背景
针灸在英国广泛应用，估计每年提供400万次治疗，主要用于慢性疼痛病症，如肌肉骨骼疼痛，头痛或偏头痛。除慢性疼痛症状外，心理困扰，包括抑郁症，也一直是人们咨询针灸师的常见原因。针灸的证据依然是斑驳的，对临床效果有一些争议，尤其是针灸不仅仅是安慰剂的程度。关于针灸在日常实践中的作用的临床问题仍然存在，针灸是否具有成本效益的问题仍然存在。该领域的挑战之一是现有证据的质量。在大多数情况下，对针灸的系统评价对于有效性得出任何明确结论的能力有限。与对许多其他慢性疼痛物理疗法的研究一样，进行了太多的弱势试验，方法学标准不够健壮。然而，在这种不确定性的气氛中，针灸试验的质量和数量最近急剧增加，特别是针对慢性疼痛病症，这使得该研究计划成为大幅降低不确定性的独特机会。在这方面，我们的主要目标是使用高质量的方法和最佳证据来确定针灸对慢性疼痛和抑郁症的临床和经济影响。
针灸慢性疼痛
虽然针灸广泛用于慢性疼痛，但对于如何工作的理解有限，反过来又导致其作为治疗方式的潜在作用的一些不确定性。不确定性与针灸可能仅仅是一个戏剧性的安慰剂的关注以及当考虑到偏见问题时，任何效果的大小可能会消失的建议相关。为了解决不确定性，在我们的第一项研究中，我们根据这些比较的直接试验数据，确定针灸针对慢性疼痛的影响大小，针灸是否与假针灸或非针灸控制进行比较。在建立针灸Triallists协作（ATC）的关键演习者的协作下，我们能够进行个人患者数据（IPD）荟萃分析。为了建立符合条件的研究，我们确定了针对四种慢性疼痛病症的随机对照试验（RCT），其中分配隐藏被明确地确定为足够。对31项符合条件的RCT患者中的29项，共有17,922例患者进行了分析。在包括所有符合条件的RCT时，我们的主要结果是，所有四种疼痛状况下，针灸在统计学上显着高于假手术对照组和非针灸对照组（p \u003c0.001）。当排除强烈倾向于针灸的偏僻RCT的一组时，每个疼痛状况的影响大小相似。接受针灸的患者的疼痛少于接受假手术治疗的患者，其效果大小基于0.23 [95％置信区间（CI）0.13〜0.33]的标准平均差异（SMD），0.16（95％CI 0.07〜0.25）和0.15 95％CI 0.07〜0.24），分别为背痛和颈痛，骨关节炎和慢性头痛。针刺与非针刺对照组比较，效果分别为0.55（95％CI 0.51〜0.58），0.57（95％CI 0.50〜0.64）和0.42（95％CI 0.37〜0.46）。进行了各种灵敏度分析，包括与出版偏倚相关的敏感性分析，对研究的主要发现影响不大。总而言之，在这些高质量的研究中，针灸和非针灸对照之间发现了临床相关性和统计学上的显着性差异，这表明针灸是适合慢性疼痛的转诊选择。真实针灸与假针灸的统计学显着性差异表明针灸不仅仅是安慰剂。
骨关节炎的物理治疗
许多系统评价已经评估了膝关节骨性关节炎的各种物理治疗方法，但是在同一分析中没有研究将这些不同的物理治疗相互比较。在第二项研究中，我们使用网络荟萃分析来解决膝关节骨关节炎的有效物理治疗相对于减轻疼痛的问题。我们回顾了截至2013年1月的文献，其中涉及搜索17个电子数据库。我们确定膝关节骨性关节炎患者身体治疗的RCTs，其中疼痛报告为结果。在网络荟萃分析中，合成了直接和间接证据，以比较针灸与其他相关物理治疗缓解疼痛的疗效。总共114项试验（共22项治疗和9709例）纳入分析。许多试验被归类为质量差，几个领域的偏倚风险很高。统计学上显着优于标准（通常）护理，与标准护理相比，统计学显着优于标准（通常）护理，产生疼痛改善：干预疗法，针灸，经皮神经电刺激（TENS），脉冲电刺激，平衡疗法，有氧运动，假针灸和肌肉强化运动。灵敏度分析包括更好的研究，最常见的针灸（11项试验），肌肉强化运动（9项试验）和假针灸（8项试验）。针灸在统计学上显着高于肌肉强化运动和假针刺（SMD 0.49,95％CI 0.00至0.98;分别为0.34,95％CI 0.03至0.66）。针灸统计学显着优于标准护理（SMD 1.01,95％CI 0.61至1.43）。总之，在网络荟萃分析中，针灸被发现是减轻膝骨关节炎膝关节疼痛的最有效的物理治疗方法之一。然而，鉴于许多证据质量差，许多物理治疗的有效性存在不确定性。
评估针灸治疗骨关节炎的成本效益
为了解决针灸对慢性疼痛的成本效益问题，我们受益于ATC资料库的数据可用性。在第三项研究中，我们对ATC资料库的个别患者级数据进行了网络荟萃分析。通过综合所有数据，包括间接数据，我们得出了治疗效果估计值，这是用于资源分配决策的成本效益分析的关键参数。在这种主要的方法研究中，开发了用于分析报告异质连续结果的多个个体患者级数据集的新贝叶斯方法。成本效益分析的一个重要步骤是基于偏好的健康相关生活质量衡量标准，如EuroQol-5维度（EQ-5D）。为了合成异质结果，我们使用映射将异质结果转换和比较到EQ-5D摘要指标量表。开发的模型需要贝叶斯随机效应网络荟萃分析规范，包括可交换的疼痛型相互作用效应。还证明了对成本效益分析的影响。使用针灸治疗初级保健慢性疼痛的案例研究，包括头痛/偏头痛，肌肉骨骼疼痛和膝关节骨性关节炎的方法。通过使用与第一项研究相同的证据基础，我们的分析包括来自28项试验的大约17,500名患者（我们无法使用一次试验中的IPD），其中我们将针灸，假针灸和常规护理相互比较。映射EQ-5D估计的综合发现，针灸与常规护理相比有效，中位治疗效果估计为头痛/偏头痛的0.056 [95％可信区间（CrI）0.021至0.092]，0.082（95％CrI 0.047至0.116 ），肌肉骨骼疼痛为0.079（95％，CrI 0.042〜0.114）。针灸针灸EQ-5D的益处较小，更不确定（头痛：0.004,95％CrI -0.035〜0.042;肌肉骨骼疼痛：0.023,95％CrI -0.007〜0.053;膝关节骨性关节炎：0.022,95％ CrI -0.014至0.060）。尽管并非所有相关干预措施都进行了比较，但成本效益结果表明，针灸与常规护理相比较，具有成本效益，成本效益比（IQER）从7000英镑到14,000英镑不等质量调整生命年（QALY）跨痛苦类型的头痛/偏头痛，肌肉骨骼科学综合NIHR期刊图书馆www.journalslibrary.nihr.ac.uk xx膝关节疼痛和骨关节炎。在这种面向方法的案例研究中，我们表明异质结果映射到EQ-5D摘要指标为进行成本效益分析提供了有用的一步。
非药物治疗骨关节炎的成本效益
在这项研究计划的第四项研究中，我们进行了经济评估，以评估用于减轻膝关节骨性关节炎慢性疼痛的非药物治疗的成本效益。我们使用新的网络荟萃分析方法来综合17项积极干预和三项控制干预措施的RCT证据。数据来自第二项研究进行的系统评价，其中包括7507例患者的88例RCT。第一项研究的IPD可用于五项RCT，其中包括1329例患者。在试验中报告了广泛的与健康有关的生活质量结果。由于分析的最终目的是为了通报英国的资源分配决策，将这些仪器的数据映射到合成之前的EQ-5D偏好权重，扩展了第三项研究中开发的方法。进行灵敏度分析以解决与不良学习行为相关的潜在偏差，并探讨报告时间点对研究结果的重要性。来自NHS信托网站的试验数据，专家意见，文献资料和资料，对干预措施的资源使用情况进行了估算。与其他EQ-5D的变化有关的非干预资源使用是从另一项英国试验获得的。结果和成本是使用时间范围为8周的曲线下成本效益模型进行合成的。当所有试验都包括在综合中时，与常规护理相比，TENS具有成本效益，每QALY的ICER为2690英镑。当分析仅限于具有适当分配隐藏的试验时，与TENS相比，针灸具有成本效益，ICAL为13,502英镑/ QALY。关于用于治疗膝关节骨性关节炎的许多非药物干预的长期影响的数据有限。通过这种分析的试验中的积极和控制干预措施受到方法的不均匀性，持续时间和施用强度的影响。这些结果受到一些决定的不确定性，完美信息的预期价值相对较高，这表明额外的研究可能具有成本效益。
针灸或咨询抑郁症
抑郁症是发病的重要原因。许多患者将非药理疗法的兴趣传达给其全科医生。初级保健中针灸和抑郁症咨询的系统评价发现有限证据。第五项研究的目的是评估针灸与常规护理和咨询相比，与常规护理相比，在初级保健中继续患有抑郁症的患者。此外，需要进行成本效益分析，以了解这些疗法是否应被视为有利的卫生资源。总共755例抑郁症患者（Beck Depression Inventory-II评分≥20）被招募到英国北部27个初级保健实践中进行的RCT。分配到三个武器之一，使用比例为2：2：1针灸（n = 302），咨询（n = 302）和平时护理（n = 151）。平均患者健康问卷9项（PHQ-9）在3个月时的差异是主要结果。患者随访12个月以上，分析为意向治疗。进行了额外的定量和定性研究。关于PHQ-9评分的数据可在614个患者3个月和572例12个月。参加针灸10次，平均9次，参加咨询。与普通护理相比，针灸3个月平均PHQ-9抑郁评分差异有统计学意义（-2.46,95％CI为-3.72〜-1.21），辅助治疗与常规护理相比（-1.73,95％CI -3.00到-0.45），针灸12个月（-1.55,95％CI -2.41〜-0.70）和咨询（-1.50,95％CI -2.43〜-0.58）与常规护理相比。在控制时间和注意力的情况下，针灸和咨询之间没有发现临床疗效差异。没有报告严重的治疗相关不良事件。该试验的目的不在于分开DOI：10.3310 / pgfar05030应用研究计划GRANTS 2017 VOL。 5号3皇后的HMSO打印机和控制器2017.这项工作是由MacPherson等人制作的。根据卫生国务秘书签发的委托合同条款。为了私人研究和研究的目的，这个问题可以自由复制，并提取（或者完整的报告）可以被包括在专业期刊中，只要适当的确认和复制与任何形式的广告无关。商业复制申请应提交给：NIHR期刊图书馆，国家卫生研究所，评估，试验和研究协调中心，阿姆斯特丹，南安普敦科学园大学，南安普敦SO16 7NS，英国研究所。 xxi具体来自非特定效果。针灸和咨询被发现具有比常规护理更高的平均QALY和成本。在基础案例分析中，针灸的ICAL为每加值QALY 4560英镑，成本效益高达0.62，每QALY的成本效益门槛为20,000英镑。与普通护理相比，辅导的情景分析（不包括针灸作为比较对象，不适当或不可用）导致ICER为7935英镑，成本效益为0.91的概率。总而言之，接受短期针灸或辅导治疗的初级保健中持续抑郁症患者与常规护理相比经历统计学显着降低的抑郁症。
结论
总之，这项研究方案迄今为止提供了针对针灸及其潜在影响的最实质性证据。利用RCT的现有数据，我们使用IPD荟萃分析，发现针灸在头痛和偏头痛，背部和颈部疼痛以及膝关节骨性关节炎的慢性疼痛状况下临床有效。我们从这个数据集的证据表明，针灸是一个统计学显着更有效的干预比安慰剂。在网络荟萃分析中，我们针对膝关节骨性关节炎的一系列物理疗法的证据表明，针灸与任何其他疗法相比，具有更高质量的试验，也是最有效的疗法之一。如果仅分析高质量的试验，针灸也是成本效益的。当所有试验都包括在综合报告中，包括低质量和高质量的试验时，我们发现TENS具有成本效益。在我们对抑郁症进行的试验中，我们发现针灸和咨询的统计学显着优于常规护理，针灸也具有成本效益。我们的结果仍然存在一些不确定性。例如，关于针灸对肌肉骨骼疼痛，头痛和偏头痛的成本效益分析，并不是所有与决策相关的竞争疗法都包括在分析中。关于用于治疗膝关节骨性关节炎的许多非药物干预的长期影响的数据很少，敏感性分析表明，成本效益模型结果可能对这些影响的程度敏感。当比较针灸和普通护理抑郁症时，尽管在比较针灸和辅导时，我们对时间和注意力进行了控制。然而，我们的研究计划使用高质量的方法来提供针对慢性疼痛和抑郁症的临床有效性和成本效益的最佳数据。有力的证据符合所有利益相关者的利益，我们的结果将用于通知患者，从业人员，决策者和服务专员。
原文：
Background
Acupuncture is widely practised in the UK, with an estimated 4 million treatments provided a year, primarily for chronic pain conditions such as musculoskeletal pain and headache or migraine. In addition to chronic pain conditions, psychological distress, including depression, has been a common reason that people have consulted acupuncturists. The evidence base on acupuncture has been patchy, with some controversy regarding the clinical benefits and specifically the extent that acupuncture is more than simply a placebo. Clinical questions remain as to the effect of acupuncture in everyday practice and there are unanswered questions regarding whether or not acupuncture is cost-effective. One of the challenges in the field has been the quality of the available evidence. For most conditions, systematic reviews of acupuncture have had a limited ability to draw any definitive conclusions regarding effectiveness. As with research into many other physical therapies for chronic pain, there have been too many underpowered trials conducted and the methodological standards have been insufficiently robust. In this climate of uncertainty, however, there has been a recent and dramatic increase in the quality and quantity of trials of acupuncture, especially for chronic pain conditions, which has provided this programme of research a unique opportunity to substantially reduce the uncertainty. In this context, our primary aim was to use high-quality methods and the best evidence available to determine the clinical and economic impact of acupuncture for chronic pain and depression.
Acupuncture for chronic pain
Although acupuncture is widely used for chronic pain, there is limited understanding of how it works, which in turn fuels some of the uncertainty as to its potential role as a treatment modality. The uncertainty is associated with a concern that acupuncture might simply be a theatrical placebo and the suggestion that when issues of bias are taken into account, the size of any effect might vanish. To address the uncertainty, in our first study we set out to determine the effect size of acupuncture for chronic pain, whether acupuncture is compared with sham acupuncture or with a non-acupuncture control, based on direct trial data on these comparisons. On building a collaboration of key triallists, the Acupuncture Triallists’ Collaboration (ATC), we were able to conduct an individual patient data (IPD) meta-analysis. To establish eligible studies we identified randomised controlled trials (RCTs) of acupuncture for the four chronic pain conditions in which allocation concealment was determined unambiguously to be adequate. Data from 29 of 31 eligible RCTs, with a total of 17,922 patients, were analysed. Our primary result, when including all eligible RCTs, was that acupuncture was statistically significantly more effective than both sham controls and non-acupuncture controls for all four pain conditions (p < 0.001). When an outlying set of RCTs that strongly favoured acupuncture was excluded, the effect size across each pain condition was similar. Patients receiving acupuncture had less pain than those receiving sham controls, with effect sizes based on standardised mean differences (SMDs) of 0.23 [95% confidence interval (CI) 0.13 to 0.33], 0.16 (95% CI 0.07 to 0.25) and 0.15 (95% CI 0.07 to 0.24) for back and neck pain, osteoarthritis and chronic headache, respectively. In the comparison between acupuncture and non-acupuncture controls, effect sizes were 0.55 (95% CI 0.51 to 0.58), 0.57 (95% CI 0.50 to 0.64) and 0.42 (95% CI 0.37 to 0.46), respectively. A variety of sensitivity analyses was conducted, including those related to publication bias, with little impact on the main findings of the study. To conclude, in these high-quality studies, a clinically relevant and statistically significant difference was found between acupuncture and non-acupuncture controls, suggesting that acupuncture is a suitable referral option for chronic pain. Statistically significant differences between true and sham acupuncture indicate that acupuncture is more than a placebo
. © Queen’s Printer and Controller of HMSO 2017. This work was produced by MacPherson et al. under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK. Xix
Physical therapies for osteoarthritis
Many systematic reviews have evaluated individual types of physical treatments for osteoarthritis of the knee, but no study has compared these different physical treatments against each other in the same analysis. In a second study we used a network meta-analysis to address the question of how effective physical treatments for osteoarthritis of the knee are, when compared with each other, for relieving pain. We reviewed the literature up to January 2013, which involved searching 17 electronic databases. We identified RCTs of physical treatments in patients with osteoarthritis of the knee in which pain was reported as an outcome. In a network meta-analysis, both direct and indirect evidence was synthesised to compare the effectiveness of acupuncture with that of other relevant physical treatments for alleviating pain. In total, 114 trials (covering 22 treatments and 9709 patients) were included in the analysis. Many trials were classed as being of poor quality with a high risk of bias in several domains. Eight interventions statistically significantly outperformed standard (usual) care, producing an improvement in pain compared with standard care: interferential therapy, acupuncture, transcutaneous electrical nerve stimulation (TENS), pulsed electrical stimulation, balneotherapy, aerobic exercise, sham acupuncture and muscle-strengthening exercise. The better-quality studies, most commonly of acupuncture (11 trials), muscle-strengthening exercise (nine trials) and sham acupuncture (eight trials), were included in a sensitivity analysis. Acupuncture was statistically significantly more effective than muscle-strengthening exercise and sham acupuncture (SMD 0.49, 95% CI 0.00 to 0.98; and 0.34, 95% CI 0.03 to 0.66, respectively). Acupuncture was also statistically significantly better than standard care (SMD 1.01, 95% CI 0.61 to 1.43). To conclude, in a network meta-analysis, acupuncture was found to be one of the more effective physical treatments for alleviating osteoarthritis knee pain in the short term. However, given that much of the evidence was of poor quality, there is uncertainty about the effectiveness of many physical treatments.
Towards evaluating the cost-effectiveness of acupuncture for osteoarthritis
To address the question of the cost-effectiveness of acupuncture for chronic pain, we have benefited from the availability of the data from the ATC repository. In this third study we conducted a network meta-analysis of the individual patient-level data from the ATC repository. By synthesising all of the data, including indirect data, we derived treatment effect estimates, a key parameter for cost-effectiveness analyses to be used for resource allocation decisions. In this primarily methodological study, new Bayesian methods for analysing multiple individual patient-level data sets reporting heterogeneous continuous outcomes were developed. An essential step towards a cost-effectiveness analysis is a preference-based measure of health-related quality of life, such as the EuroQol-5 Dimensions (EQ-5D). To synthesise heterogeneous outcomes, we used mapping to convert and compare heterogeneous outcomes on to the EQ-5D summary index scale. The models developed entailed a Bayesian random-effects network meta-analysis specification, including exchangeable pain type interaction effects. The implications for cost-effectiveness analysis were also demonstrated. The methods were illustrated using a case study of acupuncture for chronic pain in primary care, including headache/migraine, musculoskeletal pain and osteoarthritis of the knee. By using the same evidence base as in the first study, our analysis included approximately 17,500 patients from 28 trials (we were unable to use IPD from one trial), in which we compared acupuncture, sham acupuncture and usual care with each other. The synthesis of mapped EQ-5D estimates found that acupuncture was effective compared with usual care, with median treatment effects estimated as 0.056 [95% credible interval (CrI) 0.021 to 0.092] for headache/migraine, 0.082 (95% CrI 0.047 to 0.116) for musculoskeletal pain and 0.079 (95% CrI 0.042 to 0.114) for osteoarthritis of the knee. The EQ-5D benefit of acupuncture over sham acupuncture was smaller and more uncertain (headache: 0.004, 95% CrI –0.035 to 0.042; musculoskeletal pain: 0.023, 95% CrI –0.007 to 0.053; osteoarthritis of the knee: 0.022, 95% CrI –0.014 to 0.060). Although not all relevant interventions were compared for decision-making purposes, cost-effectiveness results suggest that when acupuncture is compared with usual care alone it is cost-effective, with incremental cost-effectiveness ratios (ICERs) ranging from £7000 to £14,000 per quality-adjusted life-year (QALY) across the pain types of headache/migraine, musculoskeletal SCIENTIFIC SUMMARY NIHR Journals Library www.journalslibrary.nihr.ac.uk xx pain and osteoarthritis of the knee. In this methodologically oriented case study we showed that the mapping of heterogeneous outcomes on to the EQ-5D summary index provides a useful step towards conducting a cost-effectiveness analysis. Cost-effectiveness of non-pharmacological treatments for osteoarthritis
In a fourth study within this programme of research, we conducted an economic evaluation to evaluate the cost-effectiveness of non-pharmacological treatments used to reduce chronic pain in osteoarthritis of the knee. We used novel network meta-analysis methods to synthesise RCT evidence for 17 active interventions and three control interventions. Data were obtained from the systematic review carried out in the second study, which were available for 88 RCTs including 7507 patients. IPD from the first study were available for five of the RCTs, including 1329 patients. A wide range of health-related quality-of-life outcomes were reported in the trials. As the ultimate objective of the analysis was to inform resource allocation decisions in the UK, data from these instruments were mapped to EQ-5D preference weights prior to synthesis, extending methods developed in the third study. Sensitivity analyses were conducted to address potential bias associated with poor study conduct and to explore the importance of the time point of reporting for the study results. Resource use associated with the interventions was estimated from trial data, expert opinion, the literature and information obtained from NHS trust websites. Non-intervention resource use related to changes in EQ-5D was obtained from another UK trial. Outcomes and costs were synthesised using an area-under-the-curve cost-effectiveness model with a time horizon of 8 weeks. When all trials were included in the synthesis, TENS was cost-effective with an ICER of £2690 per QALY compared with usual care. When the analysis was restricted to trials with adequate allocation concealment, acupuncture was cost-effective with an ICER of £13,502 per QALY compared with TENS. There were limited data regarding the long-term effects of many non-pharmacological interventions used to treat osteoarthritis of the knee. The active and control interventions in the trials informing this analysis were subject to heterogeneity in the method, duration and intensity with which they were administered. These results are subject to some decision uncertainty and the expected value of perfect information is relatively high, suggesting that additional research may be cost-effective.
Acupuncture or counselling for depression
Depression is a significant cause of morbidity. Many patients have communicated an interest in non-pharmacological therapies to their general practitioners. Systematic reviews of acupuncture and counselling for depression in primary care have identified limited evidence. The aim of the fifth study was to evaluate acupuncture compared with usual care and counselling compared with usual care for patients who continue to experience depression in primary care. Moreover, a cost-effectiveness analysis is needed to understand whether or not such therapies should be considered a good use of limited health resources. In total, 755 patients with depression (Beck Depression Inventory-II score of ≥ 20) were recruited to a RCT carried out in 27 primary care practices in the north of England. Allocation was to one of three arms using a ratio of 2 : 2 : 1 to acupuncture (n = 302), counselling (n = 302) and usual care alone (n = 151). The difference in mean Patient Health Questionnaire-9 items (PHQ-9) score at 3 months was the primary outcome. Patients were followed up over 12 months and analysis was by intention to treat. Additional quantitative and qualitative substudies were conducted. Data on PHQ-9 scores were available for 614 patients at 3 months and 572 patients at 12 months. A mean of 10 sessions was attended for acupuncture and a mean of nine sessions was attended for counselling. There was a statistically significant reduction in mean PHQ-9 depression score at 3 months for acupuncture compared with usual care (–2.46, 95% CI –3.72 to –1.21) and counselling compared with usual care (–1.73, 95% CI –3.00 to –0.45), and at 12 months for acupuncture (–1.55, 95% CI –2.41 to –0.70) and counselling (–1.50, 95% CI –2.43 to –0.58) compared with usual care. When controlling for time and attention, no significant differences in clinical outcome were found between acupuncture and counselling. No serious treatment-related adverse events were reported. The trial was not designed to separate DOI: 10.3310/pgfar05030 PROGRAMME GRANTS FOR APPLIED RESEARCH 2017 VOL. 5 NO. 3 © Queen’s Printer and Controller of HMSO 2017. This work was produced by MacPherson et al. under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK. xxi out specific from non-specific effects. Acupuncture and counselling were found to have higher mean QALYs and costs than usual care. In the base-case analysis, acupuncture had an ICER of £4560 per additional QALY and was cost-effective with a probability of 0.62 at a cost-effectiveness threshold of £20,000 per QALY. A scenario analysis of counselling compared with usual care, excluding acupuncture as a comparator when inappropriate or unavailable, resulted in an ICER of £7935 and a probability of being cost-effective of 0.91. To summarise, patients with ongoing depression in primary care who received a short course of either acupuncture or counselling experienced statistically significant reductions in depression compared with usual care alone.
Conclusion
In conclusion, this programme of research has provided the most substantive evidence to date on acupuncture and its potential impact. Drawing on the existing data from RCTs, we used an IPD meta-analysis and found acupuncture to be clinically effective across the chronic pain conditions of headache and migraine, back and neck pain, and osteoarthritis of the knee. Our evidence from this data set suggests that acupuncture is a statistically significantly more effective intervention than placebo. In a network meta-analysis, our evidence across a range of physical therapies for osteoarthritis of the knee suggested that acupuncture is associated with more high-quality trials than any of the other therapies and is also one of the most effective therapies. Acupuncture was also cost-effective if only high-quality trials were analysed. When all trials were included in the synthesis, including both low- and high-quality trials, we found TENS to be cost-effective. In the trial that we conducted on depression, we found that acupuncture and counselling were statistically significantly better than usual care and that acupuncture was also cost-effective. There remains some uncertainty regarding our results. For example, with regard to the cost-effectiveness analyses of acupuncture for musculoskeletal pain and headache and migraine, not all competing therapies relevant to decision-making were included in the analysis. There are few data regarding the long-term effects of many non-pharmacological interventions used to treat osteoarthritis of the knee, and sensitivity analyses suggested that the cost-effectiveness model results may be sensitive to the magnitude of these effects. When comparing acupuncture with usual care for depression, there was no control for non-specific effects, although we did control for time and attention when comparing acupuncture with counselling. Nevertheless, our programme of research has used high-quality methods to provide the best possible data on the clinical effectiveness and cost-effectiveness of acupuncture for chronic pain and depression. Robust evidence is in the interests of all stakeholders and our results will be used to inform patients, practitioners, policy-makers and commissioners of services
