NIOSHTIC-2 Publications Search
Estimation of Risk and Inferring Causality in Epidemiology.
NIOSH 1985 May:22 pages
Estimating risk and determining causality in epidemiological studies were discussed. Indices of risk used in epidemiologic studies such as absolute risk, relative risk, odds ratio, standardized mortality ratio, proportionate mortality ratio, and attributable risk were described. Causal inferences were discussed. It is noted that although the concept of risk is closely associated with that of causality, very rarely does a situation exist where necessary and sufficient cause can be clearly determined. The traditional concept of causation as applied to infectious diseases is not applicable to chronic diseases such as lung cancer. In a risk assessment, the presence of a dose response relationship clearly strengthens the notion of a causal relationship. The absence of a clear dose response, however, does not necessarily weaken the case for a causal relationship. Problems in risk assessment were discussed. Epidemiology can usually determine whether a given disease is caused by a certain factor; however, it has difficulty in determining whether there are adverse health effects associated with a specific exposure. Epidemiological studies usually have considerable difficulty in detecting very low levels of risk in human populations. It is noted that since a risk free society is not attainable, epidemiological methods must be refined in order to best provide the scientific rationale for formulating sound policy and regulatory decisions in handling potential hazards.
Biostatistics; Risk-analysis; Dose-response; Risk-factors; Regulations; Mortality-rates; Health-protection; Health-hazards;
Infectious Diseases; Disease and Injury;
Proceedings of a Symposium on Epidemiology and Health Risk Assessment, Columbia, Maryland, May 14-16, 1985, Centers for Disease Control/NIOSH, 22 pages
Page last reviewed: April 12, 2019
Content source: National Institute for Occupational Safety and Health Education and Information Division