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

Abstract

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, with a median of $13,015.

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.

Introduction

A recent study in the United States estimated that annual direct medical costs for spine problems—including low back pain (LBP)—had doubled from $52 to $102 billion in only 7 years, whereas the proportion of persons with spine problems reporting limitations in activities of daily living rose nearly 20% for the same period [1]. These findings highlight the challenges faced by those involved in the management of LBP, including general practitioners (GP), specialist physicians from a variety of disciplines, allied health providers, complementary and alternative medicine (CAM) providers, as well as patients, third-party payers, and policy makers. Every day, stakeholders are faced with selecting from among the hundreds of available interventions, such as medication, education, physical therapy, manual therapy, exercise therapy, cognitive behavioral therapy, injections, surgery, and CAM [2]. Although all claim to improve LBP, they often lack strong evidence to support their efficacy and safety [2].

When faced with such difficult choices, the consideration of cost-effectiveness could perhaps be used to guide these decisions. Some have even suggested that cost-effectiveness be given priority over clinical effectiveness when confronted with numerous comparable interventions [3]. Cost utility analysis (CUA) examines cost-effectiveness using health outcomes valued using scores that typically range from 0 (worst health) to 1 (best health). Depending on how they are obtained, these scores can be viewed as “utility,” which measures one's preference for certain states of health. Utility is often estimated from instruments such as the short form 36 (SF-36) or the EuroQOL (EQ-5D) [4]. By combining utility with survival it is possible to calculate quality-adjusted life year (QALY), a popular outcome for CUAs [5].

For example, the health of a patient who survives 2 years with severe LBP at a utility of 0.5 would be rated at 1.0 QALYs. If an intervention were to increase utility to 0.75 over that period, the patient would gain 0.5 QALYs. By estimating the cost implications of that intervention (eg, physician charges, facility fees, supplies), it would be possible to estimate the incremental cost per QALY gained. If the intervention cited in the example had net costs of $10,000, its incremental cost utility would be $20,000 per QALY. Although the desirable economic efficiency of interventions remains undefined, government agencies have operationalized a threshold [6]. In the United Kingdom, for instance, the National Institute for Clinical Excellence has decided that interventions costing more than 30,000 pounds (∼50,000 US dollars [USD]) per QALY are not cost effective [7]. In the United States, no such formal threshold exists and there have been calls to avoid using one; values as high as $100,000 per QALY have been discussed [6].

By measuring utility and costs in a randomized controlled trial (RCT), it is possible to compare the efficiency of two or more interventions. When one intervention generates lower net costs and provides more QALYs, it is said to “dominate” the other. If it not only generates higher net costs but also provides more QALYs, then the incremental cost-effectiveness ratio (ICER) expresses its efficiency relative to the other intervention. The ICER is defined as the difference in costs between two interventions divided by the difference in the value of health gains between those interventions [8]. Because there is often uncertainty about cost estimates and utility scores, researchers will also report results as the cumulative frequency with which various analyses using different values result in an ICER below a particular value, often termed the willingness-to-pay threshold [9].

The primary objective of this study was to summarize the methods and results of CUAs based on clinical trials of interventions for LBP to determine whether this information could help inform decision makers about which treatment approach is preferred from a cost-effectiveness viewpoint. Secondary objectives were to outline important methodological and reporting issues pertinent to interpreting results of CUAs and to identify the areas of priority research for future CUAs.