Scoping review of systematic reviews of complementary medicine for musculoskeletal and mental health conditions.

OBJECTIVE: To identify potentially effective complementary approaches for musculoskeletal (MSK)-mental health (MH) comorbidity, by synthesising evidence on effectiveness, cost-effectiveness and safety from systematic reviews (SRs).
DESIGN: Scoping review of SRs.
METHODS: We searched literature databases, registries and reference lists, and contacted key authors and professional organisations to identify SRs of randomised controlled trials for complementary medicine for MSK or MH. Inclusion criteria were: published after 2004, studying adults, in English and scoring >50% on Assessing the Methodological Quality of Systematic Reviews (AMSTAR); quality appraisal checklist). SRs were synthesised to identify research priorities, based on moderate/good quality evidence, sample size and indication of cost-effectiveness and safety.
RESULTS: We included 84 MSK SRs and 27 MH SRs. Only one focused on MSK-MH comorbidity. Meditative approaches and yoga may improve MH outcomes in MSK populations. Yoga and tai chi had moderate/good evidence for MSK and MH conditions. SRs reported moderate/good quality evidence (any comparator) in a moderate/large population for: low back pain (LBP) (yoga, acupuncture, spinal manipulation/mobilisation, osteopathy), osteoarthritis (OA) (acupuncture, tai chi), neck pain (acupuncture, manipulation/manual therapy), myofascial trigger point pain (acupuncture), depression (mindfulness-based stress reduction (MBSR), meditation, tai chi, relaxation), anxiety (meditation/MBSR, moving meditation, yoga), sleep disorders (meditative/mind-body movement) and stress/distress (mindfulness). The majority of these complementary approaches had some evidence of safety-only three had evidence of harm. There was some evidence of cost-effectiveness for spinal manipulation/mobilisation and acupuncture for LBP, and manual therapy/manipulation for neck pain, but few SRs reviewed cost-effectiveness and many found no data.
CONCLUSIONS: Only one SR studied MSK-MH comorbidity. Research priorities for complementary medicine for both MSK and MH (LBP, OA, depression, anxiety and sleep problems) are yoga, mindfulness and tai chi. Despite the large number of SRs and the prevalence of comorbidity, more high-quality, large randomised controlled trials in comorbid populations are needed. © Author(s) (or their employer(s)) 2018. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

A comprehensive scoping review covering a wide range of individual musculoskeletal and mental health conditions, in the notable dearth of systematic reviews (SRs)/randomised controlled trials in comorbid populations.

This review used systematic searching and screening techniques.

Results are limited by our pragmatic choice to only include high-quality SRs (high scoring on AMSTAR checklist (Assessing the Methodological Quality of Systematic Reviews)) and definition of complementary medicine as including a practitioner in its delivery.

Date and language limitations may have excluded some topics.

Introduction

Musculoskeletal (MSK) and mental health (MH) disorders are two conditions resulting in some of the highest burden in terms of disability, accounting for 30.5% and 21.1% of years lived with disability, respectively.1 Both affect one in four UK adults at some point in their life.2–4 MSK disorders lead to very high healthcare expenditure and loss of work,5 in the UK accounting for 7.5 million working days lost annually6 and 60% of occupational sick leave.2 In England, it has been estimated that MH costs £105 billion, and treatment costs are expected to rise substantially.7

Low back pain (LBP) is one of the most common health conditions,8–10 with a mean global general population prevalence (regardless of prevalence period) of 31%10 and a UK prevalent population of 17.3 million, with 3.1 million adults suffering during an entire year,11 costing the National Health Service (NHS) over £500 million/year.12 In the UK, 8.75 million people sought primary care treatment for osteoarthritis (OA) (over a 7-year consultation period)13 which is estimated to have resulted in £14.8 billion indirect costs.14

The most common MH conditions are depression and anxiety.3 4 In England, 19% of the general adult population reported ever being diagnosed with depression and 6% ever being diagnosed with generalised anxiety disorder.15 Insomnia is another common MH problem, affecting 6.4% of the UK general population.16

There is an increasing recognition of the importance of multimorbidity, that is, the coexistence of two or more physical and/or mental long-term conditions in an individual, which is becoming the norm in the UK.17–19 In a cross-sectional study in Scotland, 8.3% of primary care patients (36% of those with multimorbidity) had both a physical and a MH disorder20 and around 50% of patients with depression had pain symptoms.21 Long-term MSK disorders are closely associated with multimorbidity.22

Complementary medicine is ‘a diverse group of health-related therapies and disciplines which are not considered to be a part of mainstream medical care’, including osteopathy, chiropractic, acupuncture, herbal medicine and homeopathy.23 Estimates of 12-month prevalence of use of any complementary medicine are 0.3%–86% in Europe24 and 26% in England.25

Many MSK and MH conditions, and comorbid MSK and MH, have limited conventional treatment options, ‘effectiveness gaps’, which are potential key areas for complementary medicine.26 27 Both MH28–30 and MSK24 31 are common reasons for patients to use complementary medicine. We used scoping methodology to identify priority areas where complementary medicine may be useful.32 The overall aim of scoping studies is ‘to map rapidly the key concepts underpinning a research area and the main sources and types of evidence available’ in a formal, systematic and transparent way (which is often lacking in research priority setting exercises).33 Scoping study methodology was used to identify gaps in the research base and priorities for future research. Scoping work is an essential step in the development of a strategy for the integration of complementary medicine with conventional care.27 Preliminary searches showed that evidence for comorbid populations was very limited, so we chose to include independent evidence on MSK and MH conditions. This review was part of a wider scoping study, which obtained a breadth of perspectives—the results of this review were combined with findings from consultation with conventional and complementary medicine practitioners, a public survey and case studies of NHS provision of complementary medicine, to inform the final choice of conditions and complementary medicine approaches for a randomised controlled trial (RCT).

Methods

The scoping study followed Arksey and O’Malley’s32 framework plus the refinements suggested by Levac et al,34 Daudt et al 35 and Colquhoun et al.36 Scoping reviews are broader and more exploratory than systematic reviews (SRs). They review a large number of papers to map the evidence—focussing on breadth rather than detail; have few preconceived ideas regarding focus (particularly types of MSK/MH in this context); do not necessarily assess individual study quality; and use a narrative approach to analysis, rather than synthesising or aggregating quantitative data.32 The reader is advised to refer to the original SRs we have included for more detailed content.

Aims

This study aimed to identify which practitioner-based complementary approaches have evidence for both MSK disorders and MH conditions in order to identify research priorities in terms of treatment choice and specific patient groups/conditions for a future pragmatic trial of comorbid MSK and MH in UK primary care. The specific aims were to:

Collate the evidence on effectiveness, cost-effectiveness and safety from recent SRs of trials of complementary medicine for MSK and MH conditions.

Identify areas where there is high-quality evidence of effectiveness but sufficient uncertainty to justify a trial.

Identify areas where there is also some evidence of cost-effectiveness and safety.

Searches

As this was a scoping review, we used a wide range of methods to identify SRs (published, unpublished, in progress). Literature databases were searched (see below for details of searches): AMED, Medline, EMBASE, CINAHL, PsycINFO; Index to theses; Cochrane library; Epistemonikos; SIGLE. Three trial registries were searched (UKCRN, Health service research projects in progress, Prospero). Reference lists of ‘overviews of SRs’ and reports (from professional/patient organisations) were searched. We emailed key authors in the field. We also asked the project team, steering group and public and patient involvement (PPI) group, and used websites/social media (university websites, Research Gate and Twitter) to access the wider research community. We asked patient and professional organisations to ask their members for any information (by email, newsletters or social media). These included: disease-specific organisations, for example, Arthritis Research UK, Back Care and MIND; complementary medicine organisations, for example, Research Council for Complementary Medicine and Complementary and Natural Healthcare Council; primary care organisations, for example, UK Royal College of General Practitioners. We presented at one conference and distributed a flyer at another. We contacted the authors of eligible conference abstracts and protocols to ascertain if the SR was available, that is, submitted or in press.

For complementary medicine, we used MeSH terms or subject headings, plus additional terms for complementary medicine which were not indexed. To ensure relevance to UK primary care, we only included complementary approaches which involve practitioner-led treatment. We therefore excluded herbal medicine, as the majority of studies do not involve a practitioner. We included manipulation, manual therapy and mobilisation as techniques commonly practised by some complementary medicine practitioners (as well as conventional practitioners). For MSK conditions, we used ‘musculoskeletal diseases/disorders’ as index terms, adding in any which were missing. For MH, we searched for common, minor MH disorders and symptoms, excluding long-term/severe mental illnesses. Thus, words in titles/abstracts were searched for (rather than subject headings), with reviews needing to include patients with MH symptoms/diagnoses rather than just measuring MH outcomes. See online supplementary appendix 1 for a full list of included and excluded non MSK or MH conditions.

Database searches were conducted in June 2015 and updated in February 2016, in collaboration with a librarian. We used filters for SRs. We excluded reviews published pre-2005, in order to make the final number of papers manageable with minimal impact on our results (given that the majority of SRs were published in 2010 and after (122/158 SRs)). We only included English language reviews. Databases were searched for two topics: complementary medicine+musculoskeletal disorders and complementary medicine+mental health. See online supplementary appendix 1 for details and supplementary appendix 2 for an example search.

We used the inclusion/exclusion criteria (see online supplementary appendix 1) to screen the results from each search. This was a multistage process involving the whole project team. AL screened the titles of all search results, and DS checked the results, with any disagreements discussed and inclusion criteria amended accordingly. AL then screened the abstracts, which were also independently screened by another author. Results were compared, any disagreements discussed and the criteria again amended accordingly. Any remaining disagreements were discussed and resolved with a third author. AL then screened the full text of the agreed papers.

Data extraction

AL extracted the following data from each SR: author, year, location, complementary medicine/s, health condition/s, for MSK reviews whether MH was included and for MH reviews whether findings applied to MSK populations, methodological features (sources searched, inclusion/exclusion criteria, quality assessment), results and conclusions (effectiveness, safety, cost-effectiveness). Where possible, these data were copied and pasted directly, to avoid any misinterpretation.

Quality appraisal of SRs

Although not essential,32 quality assessment is recommended in scoping reviews.37 We used a limited quality appraisal of SRs to help inform prioritisation, using AMSTAR (Assessing the Methodological Quality of Systematic Reviews), a validated checklist.38 Due to time constraints and the number of SRs to be appraised, a simple procedure was followed, using the ‘search’ function in the text of SRs to find particular terms. The emphasis was on obtaining relative scores, allowing ranking of the included reviews, rather than an absolute score of quality. For each of the 11 items on the AMSTAR checklist, a score of 0 was given if they did not meet the criteria or if information was unavailable, a score of 1 if the criteria was met. The scores were summed to give a score for each SR out of a total of 11.

Any reviews which scored 5 or less (half of the possible total) on AMSTAR were not considered in the evidence synthesis.

Evidence synthesis

In line with Arksey and O’Malley’s32 framework for conducting scoping reviews, we used a ‘descriptive‐analytical’ method, based on the narrative tradition, applying a common analytical framework to the reviews. Evidence synthesis was performed by AL and all authors met regularly to discuss the process and emerging results.

Reviews were placed into a matrix of health condition/symptom versus complementary medicine, for example, LBP and acupuncture (see tables 1 and 2).

Where there was more than one SR in an area, one (occasionally more than one—where they studied different interventions, eg, types of manipulation) was prioritised, based on recency, breadth, quality (AMSTAR score) and level of analysis (eg, meta-analysis).

There were three steps to the evidence synthesis: step 1: extraction and narrative synthesis of data from prioritised SRs: step 2: ranking of areas (condition vs complementary medicine) using an Excel framework (see below for ranking criteria); and step 3: computation of sample size.

Step 1 data extraction

For each area (condition vs complementary medicine) in the matrix, data were synthesised for:

Effectiveness: the conclusions on effectiveness were extracted from the prioritised SR/s, including information on comparator and outcomes.

Quality of studies: in line with guidance on scoping reviews,32 we did not assess the quality of the individual RCTs in each SR, but synthesised the SR authors’ conclusions about quality, from the prioritised review/s. We had three categories of quality: poor, moderate or good. Moderate included SRs reporting mixed results.

Safety: any information on safety was synthesised across all the included SRs.

Cost-effectiveness: any information on cost-effectiveness was synthesised across all the included SRs.

Step 2 ranking

Using the information extracted in step 1, we ranked the areas in the matrix (condition vs complementary medicine), based on consideration of the following criteria:

Effectiveness: areas were ranked highly where there was enough moderate/good quality evidence to establish proof of concept, but still key gaps. Data on effectiveness came from the conclusions of the prioritised SRs regarding effectiveness and quality.

Safety: areas where there was evidence of harm were ranked lower.

Cost: areas where there was any indication that the intervention may have potential for cost-effectiveness were ranked highly.

Comorbidity: any interventions that showed an indication (of any quality) for comorbid MH and MSK patients were ranked highly.

Step 3 sample size

Once a ranked list of areas had been compiled, information on total sample size for the prioritised SR/s (ie, total number of participants in the trials included in the SR/s) was added to allow further discrimination, classified as: small (<500 participants), medium (501 to 3000) or large (>3000).

Patient and public involvement

The wider scoping study included a PPI group of seven people with experience of MSK/MH issues, who advised the research team on a variety of issues as the study took shape and progressed. They met four times over the study period and provided input into the literature review questions, search strategy and selection criteria. The PPI group also provided ideas on ways to disseminate the findings to the general public and to patient groups.

Results

Search results

Figure 1 shows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses flowchart of the searches.39 Searches resulted in 8393 references—4304 for MH and 4089 for MSK, plus three for MH identified elsewhere (email update sent by journal publisher). After removing duplicates, Cochrane reviews which were withdrawn, Cochrane reviews which had been updated or those not in English, 5230 records remained for title screening. This resulted in 486 MH and 739 MSK records for the abstract screening phase. We then excluded 49 (13 MH and 36 MSK) records, that is, editorials, commentaries, abstracts etc. After abstract screening, there remained 444 studies (180 MH and 264 MSK). At this stage, a further 12 records were added—three from grey literature (Prospero/Sigle/HSRProj), five from conference abstracts and four from the reference lists of overviews of SRs and other key reports. This produced 456 records to be screened at the full-text stage, after which 206 records remained: 48 MH, 158 MSK.

Figure 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flowchart. MH, mental health; MSK, musculoskeletal; SR, systematic review.

During screening at the title stage (AL and DS), there were 49 disagreements (out of 4411 records; 1%). During abstract screening (AL and each of the six authors), there were a total of 296 initial disagreements (out of 1165 records; 25%). Most were resolved by discussion without requiring a third reviewer.

Of the 158 MSK SRs, 84 were included as high quality (scored six or more on AMSTAR): LBP,40–69 fibromyalgia,70–81 neck pain,40 82–90 OA,91–101 rheumatoid arthritis,97 102–105 general MSK pain/conditions,106–110 shoulder pain/disorders,111–115 myofascial trigger point pain106 116 117 and other musculoskeletal (MSK) conditions.106 118–123

The 158 reviews covered 17 different and 15 different disorders. The overall matrix is shown below in table 1. LBP was by far the most common topic, addressed by 64/158 (40%) of reviews. Fibromyalgia was the next most common, followed by OA then neck pain/disorders. Acupuncture was the most common complementary medicine studied (58/158 SRs; 37%), followed by manual therapy (including manipulation and mobilisation). The mean AMSTAR score for methodological quality of MSK reviews was 6/11 (ranging from 0 to 11). Online supplementary appendix 3 shows which items on the AMSTAR checklist were reported by the reviews.

Table 1

Matrix of musculoskeletal reviews (numbers=number of reviews)

	Low back pain		Fibromyalgia		Neck pain/ disorders		Osteoarthritis		Rheumatoid arthritis		MSK pain/ conditions		Myofascial pain		Shoulder pain/ disorder		Other		Total
	Total number	High quality	Total number	High quality	Total number	High quality	Total number	High quality	Total number	High quality	Total number	High quality	Total number	High quality	Total number	High quality	Total number	High quality
Acupuncture related
Acupressure	2	2			2	1					1	1					1	0	6
Acupuncture	12	6	9	6	5	2	10	6	4	2	4	2	6	3	2	1	6	5	58
Moxibustion	1	0	1	0			1	0	2	1							1	0	6
Manipulation
Chiropractic	2	1	3	1			1	0					1	0					7
Manual therapy/spinal manipulation/mobilisation	12	9	8	4	14	9	2	1			1	0	1	0	5	3	3	2	46
Osteopathy	3	2	2	1							1	1							6
Other
Integrative therapies/multiple complementary medicine	2	2	2	2															4
Tai chi	1	1	1	0	1	0	7	5	3	1	1	1							14
Qigong					1	0	1	0	1	0									3
Pilates	10	5																	10
Complementary medicine exercise			2	2															2
Yoga	8	4	1	1	1	0	3	0	1	0	1	1							15
Mindfulness/meditation	2	0	5	3					2	0	4	2							13
Relaxation			3	1			1	0	1	0	1	0							6
Hypnotherapy	1	0	1	1			1	0	1	0									4
Spa therapy/balneotherapy	2	1	2	1					1	1									5
Traditional Chinese Medicine bodywork	2	1			1	0					1	0							4
Massage/therapeutic touch	4	2			4	2	1	0	1	0	1	0			1	0	1	1	13
Homeopathy			2	0															2
Total*	64	36	42	23	29	14	28	12	17	5	16	8	8	3	8	4	12	8	158 (84)

*Note that reviews can occur in more than one row/column.

Of the 48 MH SRs, 27 were high quality (scored six or more on AMSTAR): depression,124–134 anxiety,124 125 129 133 135–142 sleep,143–150 stress124 125 140 and other.124 The 48 reviews covered 18 different and five different categories of MH symptom or disorder. The overall matrix is shown in table 2. Many of the reviews included more than one MH condition. Of the 48 reviews, 36 (75%) included depression. Meditation, yoga, acupuncture and meditative movement were the most common complementary medicine. Online supplementary appendix 3 shows which items on the AMSTAR checklist were reported by the reviews.

Table 2

Matrix of mental health reviews (numbers=number of reviews)

	Depression		Anxiety		Stress		Insomnia/sleep		Distress (psychological)		Total*
	Total number	High quality	Total number	High quality	Total number	High quality	Total number	High quality	Total number	High quality
Acupuncture-related
Acupuncture	9	3	2	1			2	1			6
Auricular acupuncture							3	1			4
Multiple
Complementary medicine (multiple) (did not separate complementary medicine in analysis)	1	0	1	0			2	1			4
Meditation/mindfulness
Mindfulness	3	2	1	1	1	1			1	1	5
Meditation			3	3	1	1	1	0			9
Exercise
Meditative movement							3	3			6
Qigong	4	1	2	1	1	1					5
Tai chi	4	2	1	0			3	1			5
Yoga	4	2	3	3	1	0	1	0			8
Other
Homeopathy							1	1			2
Hypnosis	2	0					2	2			4
Music therapy	3	2	1	1							2
Relaxation	3	1	1	1			1	1			4
Reflexology							1	0			1
Reiki	1	1	1	1							2
Spinal manipulation											0
Massage/therapeutic touch	2	1	2	2							4
Morita therapy			1	1							2
Total*	36	15	19	15	4	3	20	11	1	1	48 (27)

*Note that reviews can occur in more than one row/column

The following sections present a summary of the areas where we found moderate/good quality evidence of effectiveness (from the prioritised SRs). For those conditions with evidence for more than one intervention, we have presented the results in tables.

Evidence for MSK disorders

The areas where there was moderate/good quality evidence were LBP, myofascial trigger point pain, OA, neck pain, fibromyalgia and lateral epicondylitis. It should be noted that the SR quality of evidence is based on the overall assessment of quality of the included RCTs by the SR authors, which is subject to some limitations.

Low back pain

Moderate/good quality evidence for LBP (see table 3) was found as follows:

Table 3

Evidence for low back pain

Comparator	Quality	Population size	Complementary medicine	Safety	Cost-effectiveness
Active intervention	Good	Medium	Osteopathy65	Some evidence of harm65*	NR
	Moderate	Large	Spinal manipulation/mobilisation (chronic only)40 43 44	Some evidence of safety44†	Mixed results—some evidence of CE47‡
	Moderate	Medium	Yoga62	Some evidence of safety51 60 62§	No data
	Moderate	Small	Balneotherapy52	Some evidence of safety52	NR
Usual care	Moderate	Large	Acupuncture40 50 (for pain, function, well-being, disability, range of movement, quality of life)	Some evidence of safety53 54¶	Some evidence of CE40**
	Moderate	Small	Spa therapy52	Some evidence of safety52††	NR
			Tai chi54	NR	NR
Placebo	Moderate	Large	Acupuncture (for pain and quality of life; negative evidence for disability).40 50	Some evidence of safety53 54¶	Some evidence of CE40**

*Two studies reported minor adverse events.

†One review identified serious adverse events, but this appeared unrelated to treatment. Minor adverse events included muscle soreness, stiffness and/or transient increase in pain.

‡Mixed results regarding cost-effectiveness,40 with limited evidence that manual therapy is more cost-effective than usual care.47

§No serious adverse events associated with the yoga.

¶5% minor adverse events and no serious adverse events.

**Short-term only.

††No adverse events from one study.

CE, cost effectiveness; NR, not reported (by systematic reviews).

Yoga: four high-quality reviews,51 60–62 one prioritised62

Osteopathy: two high-quality reviews,64 65 one prioritised65

Acupuncture: six high-quality reviews,40 41 49 50 53 54 two prioritised—as highest quality40 and most recent50

Spinal manipulation/mobilisation: nine high-quality reviews,40–48 three prioritised (due to differences in types of back pain and interventions included)40 43 44

Spa/balneotherapy: one high-quality review52

Tai chi: one high-quality review54

Osteoarthritis

Moderate/good quality evidence for OA (see table 4) was found as follows:

Table 4

Evidence for osteoarthritis

Comparator	Quality	Population size	Complementary medicine	Safety	Cost-effectiveness
Placebo	Good	Large	Acupuncture92 94*	Some evidence of harm91 92†	NR
Usual care	Moderate	Small	Tai chi99‡	Some evidence of safety98–100§	NR

*Compared to pulsed electromagnetic fields, muscle-strengthening or aerobic exercise, weight loss.

†Up to 45% frequency of minor adverse events.

‡Only in the short term.

§Systematic reviews (SRs) concluded tai chi is safe.

NR, not reported (by SRs).

Acupuncture: six high-quality reviews,91–96 two prioritised92 as a Cochrane review and94 as most recent.

Tai chi: five high-quality reviews,97–101 one prioritised/99

Neck pain

Moderate/good quality evidence for neck pain (see table 5) was found as follows:

Table 5

Evidence for neck pain

Comparator	Quality	Population size	Complementary medicine	Safety	Cost-effectiveness
Active intervention	Good	Medium	Manual therapy87*	Some evidence of potential harm82 83†	Some evidence of CE87 88‡
	Moderate	Large	Manipulation82§¶	Some evidence of potential harm82 83†	Some evidence of CE87 88‡
	Moderate	Small	Manual therapy88**	Some evidence of potential harm82 83†	Some evidence of CE87 88‡
Placebo	Moderate	Medium	Acupuncture89	Some evidence of safety89	No data40
Usual care	Moderate	Medium	Acupuncture89	Some evidence of safety89	No data40

*Manual therapy plus exercise compared with exercise alone.

†Reduced cost of therapy.

‡Minor adverse events and rare but devastating adverse effects from manipulation.

§Compared to medication.

¶Also showed effectiveness for pain compared with any control.

**Mobilisation, manipulation and soft tissue techniques.

CE, cost effectiveness.

Manual therapy/manipulation: eight high-quality reviews,40 82–88 three were prioritised due to differing inclusion criteria.82 87 88

Acupuncture: two high-quality reviews,40 89 one prioritised.89

There was moderate quality evidence in a medium-sized population for a range of other techniques.82 87 88

Fibromyalgia

Acupuncture for fibromyalgia was reviewed by—five high-quality reviews,70–73 81 one prioritised.72 The prioritised review found moderate quality evidence in a small population compared with placebo for stiffness (acupuncture) and for pain and sleep disorders (electroacupuncture). However, there was some evidence that acupuncture is less effective than usual care for MH outcomes in fibromyalgia.72 There is some evidence of safety (no serious adverse events)70–73 81 but none of cost-effectiveness.

Myofascial trigger point pain

Acupuncture for myofascial trigger point pain was reviewed by—three high-quality reviews.106 116 117 The prioritised review found good quality evidence in a medium-sized population compared with placebo and other active interventions for pain in the short and medium term but not in the long term.117 However, safety, cost-effectiveness or impact on MH outcomes were not reported.106 116 117

Lateral epicondylitis

One high-quality review120 found moderate quality evidence for manual therapy (mobilisation with movement) but the comparator and sample size were not reported so we are unable to include it in our synthesis.

Evidence for mental health

No reviews provided good quality evidence for complementary medicine for any MH condition nor any evidence of cost-effectiveness, mainly because the latter was not reported.

Depression

Moderate quality evidence for depression (see table 6) was found as follows:

Table 6

Evidence for depression

Comparator	Quality	Population size	Complementary medicine	Safety	Cost-effectiveness
Active intervention	Moderate	Medium	Meditation125*	Some evidence of safety§ 125	No data124
Usual care	Moderate	Medium	MBSR (mindfulness-based stress reduction)124*	Some evidence of safety§ 125	No data124
	Moderate	Medium	Tai chi134†‡	No data134	No data134

*Also some evidence of effectiveness for both in musculoskeletal (MSK) populations.124 125

†Older adults only.

‡May be less effective in comorbid populations (including some MSK conditions).132

§Nine trials found no adverse events.

Mindfulness/meditation: two high-quality reviews,124 125 both prioritised as included different types of meditation/mindfulness/mindfulness-based stress reduction (MBSR) or populations.

Tai chi: two high quality reviews,132 134 both prioritised as Chi et al 134 only reviewed studies of older people (Liu et al 132 found only poor quality evidence).

Relaxation: one high-quality review129 but comparator and sample size were not reported so it is not included here.

Anxiety

Moderate quality evidence for anxiety (see table 7) was found as follows:

Table 7

Evidence for anxiety

Comparator	Quality	Population size	Complementary medicine	Safety	Cost-effectiveness
Active intervention	Moderate	Medium	Meditation138*†	Some evidence of safety125 137 138	NR
	Moderate	Medium	Yoga136‡	No data135	NR
Placebo	Moderate	Medium	Meditation138*†	Some evidence of safety125 137 138	NR
Usual care	Moderate	Medium	Meditation138*†	Some evidence of safety125 137 138	NR
	Moderate	Medium	MBSR124†	Some evidence of safety125 137 138	NR

*Moving meditation (tai chi, qi gong or yoga) was more effective than static meditation.

†Mindfulness-based stress reduction (MBSR)/meditation was also effective for anxiety (as an outcome) in a musculoskeletal population.124 125

‡For women with breast cancer only.

NR, not reported (by systematic reviews).

Meditation, including moving meditation (yoga, tai chi and qi gong): three high-quality reviews,125 137 138 one prioritised.138

MBSR: one high-quality review.124

Yoga: three high-quality reviews,129 135 136 one prioritised.136

Sleep disorders

Three high-quality reviews considered meditative/mind–body movement for sleep.143–145 The prioritised review found moderate quality evidence in older adults for meditative movement in a medium-sized population, compared with usual care and other active interventions.143 There was some evidence of safety (no adverse events)143 and effect on sleep quality in an MSK population.145 No cost-effectiveness data reported.

Stress and distress

There was one high-quality review of mindfulness for stress and distress.124 For stress, they found moderate quality evidence in a medium-sized population compared with usual care and placebo, and some evidence of effectiveness on stress as an outcome in an MSK population.124 There were no data on cost-effectiveness and no data on safety. For distress, they found moderate quality evidence in a medium-sized population compared with usual care, and some evidence of effectiveness on distress as an outcome in an MSK population.124 There were no data on cost-effectiveness and safety was not reported.

There was only poor quality evidence for all other combinations of MH condition/complementary medicine/comparator.

Evidence for effectiveness in comorbidity/MH outcomes in MSK

Only one SR provided data on comorbid MSK and MH conditions (depression and a variety of comorbid MSK conditions).132 Three SRs provided data on MH outcomes in MSK populations (but did not specify an MH score/diagnosis as inclusion criterion). Tai chi for depression and comorbid conditions (including MSKs) may be less effective than for depression alone132; MBSR/meditation was effective for MH outcomes (depression, anxiety, stress/distress) in an MSK population124 125; meditative/mind–body movement was effective for sleep outcomes in an MSK population.145

Complementary medicine with evidence for both MSK and MH conditions

From the above synthesis, complementary medicine with moderate/good quality evidence for both MSK and MH conditions (from separate SRs) are yoga and tai chi. Yoga appears to be effective for both LBP and anxiety compared with other active interventions, although the anxiety finding was only from studies of women with breast cancer. Meditative movement, which includes yoga, is effective for sleep disorders, although only for older adults (there is also some evidence in chronic pain and fibromyalgia populations). Tai chi appears to be effective compared with usual care for LBP, knee OA and depression, in older people (although it appears less effective in comorbid depressed patients). Meditative movement, which includes tai chi, is effective for sleep disorders, although only for older adults (there is some evidence in chronic pain and fibromyalgia populations as well).

Update

Since completing this review, in July 2017, we searched for recent SRs on the highest ranking topics (results are not included in tables 3-7). These either provide additional evidence for MH outcomes in an MSK population, or evidence supporting our original conclusions.

Yoga: a new Cochrane review for LBP has been published which suggests there is moderate-quality evidence for yoga compared with other active interventions for pain, and for function at 6 months only, although they also found moderate-quality evidence of harm for yoga (primarily exacerbation of back pain).151 They also found very low-quality MH evidence for yoga improving depression (as an outcome not a diagnosis) in LBP (comparator not stated).151

MBSR/mindfulness/meditation: a new SR provides evidence for mindfulness and MH, including high-quality evidence (sample size not reported) for mindfulness helping with depression (compared with all controls—usual care, and active interventions (education, stress management); (as an outcome not a diagnosis) in LBP (and fibromyalgia).152 Another SR of mindfulness for LBP was recently published but cannot be included as there were no conclusions regarding study quality.153

Acupuncture (compared with other active interventions): Two new SRs confirm the effectiveness of acupuncture for OA of the knee, with moderate-quality evidence in a large sample compared with active interventions and placebo.154 155

Discussion

This scoping review has highlighted the large and increasing number of SRs of complementary medicine for MSK and MH disorders, covering 29 different complementary medicine approaches, but identified no high-quality evidence for complementary medicine for MSK–MH comorbid populations. Three reviews showed that MBSR/meditation/meditative movement may improve MH outcomes in MSK populations. Acupuncture, yoga, tai chi and mindfulness/meditation have evidence for both MSK and MH conditions.

We have identified moderate/good-quality evidence (as reported by SR authors) for effectiveness (any comparator) in a moderate/large population for: LBP (yoga, acupuncture, spinal manipulation/mobilisation, osteopathy), OA (acupuncture, tai chi), neck pain (acupuncture, manipulation/manual therapy), myofascial trigger point pain (acupuncture), depression (MBSR, meditation, tai chi, relaxation), anxiety (meditation/MBSR, moving meditation, yoga), sleep disorders (meditative/mind–body movement) and stress/distress (mindfulness). The majority of these complementary approaches had some evidence of safety—only three had evidence of harm—osteopathy for LBP, acupuncture for OA and manipulation/manual therapy for neck pain. There was some evidence of cost-effectiveness for spinal manipulation/mobilisation and acupuncture for LBP, and manual therapy/manipulation for neck pain, but few SRs reviewed cost-effectiveness and many found no data.

Regarding comorbidity, from our original searches, the only evidence for MH outcomes in an MSK population (not necessarily with MH conditions) was for MBSR/meditation/meditative movement/mindfulness. Tai chi and yoga were effective for MSK and MH conditions in separate SRs (not comorbid populations). Our 2017 update found new (but poor quality) evidence for yoga in MSK–MH comorbidity and more evidence for mindfulness in MSK–MH comorbidity.

Some of these studies showed effectiveness compared with an active comparator (spinal manipulation/mobilisation and yoga for LBP; meditation for depression and anxiety), providing some evidence that these interventions have effects over and above non-specific attention.

The quality of the SRs was very variable. We used a quality tool, AMSTAR, which itself has some limitations, as it includes both methodological and reporting items which are not all necessarily equal in importance, and some of which can be difficult to interpret. Future reviews may consider prioritising certain items on AMSTAR, using the new version of AMSTAR,156 or using an alternative critical appraisal tool such as the Oxford Centre for Evidence-based Medicine checklist (https://www.cebm.net/2014/06/critical-appraisal/). An assessment of risk of bias, using, for example, Risk Of Bias In Systematic reviews (ROBIS), may be useful in addition to these quality assessments.

The quality of included trials also varied greatly; it is notable that none of the MH SRs concluded that trial quality was ‘good’ overall. In addition, assessment of individual RCT study quality is likely to have varied between SR authors, a common issue with complementary medicine studies which rarely fit conventional quality criteria (particularly blinding). Very few of the MSK SRs and none of the MH SRs mentioned cost-effectiveness. Safety was rarely adequately reported in SRs or the studies they included. Future SRs in the areas of complementary medicine and MSK/MH need to ensure they include a priori design/protocol registration, a list of excluded studies and conflict of interest statements, and use duplicate study selection and data extraction, assess publication bias and search for grey literature.

Both a strength and limitation of this scoping review was its comprehensive scope which included a wide range of complementary medicine approaches and types of MSK and MH condition, but limited the depth of analysis and discussion.

Other key strengths were the systematic searching techniques, prioritising high-quality SRs using MeSH terms where possible, and search techniques to capture grey and unpublished literature. The whole project team was involved in independently screening the literature search results, although a limitation was that only one author conducted data extraction. We included a quality appraisal of the reviews, although time constraints meant this was an abbreviated assessment. One of the key limitations was the pragmatic decision to only include SRs, which may have led to the exclusion of some complementary medicines, and means results are subject to the limitations of SR and trial methodology, study designs which are sometimes challenging for complementary medicine.157–159 Another key limitation is our definition of complementary medicine as including a practitioner in its delivery (which relied on review authors reporting these details), which excluded over-the -counter products or self-care practices, herbal medicine in particular, but was necessary to limit the scope of the review to complementary medicine which may be appropriate for an integrated care model using referral from a general practitioner. We are aware that by excluding reviews published pre-2005, we may have excluded some topics. We were unable to include non-English publications.

Although complementary medicine is commonly used by patients in the UK,25 it is not widely available via the NHS. Given the high burden of MSK and MH conditions to individuals and society in the UK, and the prevalence of comorbidity, complementary medicine may be worth considering for some patients, where there is evidence of effectiveness, cost-effectiveness and safety.

One of the main aims of this study was to inform the design of a future trial to assess the effectiveness and cost-effectiveness of an integrated medicine approach for patients with comorbid MSK and MH disorders. We will combine the results of this review with the views of conventional and complementary medicine practitioners, the public and NHS providers, to inform the final focus for an RCT. Our proposed trial would be pragmatic and focus on overall effectiveness, so interventions with evidence compared with any comparator are potential candidates.

SRs of complementary medicine consistently conclude that further high-quality trials are needed, with longer follow-up and larger sample sizes, assessing cost-effectiveness and use of appropriate, sensitive, validated outcome measures.

Based on our criteria of study quality, safety, cost-effectiveness and evidence for MSK MH comorbidity, this scoping review suggests that the most promising complementary medicine for future research in the area of MH–MSK comorbidity are: mindfulness (evidence for MH in MSK populations and evidence of safety), yoga (evidence for MSK and MH conditions) and tai chi (evidence for MSK and MH conditions, evidence of safety). The comorbid populations that would benefit most are less clear, as so few reviews focus on comorbidity, but may include LBP, OA, depression, anxiety and sleep.