|
Abstract:
|
Prevalence odds ratios (PORs) and prevalence ratios (PRs) are two effect measures used in epidemiologic studies of cross-sectional data to describe the magnitude of an association. POR approximates PR when the outcome or condition is rare. For a common condition, the measures may be discrepant, as POR is typically inflated compared to PR. This study estimates POR and PR for the outcome of meeting the aerobic physical activity guideline among U.S. adults aged ?18 years with mobility disability using cross-sectional data from the 2015 Behavioral Risk Factor Surveillance System (n=32126). SAS-Callable SUDAAN was used to conduct multiple logistic regression to estimate PORs and PRs, adjusted for sex, age group, race/ethnicity, education level, arthritis, and body mass index. Prevalence of meeting the guideline among this population is 30.3%, and ranges from 25%-37% among subgroups. Results show that the difference between each covariate's POR and PR ranges from 0.01 - 0.28 (absolute scale) and 1%-10% (relative scale). This information will help researchers understand the difference between POR and PR and the unique considerations when interpreting results derived from each measure.
|
