Evolutionary Health
Co-Evolution of Disease & Living Conditions
Health Effects
What is Risk?
Environmental Risk
Risk Assessment
Risk Abatement
Risk Perception
Risk Management
Uncertainty & Other Features of Risk Assessment
Precautionary Principle
Appendix 1: Contaminants
Appendix 2: Environmnet & Reproductive Health
Internet Links
Other Resources
Health & Risk System PDF
Printer-Friendly Web Version

Risk Assessment


List the major aspects that determine a person's health. Rank these by importance, taking into account the most important factors.

What do you learn from this exercise about the way in which the natural environment affects our health?

Risk assessment is the main method used by agencies such as the EPA to estimate the environmental impact of pollution, to compare different alternatives for managing pollution, and to set regulatory goals and guidelines that industry and municipalities must meet. Why? Because you can calculate a number known as a risk, and numbers are easier to work with when doing comparisons. For example, we said the wastewater treatment plant did not get rid of the pollution, but merely altered its form and where it ended up. Is that good or bad? One way to answer that question is to determine the total risk before treatment (discharge direct to water body) versus the total risk after treatment including the risk from air emissions, sludge disposal, disinfectant byproducts in the water body, and treated wastewater disposal.

That brings up another question—calculate the risk to what? Most of the risk assessments performed currently are for the risk to human health, however you could think of risk in terms of risk to a particular species in the ecosystem/biosphere, etc. The general framework for a human health risk assessment is presented below since this is what we currently know the most about.

Figure 5: Risk Mangagement Scheme

Figure 5 shows the general scheme involved in assessing and managing risk. The first step is identifying the concerns and magnitude to see what type of hazard is posed by an exposure event that poses risk. The information about biological effects of the agent may already be available—or, as in the case of many new synthetic chemicals, and microbes, it may not be. These may be a whole scientific project that may be needed to obtain this information.

Rather than describing the steps in theory, the next section goes through a specific example of how a risk assessment calculation is done.

Sample Risk Assessment - Groundwater Contamination
At many industrial sites, chemicals are stored in underground storage tanks (USTs). At a particular site, an UST containing chemical X is known to be leaking. An existing monitoring well close to the tank showed that chemical X has reached the groundwater below. A neighboring town gets part of its water supply from the contaminated aquifer and part from another source. How would you determine if there is a risk to the town? List the various steps you need to do and what information you must have to do a quantitative risk assessment.

The general EPA framework for quantitative risk assessment is as follows:

1. Hazard Identification
This is the gathering of existing information about a particular chemical's risk:

  • What chemicals was subject exposed to?
  • Can the chemicals have adverse health effects?
  • What are those effects - mutagenesis (mutations to genetic material), carcinogens (cancer causing), teratogens (birth defects), acutely toxic (short-term illnesses and death)?
  • What are the various pathways by which a chemical could affect subject population - ingestion, inhalation, dermal?

Note that the effects are determined by two methods: toxicity testing with animals and epidemiological studies of previous exposures. However, there are problems with both methods including:

a) toxicity testing with animals: expose animals (rats, mice, guinea pigs) to varying (but large) doses of chemical and monitor mutagen occurrence, tumors in organs, and death rates. Problems are 1) extrapolate (often linearly) the effects caused by large doses (needed so can see effects quickly) to low doses that humans are more typically exposed to, 2) humans have different metabolic rates than the animals tested.
b) epidemiological studies: study (over the long term) the effects on individuals who have been inadvertently exposed to a chemical and comparing their health with the general population. Problems are 1) since it is not a controlled study there are many confounding factors such as diet, smoking habits, genetics, etc., 2) takes a fairly long time to obtain conclusive results since you typically need to monitor subjects over a lifetime


2. Exposure Assessment
To have a risk, the subject population has to be exposed to the chemical.

  • What are the possible pathways? (e.g., dermal, ingestion, inhalation of water air, soil, and/or food)
  • How much contact is there? (How much chemical is in environment, what happens to the chemical along a pathway (degradation), how much chemical reaches an exposed person?)
  • How much chemical gets into a person's body (dose), considering body weight, length of exposure, bioconcentration, etc., in units of mg/kg/day?

The EPA typically assumes the subject is a 70 kg male who lives 70 years in the same location. Problems with this assumption are many including the role of ethnicity, sex, and age.

3. Dose-Response Assessment
This includes taking what is known about the chemical scientifically and accounting for the actual dose a person is exposed to and determines the risk.

Toxicologists develop relationships between toxicity testing information and risk that are called dose-response curves. These curves are actually linear - for carcinogens they go through 0 and for acutely toxic chemicals there is a threshold dose. This data exists for several chemicals and can be obtained from the National Institute of Health, IRIS database, EPA. For carcinogens the relationship is known as the potency factor (mg/KG/day)-1. For acute toxics the relationship is known as LOEF, NOEL, RfD.

4. Risk Characterization
This is where you take all the information (qualitative and quantitative) from above and determine if there is a risk that needs to be managed. Some suggestions are as follows:

  • Calculate individual risk
  • Calculate population risk - typically EPA sets maximum risk at 1 in a million for most regulations
  • Compare risk to existing death rates, etc.
  • Compare risk to regulated standards
  • Consider uncertainty, problems with data, conservatism, etc.

Let’s decide if there is a significant risk to the community. X concentration in the groundwater below UST is 300 mg/L. Groundwater flows at 1 ft/day and the supply well is 1 mile from the UST. The pump rate is 75,000 gpd. The groundwater is mixed with uncontaminated water for a total supply of 1 Mgpd for 50,000 people. The chemical decays with a half-life of 10 years assuming a first order decay rate, and the rate that X flows is half the rate of the groundwater. Assume no dispersion. The potency factor for X is .02 (mg/Kg/day-1). Note that in the US the cancer death rate is 193/100,000.

Hazard Identification:
X is a carcinogen
Cpf = 0.02 (mg/kg-day)-1
Pathways = ingestion, dermal, if volatile then inhalation

Exposure Assessment:
only consider ingestion
chemical decays and absorbs to soil so lower concentration at well
velocity is 0.5 ft/day due to sorption
dose has to account for amount of water per day drank - EPA scenario assumes 2 L per day for a 70 kg male over 70 years
Time to get to well = 1mile/0.5 ft per day = 5280/0.5 = 10,560 days
k = 0.693/(10*365) = 1.9 * 10-4 day-1 (C= C0e-kt)
C = 300 e-(1.9*10-4*10560) = 40 ug/L
In drinking water conc = (75000*40+0)/ 1000000 = 3 ug/L
Daily dose = 3*2/70 = 85.7 * 10-6 mg/kg/day

Dose Response Assessment:
risk = 0.02*85.7*10-6 = 1.7 * 10-6 or approximately 2 in a million

Hazard Charactarization:
lifetime individual risk = 2 in a million
EPA standard typically 1 in a million
Population risk = (50000*1.7)/(70*1000000) = 0.001 people will die annually of cancer in town
existing US cancer death rate is 193 in 100000 so for 50000 = 96.5 people per year
conservative with lots of uncertainty




  ©Copyright 2003 Carnegie Mellon University
This material is based upon work supported by the National Science Foundation under Grant Number 9653194. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.