Article Text

Download PDFPDF

General group exercise in low back pain management in a military population, a comparison with specific spine group exercise: a service evaluation
  1. Joanna E Surtees1 and
  2. N R Heneghan2
  1. 1 Primary Care Rehabilitation Facility, RAF Waddington, Lincoln, UK
  2. 2 Centre of Precision Rehabilitation for Spinal Pain (CPR Spine), School of Sport, Exercise and Rehabilitation Sciences, University of Birmingham, College of Life and Environmental Sciences, Birmingham, UK
  1. Correspondence to Dr N R Heneghan, Centre of Precision Rehabilitation for Spinal Pain (CPR Spine), School of Sport, Exercise & Rehabilitation Sciences, University of Birmingham, Birmingham B15 2TT, UK; n.heneghan{at}


Objective To investigate whether general group exercise (GGE) offers the same outcomes compared with a specific spinal group exercise (SSGE) for chronic low back pain (CLBP) in a military population.

Design Retrospective service evaluation using routine service activity data.

Setting A UK military rehabilitation centre.

Participants A total of 106 patients with CLBP.

Interventions Three-week intensive (5 days per week, 15-day intervention) rehabilitation course for patients with CLBP. Six SSGE groups (n=64); CLBP only. Six GGE groups (n=42); CLBP patients grouped with chronic lower limb (LL) injuries.

Outcome measures Oswestry Disability Index (ODI), Numerical Pain-Rating Scores and the Modified Multi-Stage Fitness Test (Mod-MSFT). Long-term effects were measured by Medical Employment Standard (MES) status and physiotherapy follow-up at 3 and 12 months.

Results A between-group analysis showed no significant difference in GGE compared with SSGE. Mean changes (SD) in pain were −2.71±2.35 and −1.20±1.99 (p=0.018), ODI were −3.6±5.7 and −4±8.5 respectively (p=0.649) and Mod-MSFT 28.4±30.8 and 29.7±31.7 respectively (p=0.792). At 3 months, a greater proportion of the GGE were having ongoing physiotherapy; GGE=50%, SSGE=30.2%, (p=0.016) although some differences were evident across MES with 32.5 % of GGE compared with 20.6 % of SSGE being medically fit with no restrictions. At 12 months, groups were largely comparable for follow-up physiotherapy and MES; 22.5% of GGE and 20.6% of SSGE continued to have physiotherapy input; 47.5% of GGE and 50.8% of SSGE were medically fit with no restrictions.

Conclusion Patients with CLBP who completed a 3-week rehabilitation programme had comparable outcomes when grouped with patients with LL, although only improvements in pain in the GGE group achieved a meaningful change. Further evaluation of potential costs and savings to service costs is now required.

  • chronic low back pain
  • group therapy
  • exercise
  • service evaluation

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Key messages

  • Comparable outcomes are achieved if patients with chronic low back pain (CLBP) complete a 3-week exercise programme when grouped with patients with lower limb injuries compared with a group comprising patients with CLBP only.

  • Efficiency savings could be made by grouping mixed military musculoskeletal presentations together for group exercise therapy earlier in the care pathway.


As well as the leading cause of disability in the general working-age population globally, chronic low back pain (CLBP) is also the most common musculoskeletal presentation in the US and UK armed forces.1 2 In 2009, 7% of UK armed forces medical discharges were a consequence of low back pain (LBP).3 Where exercise is recommended for management of LBP4 in the general population, exercise rehabilitation has been identified as a research priority by the UK Defence Directive of Rehabilitation (DDR).5

Recent National Institute for Health and Clinical Excellence guidelines recommend group exercise as a management option for CLBP.6 Recent systematic reviews (2015 and 2017) comparing group exercise programmes to one-to-one physiotherapy for management of chronic musculoskeletal conditions, including CLBP, reported similar clinical outcomes for improvements in pain intensity and functional disability.7 8 While both reviews endorse the use of physiotherapist-led group exercise as a cost-effective management approach for CLBP, another Cochrane review concluded that no one form of exercise afforded superior outcomes9; mirroring earlier review findings from Searle et al,10 with evidence of support for a wide variety of exercise interventions, including yoga, Pilates, strength/resistance training and stability/coordination exercises.10

Group exercise therapy is a core component in CLBP management within the Defence Medical Rehabilitation Programme (DMRP).11 12 Regional Rehabilitation Units (RRU) provide 3-week intensive rehabilitation courses (5 days per week, 15-day intervention) for patients with chronic musculoskeletal conditions, including CLBP, with 15 patients enrolled on each course. This long established mode of delivery has historically differentiated upper limb, lower limb (LL) and specific spinal group exercise (SSGE) courses. Patients are referred to the RRU from physiotherapists working at Primary Healthcare Rehabilitation Facilities (PCRF) (Figure 1).

Figure 1

Summary of Defence Medical Rehabilitation Centre (DMRC) model. PCRF, Primary Healthcare Rehabilitation Facilities; RRU, Regional Rehabilitation Units

In June 2015, the investigating RRU was unable to fill the SSGE course quota; moreover, LL courses were routinely full. Consequently, patients with LL had to wait longer for a LL course, exceeding Key Performance Indicators outlined in the DMRP. Based on the available research evidence at that time, amalgamating the LL and SSGE into a general group exercise (GGE) course was a justifiable course of action to afford positive outcomes for patients with CLBP. Additionally, the ability to offer courses more regularly could reduce the socioeconomic burden of CLBP in a military population, hastening return to duties and reducing healthcare usage.

From a detailed literature search, the authors identified one study that investigated the effectiveness of exclusively LL exercise in the treatment of CLBP. Cai et al 13 demonstrated that LL strengthening was equally effective to lumbar extensor or lumbar stabilisation exercises for improving lumbar multifidus muscle activation and superior for running functional outcomes in a recreational running population with CLBP (n=84). No study was identified specifically investigating outcomes in exercise groups comprising both CLBP and LL conditions or in a more representative population. In view of these findings, and in line with the identified DDR research priorities, the aim of this service evaluation (SE) was to evaluate the outcomes of the GGE for patients with CLBP compared with existing data for SSGE.



A retrospective SE was designed using routine service activity data. In the absence of reporting guidelines,14 the Standards for Reporting Implementation Studies document was used to inform the methods of the SE.15 An a priori protocol was developed with expertise from the University of Birmingham and approved by the Academic Department of Military Rehabilitation.

Inclusion criteria

All patients with CLBP accepted for residential rehabilitation from December 2014 to June 2015 were admitted to six established SSGE courses; patients accepted from December 2015 to June 2016 were admitted to six GGE courses. Inclusion criteria: Army, Royal Air Force (RAF) or Royal Navy (RN) personnel, aged 17–55. All patients were seen by a general practitioner and physiotherapist at PCRF and, in 69% of cases, an exercise therapist. All patients had CLBP (>3 months)16 and a diagnosis confirmed by a Sports and Exercise Medicine Physician. For the purpose of this SE, clinical presentations were categorised as non-specific LBP, radiculopathy, sacroiliac joint, trauma and postoperative spinal surgery. All potential participants were screened to confirm eligibility for participation in active exercise in a group setting with no medical contraindications for example, cardiovascular, respiratory, neurological or mental health conditions.

Course design

There is no published evidence supporting the RRU 3-week residential rehabilitation course as opposed to different models or timescale. A duration of 3 weeks for the course is pragmatic, with a need to balance time to deliver a clinical intervention, manage waiting lists and allowing individuals’ time away from their unit and primary duties to focus on rehabilitation. The course elements include strength training and functional conditioning, sensorimotor training, motor control and dynamic stability training, range of motion, flexibility and general movement, cardiovascular conditioning, hydrotherapy, and education (pain, goal setting, anatomy and physiology, diet and nutrition, principles of training and relaxation).17

Each course is led by a designated senior physiotherapist and exercise therapist. Exercise prescription, progression and intensity is controlled and monitored by the physiotherapist and exercise therapist, and always conducted in a group environment. In line with normal service delivery, one-to-one treatment was available, if required. The main difference between the SSGE and GGE was that unlike the GGE the SSGE had daily, mat-based, spinal mobility sessions.

Outcome measures

Pain: Numerical Pain-Rating Score (NPRS)

A valid and responsive self-report measure of pain intensity (0–10) where 0=no pain and 10=worst possible pain with values recorded at pre-rehabilitation and post-rehabilitation course.18

Disability: Oswestry Disability Index (ODI)

The ODI is a back-specific patient-reported questionnaire, consisting of 10 questions that assess the level of pain interference with physical activities of daily living.19 Test–retest reliability is reported to be excellent intraclass correlation coefficient (ICC) 0.88 (95%CI 0.77 to 0.94) and ICC 0.94 (95%CI 0.89 to 0.97).20 21

Long-term outcomes of the intervention were assessed using individual’s Medical Employment Standard (MES) status at 3 and 12 months. MES categories are listed below,22 although this SE had a specific focus on medical fitness with no restrictions (P2):

Fitness: Modified Multi-Stage Fitness Test (Mod-MSFT)

A measure physical function, the Mod-MSFT is a modification of the established MSFT.23 It was first used with traumatic brain injuries demonstrating excellent reliability and validity.24 25 Markers are place at 0, 10 and 20 m where the test involves walking, and then running the 20 m distance in time to a shortening frequency of beeps, played out on an audio device. The test is terminated by the patient due to pain or fatigue and has been used in a CLBP military population.26 The MSFT is used by the RAF and RN as measurement of physical fitness; achieving an age and sex appropriate pass mark is essential to achieve medical fitness with no restrictions (P2) MES.22 27

Healthcare use including ongoing physiotherapy interventions was also evaluated at 3 and 12 months.28


ODI and Mod-MSFT were recorded by the individual course physiotherapist at the start and end of each course. The researcher extracted baseline demographic characteristics and all outcome measure data from a manual search of electronic defence medical records of all individuals participating in the six SSGE and six GGE courses.

Data analysis

Data were analysed using primarily descriptive methods, using the statistical analysis software SPSS V.21. The alpha level was set at 0.05. Prior to statistical analysis, the Shapiro-Wilk test for normality was used due to the small sample size. As a result, the Mann-Whitney U test was selected as an appropriate non-parametric test.29


A total of 106 patients with CLBP met the inclusion criteria and were included in the evaluation. The personal characteristics of participants are presented in Table 1.

Table 1

Characteristics of participants in specific spinal group exercise (SSGE) and general group exercise (GGE)

Table 1 shows there was no between-group statistical significant difference for the following characteristics: age (p=0.864), waiting time (p=0.864) or male/female ratio (p=0.170). The most common clinical diagnosis was non-specific CLBP; 73 of the 106 sample; SSGE 70.3%, GGE 65.9%. There was no statistically significant difference between groups in the clinical presentation (p=0.413). The frequency of non-specific LBP was lower than the commonly reported 90% of all presentations of LBP.30 Waiting time was measured from first presentation at PCRF to the first day of the course. The most common LL presentations within the GGE were postop anterior cruciate ligament reconstructions, anterior knee pain and hip pain.

A between-group analysis, summarised in Table 2, showed no significant difference in the GGE outcomes compared with the SSGE group. Pain mean change was −2.71 and-1.20 (p=0.018), ODI mean change −3.6±5.7 and −4±8.5 respectively (p=0.649) and mod-MSFT mean change 28.4±30.8 and 29.7±31.7 respectively (p=0.792).

Table 2

Pre-intervention and post-intervention measures for pain, disability and physical function

Physiotherapy and functional status at 3-month and 12-month follow-up

At 3 months a greater proportion of the GGE group were still having ongoing physiotherapy care (50%) compared with 30.2% in SSGE, although more of the GGE were medically fit with no employment restrictions (32.5%) compared with 20.6% of the SSGE group. Notwithstanding some differences across categories of MES, at 12 months, groups were largely comparable with 22.5% of the GGE and 20.6% of the SSGE continuing to have physiotherapy and 47.5 of GGE and 50.8% of SSGE deemed medically fit with no employment restrictions (see Table 3).

Table 3

Physiotherapy and functional status at 3-month and 12-month follow-up

Missing data

Full data sets were available for ODI evaluation, although 3 values were missing for post-course Mod-MSFT SSGE, and 11 for NPRS pre and post SSGE, 13 pre and 19 post GGE.


The aim of this SE, the first of its kind, was to examine outcomes in patients with CLBP completing a 3-week course when grouped with patients with LL compared with a group of patients with CLBP only. Given the inherent difficulties of conducting clinical trials in a military setting, where participants may be deployed or posted at short notice, use of SE offers an alternative approach to evaluate practice and implement changes in a timely manner. Additionally, this offers a means to systematically assess activities and outcomes to examine efficiency and effectiveness of a service.31 While cost effectiveness is a key driver, a new multicriteria decision analysis model incorporates a more comprehensive evaluation inclusive of access, equity, effectiveness of treatment and impact on future services.32 This SE therefore provides a robust evaluation of the impact of changes in patient outcomes in a military setting. In summary this SE found more than two point difference in pain scores in favour of GGE, although no between-group difference was found with respect to disability or physical function. A greater proportion of the GGE were still having ongoing physiotherapy care at 3 months although 32.5%% of the GGE were medically fit with no restrictions compared with only 20.6% of the SSGE. At 12 months, groups were largely comparable with respect to follow-up physiotherapy and those who were deemed medically fit with no or minor employment restrictions.


Selection, administration and interpretation of outcome measures are important facets in evaluation.33 In this SE outcome measures were informed by DDR policy, with four of the five well-established areas for measuring outcomes in LBP included: disability, back specific function, generic health status, pain and work represented.34


Notwithstanding the extent of the missing data for NPRS change, scores achieved those reported in the wider literature for MCID.18 It is interesting that this sizeable change was observed during the 3-week course which would suggest that the non-physical factors such as beliefs, knowledge, and so on had a role in pain perception; reflective of the multidimensional nature of LBP.35 Caution should be taken when interpreting these findings given the extent of the missing data for pain.


Based on the reported requirements for the general population neither group achieved a meaningful change in disability scores. However, where both groups were largely of minimal disability based on ODI and the absence of a population-specific measure of disability, where a sensitive and specific has not yet been established, groups did meet the values required at the lower end of the range for the reported MCID from other populations, ranging 2.92 to 15.3636 to a 10-point change combined with a 30% improvement from baseline.37


The only MCID documented in the literature for a shuttle-based test is the Shuttle Walking Test.38 The authors found a change of 76 m would be required to represent a 95% CI. However, this was in a population of 29 patients with a diagnosis of spinal stenosis, with a mean age of 69; notably higher than the mean age reported here. Moreover, ceiling effects of this test were found39; out of a total of 90 patients, 31 had achieved 11 of the 12 levels at baseline assessment. Furthermore, it has a different format to the Mod-MSFT being conducted over a 10 m not 20 m distance.

The only comparable study using the Mod-MSFT in a CLBP military population involved 56 subjects completing an equivalent 3-week course at the Defence Medical Rehabilitation Centre (DMRC). Roberts et al 26 reported a mean change of 120 m, considerably less than distances in this SE; 284 m (GGE) and 297 m (SSGE). Given their lower mean pre-course and post-course distances of 1040 and 1160 m respectively, this suggests a lower functioning group compared with this SE (GGE=1715 m; SSGE=1667 m). As a tertiary tier of the care pathway (Figure 1), these findings for DMRC are not unexpected. In the absence of other published data, the mean post-course scores achieved by both groups is 500 m below the pass mark required for males of a comparable age in the RAF and RN MSFT,27 suggesting that full MES was still not achieved by the end of the course.

MES was used to evaluate work disability rather than the Functional Assessment Tool (FAA) reported elsewhere in the literature which limits the ability to draw comparison with this current service evelaution.40 MES selection was chosen primarily due to the known problem of clinicians assessing FAA in practice, rather than the patient.41 That said the MES correlates well to the FAA and critically has been linked to military operational effectiveness.

Implications for practice and policy

It has been documented that prompt recovery (in non-specific LBP) is most likely to occur during 3 months post onset, with only gradual improvements thereafter.42 43 Moreover, studies have found that 62% of all patients continued to complain of pain at 12 months.44 This raises the question as to what we can realistically expect given patients commenced the course, on average, 8+ months from initial presentation. Only 69% of patients saw an Exercise Therapist at PCRF pre course; one of their primary functions is group exercise. There may be greater potential for improvements if rehabilitation courses are offered earlier in the care pathway, and which may in turn ameliorate some of the cost burden of managing more established chronic pain presentations.

With a lack of comparable data for those individuals who did not attend an RRU course, this study raises the question of whether the right patients are being selected for course participation as part of the DMRP tier approach. Despite DMRP referral guidelines and timescales, referral patterns to the RRU are patient-dependent and therapist-dependent, informed by therapist expertise, patient operational demands and clinical presentation. This may also explain the variability in waiting times seen in this SE. With a review citing 1501 potential prognostic factors associated with poor recovery from LBP,43 decision-making for patient referral is complex. With the recent introduction of the STarTBack tool into the DMRP, this may now better differentiate different presentations of LBP and inform targeted management.11 45

Maher46 summaries the challenges clinicians face where no single treatment cures CLBP, and the abundant unregulated, non-evidence-based management options that bombard patients confuse the issue further. While there is no specific evidence for the 3-week model, elements found within the course are well evidenced. This SE goes someway to justify the need for more research into the modes of rehabilitation delivery in the UK military setting to assist with clinical decision-making.

Strengths and limitations

One of the main limitations is that pain, ODI and Mod-MSFT data were not available at the 3-month or 12-month follow-up points. The habitual use of outcome measures in clinical practice has challenges and is widely reported in the literature.47 The lack of routine outcome measures recorded across the DMRP limits the impact of the findings of this SE and warrants further investigation. Moreover, population-specific measures with established measurement properties are required to further inform practice decisions.48 This SE has highlighted the inconsistent recording of the numerical pain rating scale18 pre course and post course, despite being an outcome measure documented in DDR policy. Finally evaluation of the impact of a GGE course on patients with LL was beyond the scope of this SE, although it could be useful to strengthen proposed service changes.


Patients with CLBP who completed a 3-week rehabilitation programme had comparable outcomes when grouped with patients with LL injuries, although only improvements in pain in the GGE group achieved more than the MCID on completion of the course. At 12-month follow-up, both groups were largely comparable with respect to achieving medical fitness with no or minor employment restrictions. This service evaluation supports the need to further consider timing for rehabilitation in the care pathway, comprehensive use of patient-reported outcomes and further evaluation of potential costs and savings to service costs.



  • Contributors JS and NRH: service evaluation design, analysis and manuscript development. JS: data extraction and analysis.

  • Funding The author received no specific grant from any funding agency in the commercial, public or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent Not required.

  • Ethics approval School of Sport, Exercise and Rehabilitation Sciences, University of Birmingham.

  • Provenance and peer review Not commissioned; externally peer reviewed.