Advertisement

Abstract

Background The delivery of acupuncture alongside mainstream interventions and the cost-effectiveness of “alternative” treatments remain areas of controversy.

Objective The aim of this study was to assess the cost-utility of adding acupuncture to a course of advice and exercise delivered by UK National Health Service (NHS) physical therapists to people with osteoarthritis of the knee.

Design A cost-utility analysis was performed alongside a randomized controlled trial.

Methods A total of 352 adults (aged 50 years or older) were randomly assigned to receive 1 of 3 interventions. The primary analysis focused on participants receiving advice and exercise (AE) or advice and exercise plus true acupuncture (AE+TA). A secondary analysis considered participants receiving advice and exercise plus nonpenetrating acupuncture (AE+NPA). The main outcome measures were quality-adjusted life years (QALYs), measured by the EQ-5D, and UK NHS costs. Results were expressed as the incremental cost per QALY gained over 12 months. Sensitivity analyses included a broader cost perspective to incorporate private out-of-pocket costs.

Results NHS costs were higher for AE+TA (£314 [British pounds sterling]) than for AE alone (£229), and the difference in mean QALYs favored AE+TA (mean difference=0.022). The base-case cost per QALY gained was £3,889; this value was associated with a 77% probability that AE+TA would be more cost-effective than AE at a threshold of £20,000 per QALY. Cost-utility data for AE+NPA provided cost-effectiveness estimates similar to those for AE+TA.

Limitations As with all trial-based economic evaluations, caution should be exercised when generalizing results beyond the study perspectives.

Conclusions A package of AE+TA delivered by NHS physical therapists provided a cost-effective use of health care resources despite an associated increase in costs. However, the economic benefits could not be attributed to the penetrating nature of conventional acupuncture; therefore, further research regarding the mechanisms of acupuncture is needed. An analysis of alternative cost perspectives suggested that the results are generalizable to other health care settings.

Like many common disabling conditions, osteoarthritis imposes a sizeable economic burden on society, with cost-of-illness estimates accounting for 1% to 2.5% of the gross national product in several developed countries.1 In the United Kingdom, the direct cost to the National Health Service (NHS) of treating arthritis and related conditions has been estimated to be £1.1 billion (British pounds sterling; 1998–1999 prices).2 More broadly, societal costs associated with osteoarthritis alone, because of lost production as a result of work absenteeism, have been estimated to be £3.2 billion (1999–2000 prices).3 Evidence from the United States has shown that the aggregate annual cost of work absenteeism because of osteoarthritis exceeds $10 billion.4

It has previously been shown that patients generally have an aversion to drug therapies,5 and the long-term use of oral nonsteroidal inflammatory drugs is discouraged.6 Preferences for nonpharmacological treatment options have resulted in the increased promotion7 and provision8 of complementary medicine, with acupuncture being one of the more popular options. For example, in May 2009, the UK National Institute for Health and Clinical Excellence recommended acupuncture as an appropriate treatment option for patients with persistent nonspecific low back pain.9 Although the concept of integrating conventional and complementary health care approaches has been viewed positively,10 the delivery of acupuncture and the cost-effectiveness of alternative strategies remain areas of controversy. Previous trials have been criticized for small sample sizes, inadequate masking, inadequate follow-up, and lack of a credible sham intervention.1114 Recent clinical trials have made some progress in addressing these shortcomings.1520 Systematic reviews have concluded that acupuncture is more effective than a placebo for knee osteoarthritis, although a wide range of placebo or sham acupuncture interventions have been evaluated, with various degrees of credibility.13,14

In the United Kingdom, guidelines for osteoarthritis have emphasized that the literature on health economics is limited and of questionable generalizability to the UK NHS or other health care systems,3 and the specific information provided about acupuncture in the guidelines of the UK National Institute for Health and Clinical Excellence has been subject to academic debate.21,22 Current national and international recommendations for the management of osteoarthritis emphasize the core treatments of advice and education to enhance understanding of the condition, plus activity and exercise.23 Acupuncture has been shown to be a cost-effective treatment option for low back pain (in the United Kingdom and Germany),24,25 osteoarthritis pain26 (in Germany), and chronic neck pain (in Germany),27 but no cost-effectiveness evidence currently exists for acupuncture treatment of knee osteoarthritis in the United Kingdom. We previously investigated whether acupuncture is a clinically useful adjunct to a course of sessions of advice and exercise, delivered by NHS physical therapists, for knee osteoarthritis in older adults.28,29 In this article, we address different research questions relevant to consideration of the economic consequences of a new treatment, reporting the results of a cost-utility analysis performed alongside a clinical trial.

Method

Clinical Trial

Comprehensive details about the clinical trial are described in detail elsewhere.28,29 Participants were adults aged 50 years or older who had been referred to 1 of 37 NHS physical therapy centers with a clinical diagnosis of knee osteoarthritis, and who were recruited between November 2003 and October 2005. Of 1,061 potentially eligible participants, 352 were randomly allocated to receive advice and exercise (AE; n=116), advice and exercise plus true acupuncture (AE+TA; n=117), or advice and exercise plus nonpenetrating acupuncture (AE+NPA; n=119). None of the participants had previously received acupuncture. The primary clinical outcome was change at 6 months in the score on the pain subscale of the Western Ontario and McMaster Universities Osteoarthritis Index.30 All researchers involved in the collection and analysis of data were unaware of the treatment allocation.

Interventions

The AE package was provided over 6 sessions and consisted of a written advice leaflet modeled on the Arthritis Research UK leaflet on knee osteoarthritis (www.arthritisresearchuk.org.uk) and an individualized exercise program that focused on lower-limb strengthening, stretching, and balance. Exercise was advanced and supervised over the 6 sessions, and every participant had a home exercise program. In the same treatment sessions, participants allocated to the AE+TA group received the AE package plus acupuncture at traditional Chinese acupuncture points. Physical therapists selected between 6 and 10 points from 16 commonly used points, and needles were manipulated in an attempt to achieve needle sensations. Participants allocated to the AE+NPA group received the AE package plus acupuncture with nonpenetrating blunt-tip needles,31 using the same treatment protocol as for the AE+TA group. Treatments were delivered over a 6-week period.

Appropriate Comparators for Economic Evaluation

The primary purpose of an economic evaluation is to inform decision makers about competing claims for health care resources; therefore, such an evaluation often is conducted alongside a pragmatic clinical trial.32 In the present study, the nonpenetrating acupuncture therapy was an exploratory intervention, included for the purpose of investigating the efficacious needling effects of acupuncture. We obtained ethics approval to inform participants that they “may receive acupuncture, using 1 of 2 different types of acupuncture needle”; participants did not consent to a placebo-controlled trial.28 Thus, the trial results relating to the nonpenetrating acupuncture treatment could not be expected to be replicated in routine clinical practice, in which people would be aware of the nature of the needles. Despite the possibility of “acupressure” benefits induced through nonpenetrating needles, our nonpenetrating acupuncture protocol is not implementable in practice and, therefore, cannot be used to make any claims for NHS health care resources. Accordingly, the appropriate primary research goal of the economic evaluation was to investigate whether true acupuncture is a cost-effective adjunct to advice and exercise.

Health Outcomes

Utility was measured by the EQ-5D questionnaire at baseline, 6 weeks, 6 months, and 12 months.33 Using a scoring algorithm based on health status valuations elicited from a large representative sample of the UK population,34 this generic health status measure provides utility values for every possible response to the 5-dimension questionnaire. Quality-adjusted life years (QALYs) were calculated by use of area-under-the-curve analysis with linear interpolation, in which periods of time spent in the study were weighted by the quality-adjusted factor obtained from the EQ-5D. In the present study, the maximum number of QALYs per participant was 1, which is equivalent to 12 months spent in full health.

Health Care Resource Use and Cost Estimation

The estimation of health care costs involved the collection of resource use data and the assignment of appropriate standardized unit costs. The number and content of treatment sessions attended by participants were obtained from standard forms completed by the study therapists. Additional knee osteoarthritis–related resource use data were collected from self-report postal questionnaires. At each follow-up, participants were asked to recall the time period “since your last questionnaire,” and the total cost estimate resulted from the summation of data from the distinct time periods. The questionnaires collected data on consultations with primary care–based health care practitioners, hospital consultants (outpatient attendance), or any other health care provider within the UK NHS or private practice. Participants were also asked to report prescribed medications and over-the-counter purchases.

Excluded from the study were people with potentially serious pathology and those who were already on a surgical waiting list for total knee replacement.28 Given the lengthy waiting list for orthopedic consultations in the United Kingdom at the time of the trial, it was unlikely that participants would have proceeded to receive hospital-based treatment during the follow-up period; therefore, no details related to hospital-based procedures or inpatient stays were collected.

In the base-case analysis, costs were measured from an NHS perspective in accordance with UK guidelines.35 Table 1 shows a list of all unit costs used in the analysis (NHS care and private care), expressed in British pounds sterling at 2004–2005 prices. NHS care was costed as standardized national averages, using data obtained from the Personal Social Services Research Unit,36 NHS Reference Costs,37 and The British National Formulary.38 Physical therapists are the largest group of health care professionals providing acupuncture in the NHS. Therefore, all acupuncture sessions were assigned a unit cost associated with NHS community physical therapy. The cost of traditional acupuncture needles is negligible and was not accounted for in the analysis.39 Costing assumptions for the study interventions (AE, AE+TA, and AE+NPA) and consultations with other health care professionals were based on average session times for the study interventions and clinical judgment of current practice.

Table 1.

Unit Costs of Health Care Resources Used in the Economic Evaluationa

Statistical Analysis

Analysis was performed according to the intention-to-treat principle. Multiple imputation was used to analyze incomplete data.40 Missing EQ-5D scores were imputed at each time point. Cost estimates were categorized into distinct components to maximize real data reported in the questionnaires; only missing components were imputed. We generated 5 possible values for each missing value by using multiple linear regression models containing age, sex, and treatment group as covariates.41 The final imputed value was the arithmetic mean of the 5 created values, and reported standard deviations and confidence intervals were adjusted to account for the additional “between-imputation” variance.

For the estimation of cost-utility, analysis focused on the joint estimation of cost and effect differences. The primary unit of analysis was the incremental cost per QALY, calculated as the incremental cost divided by the incremental QALY. Uncertainty was analyzed by using 5,000 bootstrapped replications of mean cost and QALY differences, plotting the data pairs on a cost-utility plane, where incremental QALYs were represented along the x-axis and incremental costs were represented along the y-axis.42 We performed a probabilistic threshold analysis, which involved generating a cost-utility acceptability curve to estimate the probability that AE+TA was cost-effective, compared with AE alone, over a range of cost-per-QALY willingness-to-pay values.43

To supplement the primary economic analysis, we calculated confidence intervals around disaggregated incremental cost, EQ-5D, and QALY estimates by using conventional parametric methods and bias-corrected and accelerated bootstrapping (1,000 replications). Intervals obtained with these 2 approaches were compared, when appropriate, to assess whether the parametric methods were robust to the level of skewness in the data.44 Baseline utility imbalances between treatment groups were controlled by using a multiple regression-based adjustment.45 Given the 12-month follow-up period of the study, discounting was not necessary. All statistical analyses were performed with SPSS for Windows (version 15.0)* and Stata (version 9.0).

Sensitivity Analyses

To further examine the robustness and interpretation of the study findings, we performed the following 3 additional sensitivity analyses: a complete-case analysis to explore the implications of missing data, a cost perspective that incorporated non-NHS health care resource use, and an analysis of the AE+NPA group within the base-case framework. A price premium was used to explore variations in the cost of non-NHS health care resources; this process involved simultaneously multiplying each non-NHS unit cost by a factor ranging from 1 (NHS and non-NHS costs are considered to be equal) to 3 (non-NHS costs are considered to be 3 times the unit cost of the NHS equivalent).

Role of the Funding Source

This article reports the results of a cost-utility analysis performed alongside a clinical trial. The clinical trial was supported by Project Grant H0640 from Arthritis Research UK and Support for Science funding secured by the North Staffordshire Primary Care Research Consortium for NHS service support costs. Professor Foster was funded by a Primary Care Career Scientist Award from the Department of Health and NHS research and development. The sponsors had no role in any aspect of the study, including the writing of the article. The authors had full access to all of the data and had final responsibility for the decision to submit the article for publication.

Results

Results from the clinical trial were reported previously.29 In brief, the addition of either acupuncture intervention (AE+TA or AE+NPA) did not provide a significant improvement in pain scores on the Western Ontario and McMaster Universities Osteoarthritis Index at 6 months, compared with a course of sessions of AE alone. Small benefits in favor of the addition of acupuncture were shown for pain intensity and unpleasantness, and these effects were greater and were sustained for longer in the AE+NPA group.

Health Care Costs and Health Outcomes

Complete resource use data and EQ-5D scores were available for 175 participants (75.1%) and 184 participants (79.0%), respectively; complete data for both were available for 169 participants (72.5%). Nonresponders were more likely to be men (male nonresponders=42.2%; male responders=33.1%). Of the 58 participants with incomplete resource use data, 21 (36.2%) provided sufficient data for at least 1 self-report cost category to be costed over the 12-month follow-up period. There were no differences in response rates among treatment groups.

Disaggregated details of mean resource use and mean costs for the sample of respondents for whom complete resource use data were available are shown in Tables 2 and 3; total cost estimates for the imputed data analysis are also shown. With regard to total NHS-related costs, AE+TA was associated with a higher mean cost both before and after multiple imputation (Tab. 3). The key driver for the total cost estimate was the number of study treatment sessions attended, which was greater for the AE+TA group (difference in means of 1.89; 95% confidence interval=1.54–2.24) (Tab. 2), resulting in a higher treatment session cost (Tab. 3).

Table 2.

Health Care Resource Use per Participant Over 12 Months, by Treatment Groupa

Table 3.

Health Care Costs (British Pounds Sterling [£]) per Participant Over 12 Months, by Treatment Groupa

Mean EQ-5D scores and mean QALYs are shown in Table 4; EQ-5D scores are the observed results (without controlling for baseline utility imbalances), and estimates of QALYs are reported after controlling for baseline utility imbalances. The mean QALY over 12 months was higher for the AE+TA group than for the AE group after controlling for baseline utility levels. Confidence intervals derived from bias-corrected and accelerated bootstrapping provided similar results (not reported), indicating that conventional parametric methods were robust to the level of skewness in the cost and EQ-5D data.

Table 4.

Mean (SD) EQ-5D Scores and Quality-Adjusted Life Year (QALY) Estimates Over 12 Months, by Treatment Groupa

Cost-Utility Analysis

Point estimates from the base-case analysis showed that AE+TA was more effective and more expensive than AE alone, resulting in an incremental cost per QALY ratio of £3,889 (£84.81 divided by 0.022 QALY). Figure 1 shows that AE+TA was more effective and more costly than AE in the majority of the bootstrapped replications (83% of the 5,000 replications were represented in the northeast quadrant). The use of a cost-utility acceptability curve for consideration of the willingness to pay for additional QALYs showed that AE+TA would be associated with a 77% chance of being more cost-effective than AE if society were willing to pay at least £20,000 per additional QALY (Fig. 2).

Figure 1.

Base-case cost-utility plane comparing advice and exercise plus true acupuncture (AE+TA) with advice and exercise (AE) alone. Data were based on 5,000 bootstrapped cost-effect pairs. Incremental costs and incremental quality-adjusted life years were calculated as AE+TA minus AE.

Figure 2.

Base-case cost-utility acceptability curve comparing advice and exercise plus true acupuncture (AE+TA) with advice and exercise (AE) alone. Data were based on 5,000 bootstrapped cost-effect pairs. QALY=quality-adjusted life year.

Sensitivity Analyses

All further analyses were undertaken with the sample of participants for whom complete resource use and EQ-5D data were available over the 12-month follow-up period and yielded similar results. In all scenarios, the mean cost savings attributable to the AE package were statistically significant, estimates of QALYs were in favor of the group receiving adjunctive acupuncture (either true or nonpenetrating), and the incremental cost per QALY values were within acceptable levels. The largest incremental ratio was in the base-case analysis (£3,889 per QALY), whereas the smallest incremental ratio was associated with the complete-case analysis (£2,278 per QALY). Varying the unit cost of non-NHS health care had little effect on the results.

Under the base-case framework, there were no significant differences between the 2 acupuncture treatment groups. However, AE+NPA was associated with an additional cost of £36 (95% confidence interval=−£21 to £94) and an incremental QALY of 0.001 (95% confidence interval=−0.054 to 0.056) when compared with AE+TA.

Discussion

A treatment package of AE+TA for knee osteoarthritis delivered by NHS physical therapists provides a cost-effective use of health care resources despite an associated increase in costs; that is, the improvements in health-related quality of life warrant the additional resource use. Economic analyses seek to inform decision making, and the widely adopted analytical techniques used within a cost-utility framework reflect the different paradigm compared with clinical trial research. The underlying argument is that policy (resource allocation) decisions should be based on mean costs and effects, the objective being to identify the optimal treatment strategy, irrespective of whether differences are “significant” according to arbitrary rules of statistical inference.46,47 The nature of the analytical framework adopted within an economic evaluation ensures that a preferred treatment will be identified, which is not the case with a conventional clinical trial analysis (eg, the inappropriately named “negative” trial, in which it is not possible to reject a null hypothesis). Cost-effectiveness evidence will not always be sufficient to effect a policy change, but the nature of the analysis identifies the intervention that is more likely to provide greater value for resources spent.

We estimated the incremental cost to be £3,889 (≈$6,300 [US dollars] or €4,500 [Euros] in November 2010) per QALY gained in the base-case analysis; this ratio estimate is well below commonly cited acceptable thresholds for willingness to pay (£20,000/QALY, €25,000/QALY, and $50,000/QALY).4850 Through our secondary analysis of AE+NPA, we found that the economic benefits of adjunctive acupuncture are not attributable to the actual needle penetration. Needle penetration per se is not essential for the effect; rather, it is the patients' perception of conventional acupuncture.51 These findings provide the first UK-based estimates of the cost-effectiveness of adding acupuncture to mainstream interventions for knee osteoarthritis.

Broader Context of Study Findings

In the present study, acupuncture was provided by NHS physical therapists over 6 treatment sessions. A large proportion of physical therapists in the United Kingdom use acupuncture during patient care, and most such therapists are members of the Acupuncture Association of Chartered Physiotherapists (AACP). Of approximately 48,000 physical therapists in the United Kingdom, about 6,000 are members of the AACP, making it one of the largest clinical interest groups of the Chartered Society of Physiotherapy.52 The cost-effectiveness of the delivery of acupuncture for knee osteoarthritis by other health care practitioners, such as physicians or acupuncturists outside the NHS, remains unknown.

There is no cost-effectiveness evidence for acupuncture treatment of osteoarthritis in the United Kingdom, although there are data for patients with other pain conditions, including low back pain and headache.24,53 Despite falling short of arbitrary definitions of statistical significance, the health gains (QALYs) associated with acupuncture therapy in the present study were larger than those reported in these UK studies. Non-UK studies have estimated the cost-effectiveness of acupuncture therapy for knee osteoarthritis, although cross-study comparability was compromised because of different payer perspectives and methodological assumptions.21 A German study investigated the cost-effectiveness of acupuncture and routine care compared with routine care alone for patients with osteoarthritis of the hip and knee.26 Compared with patients receiving routine care alone, patients receiving acupuncture for knee osteoarthritis had an improved quality of life associated with significantly higher costs over a 3-month treatment period (€23,976 per additional QALY). A study in the United States compared traditional acupuncture with education alone for patients with knee osteoarthritis and reported a median cost-effectiveness of $32,000 per QALY gained.54

Despite reporting evidence for the cost-effectiveness of acupuncture therapy for osteoarthritis, these studies either did not have a placebo control group26 or failed to include a placebo control group in their economic evaluation.54 Our inclusion of nonpenetrating acupuncture demonstrated the potential for misleading cost-effectiveness conclusions when assuming an efficacious needling effect. There is potential for a cost-efficient provision of health care, delivered by NHS physical therapists, in addition to advice and exercise, but such benefits are not reliant on acupuncture needles penetrating the skin. Our findings raise important questions for researchers, health care providers, and policy makers considering implementing the provision of acupuncture in addition to physical therapist–led advice and exercise packages.

The findings of the present study are applicable to patients with knee osteoarthritis and, as with all trial-based economic evaluations, caution should be exercised when the results are applied to different health care systems or groups of patients. However, our sensitivity analyses permit issues of generalizability to be considered. In our trial, small differences in health care costs between treatment groups were seen. The results of the threshold analysis showed that acupuncture would remain a cost-effective adjunctive treatment even with considerably larger differences in costs between groups. These findings provide scope for considering the generalizability of results to other health care settings; the actual mean cost estimates for osteoarthritis treatments will differ in various settings, but it is the difference in costs between groups (relative to incremental health benefits) that is important for economic evaluation.

Strengths and Weaknesses

The strengths of our analyses lie in the methods adopted (multiple imputation to overcome potential bias because of missing data, controlling for baseline utility imbalances to provide more robust incremental ratio estimates, and probabilistic assessment of uncertainty), which are in line with current expert recommendations and national guidelines.32,35 In addition, the completion rate for participants providing sufficient information to be included in a complete-case analysis was high in comparison to other UK-based musculoskeletal studies.55,56

The presentation of results in a disaggregated format (eg, unit costs, resource use frequencies, resource use costs, and quality-of-life estimates) allows readers to consider the consequences of alternative scenarios and the consequential implications for resource allocation decisions (eg, different cost perspectives and variations in costing assumptions). The robustness of the results of our sensitivity analyses and the magnitude of the incremental QALY estimates (in relation to the incremental cost estimates) suggest that potential alternative scenarios or analytical assumptions would not affect the study findings.

Aspects of the data collection process may be regarded as limitations of the present study. First, reliance on people to recall their health care usage can introduce 2 types of bias: recall bias (the failure to remember a particular event) and telescoping (the tendency to remember distant events as occurring more recently). Although criticisms of self-report resource use data are well established,57 the method provides an efficient approach to data collection in the absence of accessible routine data sources; in addition, this method was used in previous UK-based musculoskeletal economic evaluations,55,56 which have included small-sample validation tests to verify the accuracy of self-report data in comparison with primary care records.56 Establishing optimal methods for estimating resource use alongside clinical trials remains a priority for health economic research.

A second limitation relates to the absence of data regarding hospital-based care. The study eligibility criteria excluded people with potentially serious pathology and those who were already on a surgical waiting list for total knee replacement. Given the lengthy waiting times for orthopedic consultations in the NHS, study participants were deemed unlikely to incur expensive NHS hospital-based care within the 12-month follow-up period. In addition, previous research showed that patients receiving acupuncture therapy for low back pain have fewer hospitalizations than active control groups.55 The absence of some secondary care data will distort mean cost estimates in the treatment groups. However, it is unlikely to alter the overall findings given the incremental nature of the analysis and the relatively low cost per QALY estimates.

Finally, the NHS perspective of the present study followed national UK guidelines, as opposed to the “societal” perspective advocated in other jurisdictions.35,58 Typically, the societal perspective incorporates the indirect costs to society resulting from reduced productivity in the workplace.59 Although the implications of choosing the UK-standard cost perspective remain unknown, the improvements in quality of life and the absence of between-group differences in health care resource use beyond the study treatment sessions suggest that knee pain–related work absenteeism is unlikely to be more frequent in people receiving acupuncture treatment.

Conclusion

According to the present study, given a conservative threshold of £20,000 per additional QALY, there is a 77% chance that true acupuncture provided by NHS physical therapists would be a cost-effective adjunct to a course of advice and exercise. The magnitude of the difference in health care resource use (across public and private providers), relative to the difference in health benefits, suggests that the results may be generalizable to other health care settings. However, economic benefits were identified for both acupuncture treatment groups, regardless of whether the skin was penetrated or the needle was manipulated in situ. The potential mechanisms of action of acupuncture are complex, and research with sham or nonpenetrating acupuncture devices has shown that these are not truly inactive placebo interventions.20,29 Further studies identifying the specific and nonspecific effects of this type of complementary medicine are needed to address clinical effectiveness and cost-effectiveness concerns.

Footnotes

  • Dr Whitehurst and Professor Foster provided concept/idea/research design. Dr Whitehurst, Professor Bryan, and Professor Foster provided writing. Dr Thomas and Ms Young provided data collection. Dr Whitehurst, Professor Bryan, and Dr Thomas provided data analysis. Professor Hay, Dr Thomas, and Professor Foster provided project management. Professor Hay and Professor Foster provided fund procurement and facilities/equipment. Professor Hay, Ms Young, and Professor Foster provided institutional liaisons. All authors provided consultation (including review of manuscript before submission).

  • The authors thank the patients and physical therapists involved in this study. They also acknowledge their colleagues at the Arthritis Research UK Primary Care Centre (Keele University) and the Health Economics Unit (University of Birmingham), particularly Professor Jo Coast, for their comments.

  • This study was approved by the West Midlands Multicentre Research Ethics Committee (UK) and by 13 local ethics committees.

  • An oral presentation of data from this study was given at the Chartered Society of Physiotherapy Congress; October 16–17, 2009; Liverpool, United Kingdom. A poster presentation of data from this study was given at the British Society for Rheumatology Annual Meeting/British Health Professionals in Rheumatology Spring Meeting; April 28–May 1, 2009; Glasgow, United Kingdom. Professor Foster also was invited to give an oral presentation at the British Medical Acupuncture Society Conference in October 30, 2010.

  • This article reports the results of a cost-utility analysis performed alongside a clinical trial. The clinical trial was supported by Project Grant H0640 from Arthritis Research UK and Support for Science funding secured by the North Staffordshire Primary Care Research Consortium for NHS service support costs. Professor Foster was funded by a Primary Care Career Scientist Award from the Department of Health and NHS research and development (UK).

  • * SPSS Inc, 233 S Wacker Dr, Chicago, IL 60606.

  • StataCorp LP, 4905 Lakeway Dr, College Station, TX 77845.

  • Received July 22, 2010.
  • Accepted December 28, 2010.

References

View Abstract