Abstract
We assessed the causal mechanisms of a healthy lifestyle intervention for patients with chronic low back pain and knee osteoarthritis (OA), who are overweight or obese. We conducted causal mediation analyses of aggregated data from two RCTs; which included 160 patients with chronic low back pain, and 120 patients with knee OA. Participants were randomised via one central randomisation schedule, to the intervention, or usual care. The intervention consisted of brief advice and referral to a 6-month telephone-based healthy lifestyle coaching service. Participants in the back pain trial were also offered a single physiotherapy consultation. The hypothesised primary mediator was self-reported weight, and alternative mediators were diet, physical activity, and pain beliefs. Outcomes were pain intensity, disability, and quality of life (QoL). Data were analysed using causal mediation analyses with sensitivity analyses for sequential ignorability. All mediation models were specified a priori. The intervention had no effect on pain intensity, disability or physical QoL. The intervention significantly improved mental QoL, however, the intervention effect was not channelled via the selected mediators. The intervention did not reduce weight, or the alternative mediators (diet, physical activity, pain beliefs), and these mediators were not associated with the outcomes (with one exception; poor diet was associated with lower mental QoL). The sensitivity analyses showed that our estimates were stable across all possible levels of residual confounding. Our findings show that the intervention did not cause a meaningful change in the hypothesised mediators, and these mediators were not associated with patient outcomes.