Mediators of Treatment Effect in the Back In Action Trial:Using Latent Growth Modeling to Take Change Over Time Into Account

Abstract

Objectives: To test whether change in fear-avoidance beliefs was a mediator of the effect of treatment on disability outcome, and to test an analytical approach, latent growth modeling, not often applied to mediation analysis. Methods: Secondary analysis was carried out on a randomized controlled trial designed to compare an intervention addressing fear-avoidance beliefs (n=119) with treatment as usual (n=121) for patients with low back pain, which found the intervention to be effective. Latent growth modelling was used to perform a mediation analysis on the trial data to assess the role of change in fear-avoidance beliefs on disability outcome. The product of coefficients with bias-corrected bootstrapped confidence intervals was used to calculate the mediating effect. Results: A statistically significant mediating effect of fear-avoidance beliefs on the effect of treatment on disability outcome was found (standardized indirect effect −0.35; bias-corrected 95% CI, −0.47 to −0.24). Poor fit of the model to the data suggested that other factors not accounted for in this model are likely to be part of the same mediating pathway. Discussion: Fear-avoidance beliefs were found to mediate the effect of treatment on disability outcome. Measurement of all potential mediator variables in future studies would help to more strongly identify which factors explain observed treatment effects. Latent growth modelling was found to be a useful technique to apply to studies of treatment mediation, suggesting that future studies could use this approach.