Oral Epidemiology by Unknown

Author:Unknown
Language: eng
Format: epub
ISBN: 9783030501235
Publisher: Springer International Publishing

Returning to earlier examples, Dr. Kahnyne (of the 74 possible covariables) designed an analysis plan that was systematic and reproducible but also not informed by subject matter expertise. Relying on p-values from a preliminary round of pair-wise hypothesis testing as the only arbiter of what variables were maintained in the multivariable model, Dr. Kahnyne marched forward without consideration of whether included variables were confounders, colliders, or mediators.

◘ Figure 16.7 shows one plausible interpretation Dr. Denton’s investigation into whether the security of fluoride coverage encourages reckless junk food consumption. The hypothesis to test is shown as the direct path from fluoride to sugar intake. Reasonable assumptions are that fluoride and sugar are ancestors of caries (whether protective or risk factors, the paths are drawn identically). Therefore, caries status is a collider; including it in the regression model would affect the observed association between fluoride and sugar intake, biasing it away from the underlying causal relationship Dr. Denton had hoped to measure.

Fig. 16.7The hypothetical Dr. Denton adjusted her estimate of the association between fluoride varnish (FV) and sugar consumption (S) for caries status and socioeconomic position (SEP). Based on this diagram, caries is a collider, and conditioning on it will bias the association between fluoride and sugar away from the underlying causal effect

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.