Unstandardised regression coefficients for pathways in mediation analyses where regression coefficients a, b, c, c' reflect the pathways described in Figure 1.

The independent variable is baseline pain intensity and the dependent variable is MFI at 6 months follow up for each model.*Denotes coefficients which are statistically significant at the p>0.05 level. Mediation effect represents the point estimate of the indirect effect of the independent variable on the dependent variable acting through the mediator calculated through bias-corrected accelerated bootstrapping. Mediation effect is significant where the 95% CI do not span 0.