Related Experiment Videos
Relative risk and excess rate models in exposure-response analysis
1Department of Occupational and Environmental Medicine, Umeå University, Sweden. bengt.jarvholm@envmed.umu.se
American Journal of Industrial Medicine
|April 1, 1997
Summary
Standardized mortality ratios (SMR) can bias dose-response findings in environmental exposure studies. Analyzing both relative and excess risk models is recommended for accurate exposure-disease association assessments.
Area of Science:
- Epidemiology
- Environmental Health
- Biostatistics
Background:
- Environmental exposure dose and age are often correlated.
- Standardized mortality ratio (SMR) models may present biased dose-response relationships.
- Rapidly increasing incidence rates with age can exacerbate SMR model bias.
Purpose of the Study:
- To highlight potential biases in dose-response analyses using SMRs.
- To emphasize the importance of model choice in exposure-disease association studies.
- To advocate for the use of both relative and excess risk models.
Main Methods:
- Review of cohort studies analyzing dose-response relationships.
- Comparison of relative risk models (SMR) versus excess risk models.
- Assessment of bias introduced by age-related incidence rate changes.
Main Results:
- SMR-based relative risk models can yield biased or contradictory dose-response results.
- Bias is pronounced when the reference population's incidence rate increases sharply with age.
- A lack of awareness regarding SMR bias in exposure-disease association studies was observed.
Conclusions:
- Model selection for dose-response analysis should prioritize biological plausibility.
- Analyzing data with both relative and excess risk models is often advisable.
- Awareness and appropriate application of risk models are crucial for accurate epidemiological research.