Fig. 4

This shows the relationships of TV time with bodily pain severity and potential moderation of T2D status. (A) The bodily pain prediction margins of T2D status with 95% confidence intervals for the unadjusted and fully adjusted models. (B) The marginal effects of prediabetes and T2D (in reference to NGM) on bodily pain severity at different TV time thresholds for the unadjusted and fully adjusted models. The solid lines indicate the marginal effects of changes in bodily pain severity with changing TV time. The dotted lines are the confidence intervals around the lines, which determine the threshold of TV time that has a statistically significant effect on bodily pain severity in those with prediabetes (ORANGE) and T2D (RED). They are statistically significant whenever the lower and upper limits of the confidence intervals are both below or above the zero (0 - BLUE) lines. Note: NGM was set as the reference point in the regression model.