A back-pain review finds value in some care, with local costs still in the picture

A health service weighing back-pain care faces more than a question of what an intervention costs. It also needs to know what patients gain, what care it replaces and whose costs are counted. A new systematic review finds that many non-drug, non-surgical approaches offered good economic value compared with usual care in the settings studied, but its conclusions do not translate into a universal savings figure or a treatment ranking.

Stephanie Mathieson and colleagues examined economic evaluations of guideline-supported interventions for adults with non-specific low back pain, with or without leg pain. Their review in PLOS Medicine, published October 27, draws entirely on economic evidence from high-income countries. It is a review of value, not a new trial or advice for an individual patient.

The authors report searching ten electronic databases from their inception through February 9, 2026, alongside manual searches and contact with content experts. Eligible evaluations used randomized-trial data or economic models and concerned interventions supported by at least one selected World Health Organization, NICE or North American Spine Society guideline. That selection did not require agreement across all three guidelines. Two reviewers independently screened material, extracted data and assessed methods.

The review excluded sciatica-only and specified pathological populations. It accepted comparisons with usual care, sham or placebo, or no treatment; usual care itself varied. Its primary measure was the additional cost per quality-adjusted life year, a measure combining length and quality of life. The authors also considered direct and indirect costs, reported original currencies and standardized costs to 2024 British pounds using inflation and purchasing-power adjustments. They separated results by economic perspective, because healthcare, payer and societal accounts do not count the same costs.

Some interventions appeared potentially cost-saving in particular comparisons, the review reports. Findings for massage and acupuncture were inconsistent, and results varied with context, delivery and perspective. The authors summarized evidence qualitatively: differences in populations, interventions, time horizons, perspectives and country costs, along with missing measures of economic-outcome variance, made pooling inappropriate. Inadequate split-sample data also prevented planned subgroup and sensitivity analyses. The relevance to low- and middle-income health systems remains uncertain.

The team used CHEC to appraise trial-based evaluations and the Philips checklist for models, and reported uneven methodological reporting. It omitted a planned GRADE assessment, saying available frameworks did not fit the economic evidence. The paper declares no specific funding for the work and lists authors' professional, grant, advisory and commercial relationships.

Rowan Finch analysis: Value for money is not the same as cash returned to a budget. A local commissioner would need to consider its own costs, existing care and time horizon before applying these findings. Nor can comparisons against different baselines rank every therapy against every other one; an individual care decision asks a different question.

This is the newsroom's original synthesis and analysis of Mathieson and colleagues' article, which is published under CC BY 4.0. The authors and PLOS Medicine do not endorse this synthesis.