Background The recent OPAL trial found that an oxycodone/naloxone combination was not effective at relieving acute back and neck pain compared to placebo. However, just over half of the participants were compliant with the treatment regimen, defined as taking a minimum 80% of the prescribed dose over the 6-week treatment period. The study aims to estimate the treatment efficacy in people who complied with the regimen, using a complier average causal effect analysis.
Methods Complier average causal effect analysis was a pre-specific in statistical analysis plan and it was performed using (i) a propensity-weighted approach and (ii) a joint modelling approach, with compliance defined using different cut-off points (taking ≥20%, ≥40%, ≥60% and ≥80% of the prescribed dose). The primary outcome was pain severity at week 6. A secondary outcome was pain severity at weeks 2, 4, 12, 26 and 52 weeks. A per-protocol analysis was also conducted at all time points.
Results After propensity weighting, a small but significant difference favouring the placebo group for pain severity was found at week 6 (mean difference 0.78, 95% CI 0.22-1.33, p=0.006). Similarly, pain outcomes at other timepoints, and sensitivity analysis using the joint modelling approach or different cut-offs of compliance, either favoured the placebo group or showed no difference between groups.
Conclusion This secondary analysis showed results consistent with the OPAL trial main analysis. Even when complying with the treatment regimen, the opioid was no more effective than placebo for patients with acute back and/or neck pain.