Risk modelling: which models to choose?

Area of Science:

• Environmental Health
• Toxicology
• Occupational Health

Background:

• Lung cancer mortality in coke oven workers and lung tumor induction in rats exposed to diesel emissions or cadmium chloride aerosol serve as examples.
• Conventional risk models are frequently used to assess health risks associated with environmental exposures.

Purpose of the Study:

• To determine the maximum likelihood estimate and upper limit of risk for specific environmental exposures.
• To evaluate the additional safety provided by using the upper limit of the 95% confidence interval in risk assessment.

Main Methods:

• Utilized conventional risk models to analyze lung cancer mortality data from coke oven workers.
• Applied risk models to data on lung tumor induction in rats exposed to diesel engine emissions and cadmium chloride aerosol.
• Calculated maximum likelihood estimates and upper confidence limits for risk values.

Main Results:

• The additional safety factor gained by using the upper limit of the 95% confidence interval was less than 5 in most cases, often below 2.
• Model selection was identified as the most critical factor in the accuracy of risk assessment.

Conclusions:

• The choice of risk assessment model significantly impacts the derived risk values.
• An additional safety factor may be necessary to ensure adequate public protection when less conservative models are employed.