Oral Presentation Sydney Spinal Symposium 2026

Investigation of heterogeneous treatment effects of oxycodone in acute back and neck pain: an individualised treatment effect approach (141789)

Hanan McLachlan 1 , Christine Lin 1 , Christopher Maher 1 , Peter Youssef 1 2 , Sarah Cummings 3 , Rob Herbert 4 , Nicola Jones 5 , Richard O Day 6 7 , Mark Hancock 8 , Giovanni Ferreira 1
  1. Institute for Musculoskeletal Health, Sydney Local Health District, and Faculty of Medicine and Health, University of Sydney, Sydney, NSW, Australia
  2. Institute for Rheumatology and Orthopaedics, Royal Prince Alfred Hospital, Sydney, NSW, Australia
  3. Sarah Cummings Consulting, , Sydney, NSW, Australia
  4. Geebung Statistical Consulting, Blackheath, NSW, Australia
  5. Sydney Musculoskeletal Health, Faculty of Medicine and Health, University of Sydney, Sydney, NSW, Australia
  6. Clinical Pharmacology and Toxicology,, St Vincent’s Hospital Sydney, Darlinghurst, , Sydney, NSW, Australia
  7. St Vincent’s Clinical Campus, University of New South Wales, Sydney, NSW, Australia
  8. Department of Health Sciences, Faculty of Medicine, Health and Human Sciences, Macquarie University , Sydney, NSW, Australia

Background 

Despite limited evidence of benefit, opioids remain widely prescribed for acute back pain. In the OPAL randomised placebo-controlled trial, oxycodone was, on average, no more effective than placebo for acute back and neck pain. Estimating individualised treatment effects may uncover whether the average effect is masking important differences between those with the best and worst responses to oxycodone. This exploratory analysis examines whether treatment effects of oxycodone varied across individuals in OPAL and, if so, aims to identify those most likely to benefit, supporting more targeted prescribing.

Methods

Data from 346 participants in the OPAL trial were analysed for three outcomes: pain intensity and pain interference at week 6 and pain intensity at 3 days, all reported on a 0-10 rating scale. Causal forests were constructed for each outcome and conditioned simultaneously on baseline clinical (e.g. pain intensity) and sociodemographic (e.g. age, sex) characteristics. Individualised estimates of treatment effects were obtained for each participant. The presence of between-participant heterogeneity in treatment effect was tested using Chernozhukov’s test.

Results

There was no evidence of heterogeneity in the individualised treatment effect of oxycodone on the outcomes of pain intensity at 6 weeks (p=0·16), pain interference at 6 weeks (p=0·45), or pain intensity at day 3 (p=0·017, but with a poorly calibrated model). For almost all participants, exposure to oxycodone was predicted to result in worse pain intensity and pain interference outcomes at six weeks, and a small, less than 1 point out of 10, reduction in pain intensity at three days.

Conclusion

Based on common clinical and sociodemographic characteristics measured at baseline, we found no evidence of effect heterogeneity. We did not identify individuals likely to benefit from short-term oral oxycodone for acute low back or neck pain.