Dose-response is a fundamental part of risk assessment, whether it’s for a microbe, chemical, or even a food. But these models rely on a range of assumptions. When you put them together with observational data that might be confounded, an accurate estimate might not be the result especially for low exposures when the data comes from high exposures.

In our newly-published study, we quantified the biases resulting from model selection and inclusion of high consumers on the dose-response, using the association between low consumption of red meat and colorectal cancer as an example.

Article published in Advances in Nutrition:

Observational Dose-Response Meta-Analysis Methods May Bias Risk Estimates at Low Consumption Levels: The Case of Meat and Colorectal Cancer