Skip to main content
. 2014 Nov 5;24(6):474–483. doi: 10.2188/jea.JE20140025

Figure 1. Associations between moderate-to-vigorous physical activity and the prevalence of chronic low back pain and chronic knee pain among Japanese adults (n = 4559). Solid lines represent prevalence ratios (PRs), and dashed lines indicate 95% confidence intervals estimated by Poisson regression, estimated by a quadratic function of physical activity levels (metabolic equivalent of task [MET]-hours/week). Panels on the left (A) display PR adjusted for sex, age, education years, self-rated health, depressive symptoms, smoking habit, chronic disease history, and body mass index; while on the right (B), PRs are further adjusted for history of joint injuries, medication use, and consultation with physicians for pain management. The reference value for each was fixed to the values giving the lowest prevalence of each outcome. P for each quadratic function is displayed.

Figure 1.