, 4.) inhalation of vapor while showering, and 5.) ingestion whie swimming.
CDI= (0.005 mg*L-1)(2.0 L*day-1)(365 days*year-1)(70 years)
(70 kg)(70 years)(365 days*year-1)
= 1.43 * 10-4mg*kg-1*day-1

24. Risk Characterization
For low-dose cancer risk:
Risk = (intake)(slope factor)
For high carcinogenic risk:
Risk = 1-exp[-(intake)(slope factor)]
The noncancer hazard quotient or
hazard index:
HI= intake
RfD

25. Example 5-3
Using the results from Example 5-2, estimate the risk from exposure to drinking water containing the MCL for benzene
Equation 5-21 in the form:
Total exposure risk: riskj may be used to estimate the risk. Because the problem is only to consider one compound, namely benzene, i=1 and others do not need to be considered. Because the total exposure from Example 5-2 included each of the routes of concern for drinking water, that is, all j’s, the final sum may be used to compute risk. The risk is:
Risk= (1.90 * 10-4 mg*kg -1*day -1)(2.9 * 10-2 kg*day*mg -1)
=5.5 * 10 -6

26. Example 5-3 Cont.
This is the total lifetime risk (70 years) for benzene in drinking water at the MCL. Another way of viewing this is to estimate the number of people that might develop cancer. For example, in a population of 2 million:
(2*106) (5.5*10-6) = 11 people might develop cancer
This risk falls within the EPA guidelines of 10-4 to 10-7 risk. It, of course, does not account for all sources of benzene by all routes. Nonetheless, the risk, compared with some other risks in daily life, appears to be quite small.

27. Risk Management
This is performed in order to decide the magnitude of risk that is tolerable in specific circumstances
If a very high certainty in avoiding risk is required, the costs in achieving low concentrations of the contaminant are likely to be high
To reduce risk it is necessary to:
Change the environment
Modify the exposure
Compensate for the effects