NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Guideline Centre (UK). Low Back Pain and Sciatica in Over 16s: Assessment and Management. London: National Institute for Health and Care Excellence (UK); 2016 Nov. (NICE Guideline, No. 59.)

  • Guideline updates: December 2020: in the recommendation on stopping opioid analgesics NICE added links to other NICE guidelines and resources that support discussion with patients about opioid prescribing and safe withdrawal management. September 2020: NICE's original guidance on low back pain and sciatica in over 16s was published in 2016. It was partially updated in September 2020. See the NICE website for the guideline recommendations and the evidence review for the 2020 update. This document preserves evidence reviews and committee discussions for areas of the guideline that were not updated in 2020.

Guideline updates: December 2020: in the recommendation on stopping opioid analgesics NICE added links to other NICE guidelines and resources that support discussion with patients about opioid prescribing and safe withdrawal management. September 2020: NICE's original guidance on low back pain and sciatica in over 16s was published in 2016. It was partially updated in September 2020. See the NICE website for the guideline recommendations and the evidence review for the 2020 update. This document preserves evidence reviews and committee discussions for areas of the guideline that were not updated in 2020.

Cover of Low Back Pain and Sciatica in Over 16s: Assessment and Management

Low Back Pain and Sciatica in Over 16s: Assessment and Management.

Show details

9Exercise therapies

9.1. Introduction

Exercise therapies make use of various forms of physical exercise to prevent or treat low back pain. The term ‘exercise therapy’ encompasses a wide range of different exercise types, environments and theoretical models. What they have in common is the engagement of the person with a programme of physical exercise that the person is encouraged to perform on a regular basis.

Exercise therapy may be delivered by a range of healthcare professionals, on a one to one basis or in a group environment. The focus may vary from exercise using specialist gym equipment to exercises conducted at home or in the outdoor environment. Exercise may be directed at improving a variety of parameters of fitness and function including muscle strength, timing or endurance, flexibility and range of motion, precision of movement, cardiovascular fitness, functional task performance and confidence.

Biomechanical exercise includes any exercise intervention that is primarily directed at altering or improving spinal mechanics. This includes muscle strengthening, stretching, range of motion exercise, motor control exercise (including core stability programmes and Pilates) or programmes aimed at addressing specific problem movements (including McKenzie exercise and the Feldenkrais method).

Aerobic exercise includes any exercise intervention that is primarily directed at improving cardiovascular fitness and endurance.

Mind–body exercise includes any exercise intervention that includes a combined physical, mental and spiritual focus, often with connection to metaphysical and cultural philosophies. Examples include the various forms of Yoga and Tai Chi.

Mixed modality exercise includes any exercise intervention that incorporates a combination of any of the previous three categories.

9.2. Review question: What is the clinical and cost-effectiveness of exercise therapies in the management of non-specific low back pain and sciatica?

Table 70PICO characteristics of review question

Population
  • People aged 16 or above with non-specific low back pain
  • People aged 16 or above with sciatica
Intervention(s)Individual/group exercise:
  • Mind-body exercises (Yoga, Tai-Chi)
  • Biomechanical (Pilates, core stability, McKenzie, motor control, stretching, Feldenkrais)
  • Aerobics (swimming, walking programme, aerobic exercise)
  • Mixed modality exercise (aerobics and/or mind-body and/or biomechanical)
Comparison(s)
  • Placebo/Sham/Attention control
  • Usual care/waiting list
  • To each other
  • Any other non-invasive interventions in the guideline
  • Combination of interventions: any combination of the non-invasive interventions in the guideline
OutcomesCritical
  • Health-related quality of life (for example, SF-12, SF-36 or EQ-5D).
  • Pain severity (for example, visual analogue scale [VAS] or numeric rating scale [NRS]).
  • Function (for example, the Roland-Morris disability questionnaire or the Oswestry disability index).
  • Psychological distress (HADS, GHQ, BPI, BDI, STAI)
Important
  • Responder criteria (> 30% improvement in pain or function)
  • Adverse events:
    1. morbidity
  • Healthcare utilisation (prescribing, investigations, hospitalisation or health professional visit)
Study designRandomised controlled trials (RCTs) and systematic reviews (SRs). If insufficient evidence is identified, observational studies will be included.

9.3. Clinical evidence

9.3.1. Summary of included studies – single interventions

A search was conducted for randomised trials comparing the effectiveness of exercise therapies (mind-body exercises, biomechanical exercise, aerobic exercise, and mixed modality exercises) with either placebo, usual care, or other non-invasive treatments in the management of people with low back pain or sciatica.

Seventy-five randomised trials were identified from a total of 80 papers.8,12,26,37,55,67,71,72,79,80,84,93,98,104,112,132,134,138,161,168,170,173,180,194,196,199-201,211,233,265,277,278,282,295,309,313,317,318,324,325,329,331,334,335,342,348,352,364,124,125,370,371,379,395,400,410,414,416,419,430,436,447-449,457-462,466,473 Details of these studies are summarised in Table 71, Table 72, Table 73 and Table 74 below. Evidence from the study is summarised in the clinical evidence summary below (see section 9.3.5 to (a)). See also the study selection flow chart in Appendix E, study evidence tables in Appendix H, forest plots in Appendix K, GRADE tables in Appendix J and excluded studies list in Appendix L.

The Smeets 2006 trial460 (Smeets 2008459, Smeets 2009457, Smeets 2006461, Smeets 2008458) reported data from 4 arms (exercise, cognitive behavioural approaches, exercise and cognitive behavioural approaches/MBR, and waiting list control). The data extracted in this review was for the exercise versus cognitive behavioural approaches and exercise versus waiting list control. The data for cognitive behavioural approaches versus waiting list is in the psychological review, and the data for the combination arm (exercise and cognitive behavioural approaches) is in the MBR review (see section 17).

Data from Aboagye et al. 2015 was excluded as data was not interpretable due to the number of participants in each group not being provided, therefore effect size could not be estimated.3

Evidence of cognitive therapy compared to mixed exercise (biomechanical and aerobic), and behavioural therapy compared to aerobic exercise was identified and analysed in chapter 17. This review only considered supervised exercise programmes. Unsupervised exercise was considered as self management, and therefore included in the self management review.

9.3.2. Summary of included studies – combined interventions (exercise therapy adjunct)

Sixteen studies looking at combinations of non-invasive interventions (with exercise therapy as the adjunct) were also included in this review.66,97,106,115,115,281,301,309,328,341,342,411,429,479,496,499,535,561 These are summarised in Table 75 below. Evidence from these studies is summarised in the GRADE clinical evidence profile/clinical evidence summary below (see section (a)). See also the study selection flow chart in Appendix E, study evidence tables in Appendix H, forest plots in Appendix K, GRADE tables in Appendix J and excluded studies list in Appendix L.

Ding et al. 2015 had no outcomes relevant to the review protocol to be extracted.115

Szulc et al. 2015 reported data from 3 arms (exercise with self-management and manual therapy, exercise and self-management, and TENS with laser, massage and self-management). The data extracted in this review was for the exercise and self-management versus TENS with laser, massage and self-management comparison. The data for exercise with self-management and manual therapy versus exercise and self-management, and exercise with self-management and manual therapy versus TENS with laser, massage and self-management was analysed in chapter 17.

Two Cochrane reviews244,552 were identified but could not be included for the following reasons:

  • The review was not limited to RCTs;244
  • The population was stratified by chronicity of pain (acute: less than 6 weeks; subacute, 6-12 weeks; chronic, greater than 12 weeks).552

The studies included in these Cochrane reviews were individually assessed and included if they matched the review protocol.

9.3.4. Data unsuitable for meta-analysis

Table 76. Group exercise

Table 77. Biomechanical exercise

9.3.9. Combinations – exercise therapy adjunct

9.4. Economic evidence

Published literature

One economic evaluation was identified that included mind and body exercise as a comparator and has been included in this review.87 This is summarised in the economic evidence profile below (Table 126) and the economic evidence table in Appendix I.

Table 126. Economic evidence profile: Mind/body exercise interventions.

Table 126

Economic evidence profile: Mind/body exercise interventions.

One economic evaluation was identified that included mixed modality exercise as a comparator and has been included in this review.457 This is summarised in the economic evidence profile below (Table 127) and the economic evidence table in Appendix I.

Table 127. Economic evidence profile: Mixed modality exercise interventions.

Table 127

Economic evidence profile: Mixed modality exercise interventions.

No relevant economic evaluations were identified that included biomechanical exercise or aerobic exercise compared to placebo or sham, usual care or other single active interventions in the protocol. Three economic evaluations were identified that included biomechanical exercise as a comparator (Critchley 2007,94 Beam 2004,498 and Niemisto 2003 and 2005387,388) and this was part of the following interventions: 1) biomechanical exercise in combination with self-management or self-management and manual therapy (mixed modality), or self-management, biomechanical exercise and manual therapy (mixed modality) compared to self-management alone (Beam 2004498); 2) biomechanical exercise compared to mixed modality manual therapy plus self-management or compared to MBR programme ( Critchley 200794) 3) biomechanical exercise in combination with manual therapy (manipulation/mobilisation) and self-management compared to self-management alone (Niemisto 2003388/2005387).

One economic evaluation relating to biomechanical exercise, one relating to a mixed exercise intervention, and one relating to mind-body exercise were identified but excluded due to limited applicability and/or potentially serious methodological concerns.3,210,444 These are listed in Appendix M, with reasons for exclusion given.

See also the economic article selection flow chart in Appendix F.

Table 128. Economic evidence profile: biomechanical exercise

Table 129. Economic evidence profile: biomechanical exercise

Table 130. Economic evidence profile: biomechanical exercise

Unit costs

Biomechanical and aerobic exercise interventions are generally conducted in group or individually by a physiotherapist. The relevant unit costs are provided below to aid consideration of cost-effectiveness.

Table 131Unit costs of healthcare professionals

Healthcare professionalCosts per hour
Hospital physiotherapist (band 5)£32
Community physiotherapist (band 5)£30

Source: PSSRU 201399

The unit costs of community physiotherapists do not account for travel costs, such as mileage and travel time. As a result, these estimates are probably an underestimate.

Mind and body exercise interventions are not currently provided by the NHS. These types of interventions are conducted by a therapist (for example, yoga instructor) rather than a physiotherapist. No published unit costs were identified.

Mind and body exercise interventions are not currently provided by the NHS. These types of interventions are conducted by a therapist (for example, yoga instructor) rather than a physiotherapist. No published unit costs were identified although the economic evaluation included in the review estimated the costs of yoga per person in the study to be £292.61. This included teaching and equipment costs for up to 12 group sessions (maximum 15 participants) of 75 minutes. They also noted that costs would be reduced if an NHS physiotherapist ran the class.

The cost of exercise interventions will be based on:

  • The number of sessions required
  • The length of each session
  • The number of people each session is for
  • The cost of the person who would provide the session
  • The cost of any equipment or facilities required as part of the intervention.

9.5. Evidence statements

9.5.1. Clinical

9.5.1.1. Individual biomechanical exercise versus usual care

In the mixed population, individual biomechanical exercise showed a clinically important improvement compared with usual care for improvement of quality of life scores on all but one of the reported domains (2 studies; low and very low quality; n = 57), and psychological distress (1 study; very low quality; n = 54). No clinically important benefit was seen for short term pain in 5 studies (moderate quality; n = 317), however there was a clinically important benefit in pain at rest, pain during movement, and pain when walking (3 studies; moderate, moderate and very low quality; n = 30, 30 and 32 respectively). No clinical benefit was seen for longer term pain intensity (1 study; low quality; n = 99), function long term (2 studies; low quality; n = 159), or short term function (5 studies; low quality; n=253).

For this comparison in people with sciatica, there was a clinically important improvement in short-term pain (1 study; very low quality; n = 52) in those in the exercise group, but no other outcomes were reported that were relevant to this review.

In people without sciatica, there was a clinically important improvement in short term physical and mental quality of life in those undertaking biomechanical exercise compared with usual care (2 studies; low quality; n = 99). Evidence also showed a clinically important benefit for 5 other quality of life domains in the short term, and all quality of life domains in the long term (low and very low quality; 1 study; n = 60). There was a clinically important benefit in terms of short term pain from 6 studies, which could not be meta-analysed (very low and low quality; n = 17-246), however 4 studies found no benefit for this outcome (low quality; n = 260). A clinically important benefit was observed for long term pain (very low quality; 2 studies; n = 146), however further evidence that could not be pooled in the meta-analysis showed no clinically important benefit (low quality; 1 study; n = 271). Evidence for function was mixed, with evidence for a clinically important benefit for short term and long term function (2 studies; moderate quality evidence; n = 86 and very low quality; n = 60, respectively). However, evidence from 8 studies demonstrated no clinically important benefit for function at short term and long term (low and very low quality; n = 17 – 418). No evidence was available for psychological distress. Fewer adverse events were reported in those that received usual care than biomechanical exercise although only from 1 small study (very low quality; n= 40).

9.5.1.2. Individual biomechanical exercise versus active control

Evidence for individual biomechanical exercise compared with self-management, spinal manipulation, and interferential therapy was identified, mostly from small individual studies and of low or very low quality. The evidence only showed clinical benefit for biomechanical exercise for long-term leg pain (1 study; low quality; n = 71) and long-term function (1 study; very low quality; n= 71) when compared to self-management. The evidence also showed a clinical benefit of biomechanical exercise for long term, but not short term, physical quality of life when compared to spinal manipulation (1 study; low quality; n = 164). Clinical benefit of biomechanical exercise was also seen for short-term pain (1 study; moderate quality; n= 60) when compared to interferential therapy.

9.5.1.3. Group biomechanical exercise versus usual care

In the mixed population, when compared to usual care, a clinically important benefit of biomechanical exercise was demonstrated for pain in evidence from 1 study in the long term, but not in the short term (very low quality, n = 127). However, a short term clinically important benefit of pain for biomechanical exercise was suggested using core stability (1 study; moderate quality, n = 40).

In the population with low back pain without sciatica, a clinically important benefit of biomechanical exercise was found for physical and mental quality of life, when compared with usual care (1 study; moderate quality; n = 18). No clinical difference was demonstrated for short term pain, however there was a clinically important benefit for function (2 studies; very low quality; n = 52).

No evidence was available for psychological distress.

9.5.1.4. Group biomechanical exercise versus active comparators

One study compared supervised with unsupervised exercise in the mixed population, and demonstrated a clinical benefit of the supervised sessions for reducing pain intensity in the longer term but not the short term (very low quality; n = 170 and 141 for short and long term).

No evidence was available for other comparisons, populations or outcomes.

9.5.1.5. Individual aerobic exercise versus usual care

In the mixed population no clinical benefit was observed for pain or function (low quality; 1 study; n = 46). Other outcomes were not reported. However, in people without sciatica a clinical benefit of exercise was seen in terms of reducing pain intensity in the short and longer term in 1 study of deep water running (low and moderate quality; n = 49), but not in studies of treadmill walking or running (very low and low quality; n = 37 and 57). Aerobic exercise was also shown by 2 studies to improve short-term function (low quality; n = 86), but not psychological distress or quality of life (very low and low quality; n = 37and 57).

No evidence was available for the placebo comparison, nor for the sciatica population.

9.5.1.6. Individual aerobic exercise versus active comparators

One study compared individual aerobic exercise with individual biomechanical exercise in the mixed population and demonstrated no clinically important benefit for function (low quality; n = 52). Another small study compared individual aerobic exercise to group biomechanical exercise showing benefit for group biomechanical exercise at less than or equal to 4 months for SF-36 both physical and mental components, and greater than 4 months for average back pain (very low quality, n=30). No clinically significant difference was observed for depression or anxiety measured using HADS or other pain outcomes at either short or longer-term measures.

No evidence was available for other comparisons, populations or outcomes.

9.5.1.7. Group aerobic exercise versus usual care

A clinically important benefit of physical and mental quality of life was observed for group aerobic exercise when compared with usual care in people with low back pain without sciatica (2 studies; very low quality; n = 109). A clinical benefit was also found for two of the individual quality of life domains (very low quality; n = 20). No clinical benefit was observed for any exercise in any other the other critical outcomes (low and very low quality; range of n = 40-119).

No evidence was available for the placebo comparison or for the sciatica population.

9.5.1.8. Group aerobic exercise versus active comparators

When compared with self-management, a clinically important improvement in pain in the overall population was observed (1 study; very low quality; n = 18). No other outcomes were reported. When compared to group biomechanical exercise, no clinical benefit of group aerobic exercise was found for any of the critical outcomes (very low quality, n = 83-91).

One further study in the low back pain population without sciatica compared group aerobic exercise with group biomechanical exercise reported evidence demonstrating a clinical benefit for pain in the short term but not the long term for the group receiving aerobic exercise. No clinical benefit was found for function in either the short-term or long term (low quality; n = 64).

No evidence was available for other comparisons, populations or outcomes.

9.5.1.9. Individual mind-body exercise versus biomechanical exercise

Evidence from 1 small study showed short-term clinical benefit of yoga when compared to biomechanical exercise on pain and function (low quality; n= 30), whereas another study demonstrated no clinically important difference between tai chi and biomechanical exercise on short-term pain outcome (low quality; n= 40).

9.5.1.10. Group mind-body exercise versus usual care

In the people with low back pain with or without sciatica, evidence from 2 studies suggested a benefit in terms quality of life on EQ-5D for group mind-body exercise when compared with usual care at the short term (low quality; n = 325), but further evidence did not demonstrate benefit in the longer term (1 study; moderate quality; n = 313) and no clinical difference was seen at either time point when quality of life was assessed by SF12 in the same studies (moderate quality; n = 326, 313). In terms of pain, a clinical benefit with Iyengar yoga was seen when compared to usual care at greater than 4 months, but no clinical difference at less than or equal to 4 months (1 study; very low quality, n= 90). The same applied when hatha yoga was compared to usual care at either short term (2 studies, very low quality; n= 82) or longer-term (low quality; n=23). A benefit was seen for psychological distress for hatha (low and very low quality; n = 46 and 16) but not lyengar yoga (moderate to very low quality; n = 418 and 96). Whereas no clinical difference of yoga was seen was by 6 studies for short-term function time points (low quality; n= 516) or by 3 studies for longer term time points (low quality; n= 426).

For the population without sciatica, a clinically important benefit in pain reduction in the short and longer term was found for group mind-body exercise when compared with usual care in a single study (very low quality; n = 42).

No evidence was available for the placebo comparison or for the population with sciatica.

9.5.1.11. Group mind-body exercise versus active comparators

In the low back pain population without sciatica, when compared with self-management, a clinically important benefit in short-term and long-term function was identified (2 studies; low and very low quality; n = 164). When compared with group mixed exercise, no clinically important difference between treatments was demonstrated for this outcome (2 studies; moderate and very low quality; n = 164).

In a mixed population of people with low back pain with or without sciatica, group mind-body exercise showed clinical benefit for pain at both short and long term when compared to individual biomechanical exercise in a single study (moderate quality, n= 60)

9.5.1.12. Individual mixed exercise versus waiting list

For this comparison in people with sciatica, there was a clinically important improvement in short-term back pain and leg pain (1 study; low quality; n = 30) in those in the exercise group. No other relevant outcomes were reported.

9.5.1.13. Individual mixed exercise versus active comparators

Evidence for individual mixed exercise compared to unsupervised exercise from a single study in the overall population demonstrated a clinically important reduction in pain for individual mixed exercise in the longer-term (low quality; n = 40). No other outcomes or time-points for the comparison of individual mixed exercise compared to unsupervised exercise were reported.

No clinical difference between mixed exercise or biomechanical exercise was observed in terms of short term pain or function (1 study; moderate quality; n= 63).

9.5.1.14. Group mixed exercise versus usual care

When compared with usual care in the low back pain population there was no clinical benefit for function (2 studies; very low quality; n = 88). There was evidence of no clinical benefit of short term pain (1 study, low quality; n = 29), however a clinical benefit in favour of mixed exercise compared to usual care was observed (1 study; very low quality; n = 59). A benefit in terms of psychological distress measured using the HADS depression score, but not for the HADS anxiety score was observed (very low quality; n = 29). Additionally, 3 of the 8 domains of quality of life (general health, physical role and emotional role) showed a benefit of group mixed exercise (1 study; very low and low quality; n = 36).

When compared with usual care in the population with sciatica, the evidence was conflicting. A benefit of group mixed was seen for pain in the long-term, but for function in the short and long term a benefit was seen for usual care (1 study; low and very low quality; n = 44).

In people with low back pain with or without sciatica, clinical benefit in favour of exercise was demonstrated compared with usual care in the short and long-term for pain from small studies of population size less than 100 (moderate to very low quality), clinical benefit was also seen for function at less than or equal to 4 months from 2 small studies (low quality; n= 52). One study showed no clinical benefit for psychological distress (low quality; n = 29). Another small study (n= 38) demonstrated conflicting evidence for quality of life, with clinical benefit of mixed exercise on SF-35 mental (moderate quality) but no difference on SF-36 physical (low quality) when compared to usual care.

9.5.1.15. Group mixed exercise versus active comparators

In the population with low back pain without sciatica, evidence from 1 study suggested a clinical benefit for group mixed exercise for short term function (low quality; n = 21), psychological distress (low quality; n = 21), and both long term (very low quality; n = 27) and short term pain (very low quality; n = 21), when compared with cognitive therapy. Quality of life was not reported. There was no placebo/sham evidence for the mixed or sciatica populations. No clinically important benefits for mixed exercise were found when compared with self-management for function (2 studies; moderate to low quality; n = 125 and 164) or when compared with cognitive behavioural approaches in the overall population for pain, function or psychological distress (1 study; low and very low quality; n = 104).

No evidence was available for other comparisons, populations or outcomes.

9.5.1.16. Combinations of interventions – exercise therapy adjunct

The evidence (ranging from very low to moderate quality) showed that there was no clinical difference for nearly all outcomes and nearly all combinations of non-invasive interventions that had exercise therapy as an adjunct, with a few exceptions.

A single study in a low back pain population comparing exercise (biomechanical and aerobic) and electrotherapy (PENS) compared to sham electrotherapy (PENS) demonstrated evidence of clinical benefit favouring sham PENS for quality of life outcomeSF-36 physical, but clinical benefit for short-term pain ( low quality; n=93). Comparing exercise (biomechanical and aerobic) and electrotherapy (PENS) to electrotherapy (PENS) showed clinical benefit for short and longer-term quality of life SF-36 physical outcomes in a single study in a low back pain population (very low quality; n= 92).

A study in a low back pain population demonstrated clinical benefit of cognitive behavioural approaches and self-management (education) over aerobic exercise, cognitive behavioural approaches and self-management (education) on short-term function (very low quality; n= 27).

Combining biomechanical exercise with self-management in a low back pain population showed clinical benefit when compared to self-management on short-term pain (1 study; very low quality; n= 86) and, short and long-term function (1 study; very low quality; n= 86).

In a mixed population of people with low back pain with or without sciatica, combining exercise with self-management demonstrated clinical benefit on long-term number improving on function (1 study; low quality; n= 90), quality of life index (1 study; low quality; n= 90), short term physical quality of life (1 study; low quality; n = 418) and long-term pain (low quality; 1 study; n= 83) when compared to self-management. Benefit of biomechanical exercise and manual therapy was seen over manual therapy alone in a single study on short-term pain (low quality; n= 92), and over combined self-management and manual therapy in one study on physical quality of life, long term mental quality of life and short term but not long term sensory and affective pain (very low quality, n = 25).

In the population with sciatica, the combination of biomechanical exercise with self-management (unsupervised exercise) demonstrated a clinically important benefit for short term pain and function, when compared to a combination of TENS, laser, massage and self-management (1 study; moderate quality; n = 40).

9.5.2. Economic

  • No relevant economic evaluations were identified relating to individual mind-body exercise in people with low back pain or sciatica.
  • One cost-utility analysis found that group mind-body exercise + usual care was cost effective compared to usual care alone for low back pain (with or without sciatica) (ICER: £13,606 per QALY gained). This analysis was assessed as partially applicable with potentially serious limitations.
  • No relevant economic evaluations were identified relating to individual or group aerobic exercise in people with low back pain or sciatica.
  • No relevant economic evaluations were identified relating to individual or group biomechanical exercise in people with low back pain or sciatica.

One cost-utility analysis found that group mixed modality exercise (biomechanical + aerobic) was dominated (more costly and less effective) by cognitive behavioural approaches for treating low back pain (with or without sciatica). This analysis was assessed as partially applicable with potentially serious limitations.

No relevant economic evaluations were identified relating to individual mixed modality exercise in people with low back pain or sciatica.

  • One cost-utility analysis found that biomechanical exercise was dominated (more effective and less costly) by a 3 element MBR programme (physical, psychological, educational) for treating low back pain (without or without sciatica). This analysis was assessed as partially applicable with potentially serious limitations.
  • One cost-utility analysis for the treatment of low back pain without sciatica found that:
    • the combination of manual therapy and self-management was the most cost-effective compared to a combination of biomechanical exercise, mixed modality manual therapy and self-management, biomechanical exercise in combination with self-management, and self-management alone (ICER: £8,700 per QALY gained when compared to the combination of self-management, biomechanical exercise, and manual therapy). It also found that the combination of biomechanical exercise and self-management was dominated (more effective and less costly) by the combination of biomechanical exercise, manual therapy and self-management.
    • if manual therapy (manipulation) is not available, the combination of biomechanical exercise and self-management was cost effective compared to self-management alone (ICER: £8,300 per QALY gained).

This analysis was assessed as partially applicable with minor limitations.

  • One cost-consequence analysis was identified relating to mixed modality manual therapy in combination with self-management and biomechanical exercise in people with low back pain or sciatica: the combination did not show any statistically significant increase in costs or outcomes compared to self-management (education and advice to stay active). This was assessed as partially applicable with potentially serious limitations.

9.6. Recommendations and link to evidence

Recommendations
8.

Consider a group exercise programme (biomechanical, aerobic, mind–body or a combination of approaches) within the NHS for people with a specific episode or flare-up of low back pain with or without sciatica. Take people's specific needs, preferences and capabilities into account when choosing the type of exercise.

Relative values of different outcomesThe GDG agreed that the most critical outcomes for decision making would be health-related quality of life; with pain severity, function and psychological distress being individually critical outcomes as well as components of quality of life measures.
Adverse events were considered important for decision making because experience of adverse events may outweigh the possible benefits gained from an exercise therapy, similarly, any differences in healthcare utilisation was considered an important outcome likely to reflect any benefits in quality of life experienced. Mortality was not considered as a relevant treatment related outcome for this review and so was not included in the protocol.
The GDG discussed the importance of responder criteria as an outcome and agreed that although important in decision making, due to the inherent difficulties in dichotomising continuous outcomes this was not a critical outcome.
Trade-off between clinical benefits and harmsThe GDG discussed the necessity of a body of evidence to show specific intervention effects, that is, over and above any contextual or placebo effects. It was therefore agreed that if placebo or sham-controlled evidence is available, this should inform decision making in preference to contextual effects. However, if there was a lack of placebo or sham-controlled evidence, evidence against usual care will be given priority when decision making.
Although some trials were identified that had sham exercise as a comparator, on consideration of these, the GDG agreed none met the protocol criteria for appropriate sham interventions for this review. Some shams were interventions being considered elsewhere within the guideline, and are considered under the relevant comparators, whereas others were comparing different forms of exercise and have been excluded. There was consequently no evidence available for exercise compared to placebo/sham.
The GDG noted that there was some evidence of benefit for all exercise types compared to usual care or other active comparators, but no clear evidence for one type being superior to another and benefits were seen inconsistently across critical outcomes. The GDG agreed that there are known benefits to general health and wellbeing from exercise and whilst data on adverse events was very limited there was no evidence of harm and exercise, conducted appropriately, should be safe. The GDG agreed that there was both uncertainty around the effect size and the clinical importance of the comparisons supporting aerobic exercise for low back pain with or without sciatica. They discussed and agreed that aerobic exercise has many additional health benefits and therefore, would not discourage anyone from partaking in such exercise programmes, but were not able to support a recommendation for aerobic exercise alone to be specifically offered by the NHS ahead of other forms of exercise as a treatment for low back pain or sciatica from the evidence reviewed.
Mind-body exercise, such as yoga, showed some clinically important benefits in pain and function but with inconsistency across trials, outcome measures and time points. As with individual biomechanical exercise, some improvements in quality of life were observed, but due to methodological concerns regarding the trial designs, the GDG were not confident in the effect. No evidence was found for the use of mind-body exercise in the sciatica population.
Similarly for mixed exercise, some clinically important benefits in pain, function and quality of life were found compared to usual care/waiting list. The evidence for the sciatica population was inconsistent, showing a benefit in pain reduction, but deterioration in function.
Overall, the GDG felt that there was evidence of clinically important effects for critical outcomes, such as health-related quality of life, pain and function although noted the variability in comparators and study designs made it difficult to clearly determine which form of exercise was most beneficial. The GDG considered that the effect of exercise compared with usual care or self-management could be due, at least in part, to an imbalance of therapeutic attention inherent to such trials and may not necessarily or solely reflect a specific effect of the exercises given, particularly when waiting list controls were used as the comparator groups.
The GDG agreed that there was insufficient evidence that one form of exercise was superior to another and a recommendation for a specific exercise modality was not supported from the current evidence base. However they agreed that the evidence compared to usual care did show that exercise is likely to be of value, although with some uncertainty about the effect size. In the absence of a feasible sham control, the GDG agreed that this was sufficient evidence for a recommendation to consider exercise should be made for people with low back pain with or without sciatica.
Trade-off between net clinical effects and costsIndividual mind-body exercise
No economic evaluations were identified. The cost of providing this intervention will largely depend on the number of sessions provided but individual sessions will be more costly that group sessions. There was no evidence regarding the clinical benefit of individual sessions either compared to usual care or group sessions.
Group mind-body exercise
One relevant economic evaluation was included that considered yoga as an adjunct to usual care in a mixed population of low back pain with or without sciatica. This was based on the RCT reported by Tilbrook and colleagues included in the clinical review. This within-trial analysis found that the addition of yoga to usual care increased costs and improved health (increased QALYs) with an incremental cost-effectiveness ratio of £13,606 per QALY gained. The probability cost effective was 72% at a £20,000 cost effectiveness threshold. This study suggests that group mind-body exercise may be a cost-effective intervention for the NHS because, compared with usual care, the additional health benefits appear to justify the additional costs. However, other treatment options (for example, other exercise modalities, acupuncture, spinal manipulation and pharmacological treatment) are not included in the analysis and so we cannot tell from this if yoga is the most cost-effective option available.
The economic evaluation included in the review estimated the costs of yoga per person in the study to be £292.61. This included teaching and equipment costs for up to 12 group sessions (maximum 15 participants) of 75 minutes. They also noted that if the yoga teaching fee in the trial was replaced with the cost of teaching by a physiotherapist (£38 per hour) with a resulting cost per patient of £63, assuming the participant buys their own yoga mat, manual and CD, the probability of yoga intervention being cost effective increased from 72% to 88%.
This analysis only reflects the effectiveness evidence from one RCT of mind-body exercise whereas a number were included in the clinical review. In this study people received up to 12 group sessions of yoga (75 minutes, maximum 15 participants) over 12 weeks and benefits to patients in terms of QALYs were evaluated over one year. Across the studies included in the clinical review the majority of studies had a similar intensity (range 4 to 48 sessions) and treatment duration (range 4 to 24 weeks). One other study (reported by Cox and colleagues) also reported EQ-5D with a smaller benefit at 12 weeks but is a much smaller study with only short term outcomes.
Biomechanical exercise
One relevant economic evaluation was included that compared biomechanical exercise to manual therapy plus self-management and to MBR in a mixed population of low back pain with or without sciatica. In this study MBR was the least costly and more effective strategy, therefore biomechanical exercise was a dominated option.
Some evidence was available for biomechanical exercise in combination. The economic evaluation based on the UK BEAM study found that biomechanical exercise in combination with self-management was cost effective compared to usual care.498 However, when compared to other active interventions spinal manipulation plus self-management was the most cost effective option. This suggests that biomechanical exercise may be cost effective if spinal manipulation is not an option but when both are available spinal manipulation would be a more cost effective treatment than biomechanical exercise (in combination with self-management).
Individual aerobic exercise
No economic evaluations were identified. The cost of providing this intervention will largely depend on the number of sessions provided but individual sessions will be more costly than group sessions. As the clinical evidence did not show any clear benefit for individual aerobic exercise, the GDG considered this intervention unlikely to be cost effective.
Group aerobic exercise
No economic evaluations were identified. The cost of providing this intervention will largely depend on the number of sessions and the number of people per group. The clinical evidence did not show any clear benefit for group aerobic exercise, however considering the lower cost of group exercise compared to individual exercise, the GDG concluded there was uncertainty around the cost effectiveness of this intervention and it could be recommended as part of an exercise programme.
Individual mixed exercise
No economic evaluations were identified. The cost of providing this intervention will largely depend on the number of sessions provided but individual sessions will be more costly that group sessions. The clinical evidence showed no benefit associated with this intervention, therefore the GDG considered it unlikely to be cost effective.
Group mixed exercise
One relevant economic evaluation (Smeets 2009457 based on the clinical trial Smeets 2006A461) was included that considered group mixed modality exercise (biomechanical + aerobic) was dominated (more costly and less effective) by cognitive behavioural approaches for treating low back pain (with or without sciatica). This analysis only reflects the effectiveness evidence from one RCT of mixed modality exercise comprising biomechanical and aerobic exercise. The rest of the body of evidence showed some clinical benefit for group mixed exercise for pain when compared with placebo/sham. When compared to usual care there was benefit for both long and short term pain, short term function, HADs depression and for 3 of the 8 domains of quality of life. There was also evidence of some benefit on pain at both short and long-term, and function at short-term over usual care in the mixed population.
When compared with usual care in the population with sciatica, there was a clinically important benefit in pain in the long-term, but not short term, and benefit favouring usual care for function in the short and long term. In the overall population, clinical benefit was demonstrated in the short and long-term for pain, and in the short-term for function. For this reason, the GDG considered that group mixed exercise could be cost effective compared to usual care.

Summary
The GDG concluded that there was uncertainty about the cost effectiveness of exercise programmes. There will be a cost to the NHS of providing exercise programmes for people with low back pain and sciatica; this will largely depend on the number of sessions provided and whether delivered as a group or individually. If exercise programmes are effective, upfront costs may be offset by downstream cost savings due to reduced healthcare utilisation or may be justified due to the benefits to the patient. As described in the previous section, the GDG concluded that overall exercise programmes were likely to be of benefit to people with low back pain and that while the evidence varied between specific types of exercise they did not feel that the evidence was sufficient to support a strong recommendation with regards the optimal type, dose or duration of any exercise programme. They also noted that exercise has well established benefits to health beyond any effect seen in the outcomes for treating low back pain. Given this the GDG concluded that despite the uncertainties it was likely that the benefits of exercise to people with a specific episode or flare-up of low back pain with or without sciatica would justify the costs.
Costs of delivering group exercise will be lower than costs of delivering individual exercise therapy. Given the additional cost and uncertainties regarding benefits of individual exercise, it was considered appropriate to recommend group exercise.
Quality of evidenceQuality of evidence in the review ranged from a GRADE rating of moderate to very low. No studies included in the review were assessed as being at low risk of bias, reflecting the inherent difficulty of ensuring plausible blinding to exercise interventions and therefore, the risk of overestimating effects in subjective outcomes, such as pain, function and quality of life. It was also noted that the trials were relatively short term in nature, with the average exercise intervention lasting just 9.5 weeks.
In relation to the difficulties of ensuring blinding in such trials, the quality of evidence could be considered as the best possible for these interventions. The GDG considered the likelihood of effects occurring in exercise groups due to contextual factors, such as the attention given by the therapist or the expectation of success of an active treatment that might explain, at least in part, the observed effects to the likelihood of over-estimating the effect. There were also comparisons with waiting list controls included in the review, which were further down-graded for risk of bias due to the likelihood of over-estimating the effect.
The GDG recognised the difficulties in splitting the comparisons, as well as the group and individual exercise programmes, thereby creating numerous comparisons and outcomes with fewer studies in each. However, the GDG agreed that the pooling of studies with widely differing interventions, despite strengthening the body of evidence, would make it difficult to draw a conclusion about what type of exercise to offer, and to which populations.
The economic evidence was assessed as partially applicable with potentially serious limitations.
Other considerationsFor recommendations on manual therapy, psychological interventions and multidisciplinary biopsychosocial rehabilitation programmes, please see chapters 12, 15, and 17, respectively.
The GDG noted that currently exercise is offered within the NHS, most commonly delivered by physiotherapists. The type of exercise currently offered to people is very variable and depends on the person's preferences, their health care professional's preferences, the local availability of different exercise interventions as well as local commissioning policy. The local provision may include elements of some or all of the types of exercise considered in this review, and may be delivered individually or in a group environment. The recommendation to consider offering exercise in a group environment was based on the likely cost savings of that approach and the lack of clear evidence for the superior efficacy of individually delivered exercise. However the GDG discussed that there are various instances where group exercise may not be suitable or acceptable for the patient and the GDG recognised the need for clinicians to be sensitive to this, for example cultural, psychological or functional ability.
The GDG considered the evidence pertaining to exercise that came from the review of combinations of non-invasive interventions. Exercise was given both as an intervention and in some instances as a comparator.
The GDG found it difficult to tease out which type of exercise modality was effective and the frequency and duration of the exercise to be given. They agreed that it would be useful to recommend an intervention that the person with back pain would be likely to participate in and that promotes self-management.
This review was unable to inform on the intensity of exercise programme, and the GDG agreed it was important to consider tailoring the programme to the individual, including taking into account an intensity that was feasible for the individual to be able to undertake and sustain. It was noted that the majority of exercise considered in this review was delivered by clinical providers.
Copyright © NICE, 2016.
Bookshelf ID: NBK410131

Views

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...