What is it about?
Imagine that a member of your epidemiologic team approaches you and says: “I’m stumped. I think that 3 of the foods caused the outbreak, but their risk ratios are low and not statistically significant.” You are also seeing a similar problem in reports from your team on other investigations. This problem, called a shifting reference group, occurs when each exposure is compared to different combinations of other exposures. These include other causative exposures. To resolve this problem, we revisited an approach that compares each suspected individual exposure and exposure combination to the same common reference group.
Featured Image
Photo by Campaign Creators on Unsplash
Why is it important?
The common reference group has broad utility and versatility. Yet, it is infrequently used in epidemiologic practice. It provides effect measures of individual exposures and joint exposures that are easily interpreted and compared. It also corrects for confounding risk factors. Statistical testing uses basic tests for 2 x 2 contingency tables. Because it is also easy to learn and understand, epidemiologic units from the local to the national level should all find it applicable to their missions.
Perspectives
In over 35 years of mentoring epidemiologists in field investigations in multiple countries, I have repeatedly needed to explain how analyses have gone off track, and how to fix things with a common reference group. The solution has been to include this approach in the initial training. It takes about 30 minutes to explain the concept and 30 minutes more of practice on how to construct a common reference group from raw data.
Robert Fontaine
Centers for Disease Control and Prevention (retired)
Read the Original
This page is a summary of: Untangling the effects of multiple exposures with a common reference group in an epidemiologic study: A practical revisit, PLoS ONE, December 2023, PLOS,
DOI: 10.1371/journal.pone.0295915.
You can read the full text:
Contributors
The following have contributed to this page
