Can cost utility evaluations inform decision making about interventions for low back pain?

BACKGROUND CONTEXT: Low back pain (LBP) is associated with high health-care utilization and lost productivity. Numerous interventions are routinely used, although few are supported by strong evidence. Cost utility analyses (CUAs) may be helpful to inform decision makers.

PURPOSE: To conduct a systematic review of CUAs of interventions for LBP.

STUDY DESIGN: Systematic review.

METHODS: A search strategy combining medical subject headings and free text related to LBP and health economic evaluations was executed in MEDLINE. Cost utility analyses combined with randomized controlled trials for LBP were included. Studies that were published before 1998, non-English, decision analyses, and duplicate reports were excluded. Search results were evaluated by two reviewers, who extracted data independently related to clinical study design, economic study design, direct cost components, utility results, cost results, and CUA results.

RESULTS: The search produced 319 citations, and of these 15 met eligibility criteria. Most were from the United Kingdom (n=8), published in the past 3 years (n=12), studied chronic LBP or radiculopathy (n=13), and had a follow-up >12 months (n=13). Combined, there were 33 study groups who received a mean 2.1 interventions, most commonly education (n=17), exercise therapy (n=13), spinal manipulation therapy (n=7), surgery (n=7), and usual care from a general practitioner (n=7). Mean baseline utility was 0.57, improving to 0.67 at follow-up; the mean difference in utility improvement between study groups was 0.04. Based on available data and converted to US dollars, the cost per quality-adjusted life year ranged from $304 to 579,527 dollars, with a median of 13,015 dollars.

CONCLUSIONS: Few CUAs were identified for LBP, and there was heterogeneity in the interventions compared, direct cost components measured, indirect costs, other methods, and results. Reporting quality was mixed. Currently published CUAs do not provide sufficient information to assist decision makers. Future CUAs should attempt to measure all known direct cost components relevant to LBP, estimate indirect costs such as lost productivity, have a follow-up period sufficient to capture meaningful changes, and clearly report methods and results to facilitate interpretation and comparison.