Improving the understanding and management of back pain in older adults: the BOOST research programme including RCT and OPAL cohort

Article history

The research reported in this issue of the journal was funded by PGfAR as project number PTC-RP-PG-0213-20002. The contractual start date was in April 2015. The final report began editorial review in April 2022 and was accepted for publication in March 2023. As the funder, the PGfAR programme agreed the research questions and study designs in advance with the investigators. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The PGfAR editors and production house have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the final report document. However, they do not accept liability for damages or losses arising from material published in this report.

Copyright statement

Copyright © 2023 Williamson et al. This work was produced by Williamson et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This is an Open Access publication distributed under the terms of the Creative Commons Attribution CC BY 4.0 licence, which permits unrestricted use, distribution, reproduction and adaptation in any medium and for any purpose provided that it is properly attributed. See: https://creativecommons.org/licenses/by/4.0/. For attribution the title, original author(s), the publication source – NIHR Journals Library, and the DOI of the publication must be cited.

2023 Williamson et al.

Research summary

Low back pain (BP) is the most common cause of disability in the world. 1 The prevalence of severe BP increases with age,2 and it is associated with immobility, disability, frailty and falls. 3–7 In the UK, estimates suggest that one in four older adults suffer from BP. 2,8 However, many older people either do not consult, or health professionals do not prioritise BP treatment, presuming nothing can be done. 2,9,10 Many clinical trials exclude older people. 11 Little attention has been paid to understanding the presentation of BP in older people and developing effective treatments.

Older adults may also experience leg pain referred from the lumbar spine. A common clinical presentation of spinal-related leg pain in older adults is neurogenic claudication (NC). 12 It presents as an easily recognised clinical pattern of symptoms radiating from the spine into the buttocks and legs which are provoked by walking or standing and relieved by sitting or lumbar flexion. 12 BP is often present, but not always. NC is particularly problematic for older people, as not only can it cause severe pain and discomfort but it substantially reduces an individual’s confidence and ability to walk and their quality of life. 13 It is estimated that 11% of community-dwelling older adults report symptoms of NC, although there were no estimates from community-dwelling UK populations. 14 Symptoms of NC are thought to arise from pressure on nerves and blood vessels in the spinal canal caused by degenerative changes narrowing the volume of the spinal canal (referred to as spinal stenosis). NC is the most common reason that older adults undergo spinal surgery,15 and physiotherapy is recommended prior to surgical intervention. However, there was a lack of high-quality evidence to guide conservative care of people with NC, including primary care management or physiotherapy. 16,17 We identified a need to better understand the presentation of BP in older people and to improve the care they receive for their BP, including conditions like NC.

Aims and objectives

The overall aim of the Better Outcomes for Older people with Spinal Trouble (BOOST) programme was to conduct a series of linked studies to understand and reduce the burden of BP on older people by increasing knowledge and developing evidence-based treatment strategies.

We set out the following objectives:

1. to describe the presentation and impact of BP in community-dwelling older adults (work packages 2 and 3)

2. to refine a physiotherapy intervention for older people with NC to be tested in a randomised controlled trial (RCT) (work package 1)

3. to undertake a definitive RCT evaluating the clinical and cost-effectiveness of physiotherapy interventions for NC (work packages 2, 4 and 5)

4. to undertake a longitudinal qualitative study embedded within the BOOST RCT to understand the participants’ experiences of taking part in the trial, including acceptability and impact of the interventions, and issues related to implementation (work package 4)

5. to develop a prognostic tool using the Oxford Pain Activity and Lifestyle (OPAL) cohort study data to predict when older people are at risk of mobility decline and to understand the role of BP (work packages 2 and 3)

6. to integrate the findings into an implementation package (work package 6).

Work packages

We have undertaken:

1. a large population-based cohort study of community-dwelling older people recruited from primary care practices (the OPAL cohort study)

2. An RCT of physiotherapy interventions for NC, including an economic evaluation (the BOOST RCT) and prespecified subgroup analyses

3. a longitudinal qualitative study embedded within the BOOST RCT.

This research was planned in six work packages. There was considerable overlap between work packages, and they were not undertaken chronologically. The original work packages as described in the funding application are described below. Overall, this proceeded as planned, with some changes to work package 3 which we describe.

Changes to the programme

During the first year of this programme of research, we made refinements to the OPAL cohort study.

We originally planned to enrol 1000 older adults with BP in the cohort study. To identify the 1000 older adults with BP, we needed to screen a minimum of 4000 older adults who responded to the invitation to participate in the cohort study. This was based on the most recent estimates of disabling BP in the population aged 70–90 years, with a prevalence of 25–30% for these age groups, respectively. 8

We decided to enrol all invitation respondents in the cohort regardless of their BP status. In addition to the original research objective (to recognise when and what types of BP presentation were important treatment targets for older people), we would be better able to evaluate the contribution of BP to mobility decline within a broader context. This change to the original proposal was reflected in the study protocol approved by the London–Brent Research Ethics Committee (16/LO/0348) on 10 March 2016. Changes were supported by the Programme Steering Committee and reviewed and accepted by the Programme Grants Manager on 30 March 2016.

Work package 6 could not be delivered as planned due to the challenges arising from the COVID-19 pandemic. We had initially planned to run focus groups with clinicians to discuss the study results and implications. However, delays to results meant this was not completed. The OPAL cohort study was nearing completion of the 2-year follow-up mail-outs when the COVID-19 lockdown started in March 2020. Although the team managed to send out all year 2 questionnaires just prior to the lockdown, this had a major impact on the study team’s ability to process the large volume of data collected from home, which was very time-consuming using remote access. Data were available 6 months late, delaying the analysis needed to develop the OPAL prognostic tool. The BOOST trial health economic analysis was delayed by 6 months as our health economist was working at home with young children throughout lockdown. This work was essential to interpretation of the trial findings and significantly delayed our plans to publish the results. Despite these challenges, we had strong engagement with clinical collaborators throughout this research and we worked as best we could to engage with clinical staff to form our dissemination plan.

Research pathway diagram

Background

We lacked a good understanding of the prevalence and impact of back and leg pain within the community-dwelling population in the UK. There were estimates for the prevalence of BP in the UK2,8 but not for back-related leg pain. A single UK-based study (n = 30) estimated the prevalence of NC among patients (50 years and over) presenting with back and leg pain to primary care to be 87%, suggesting it was the predominant cause of spinal-related leg pain when patients sought care. 18 However, we wanted to understand the extent of the problem among community-dwelling older adults and not just those seeking care.

In addition to understanding the presentation of back and leg pain, we wanted to understand the broader impact of having back and leg pain in later life. We were interested in understanding the relationship between different presentations of back and leg pain with quality of life and age-related adverse health states including frailty, mobility decline, falls, poor sleep quality and incontinence. These health states are common in later life and associated with substantial morbidity and poor outcomes. 19 Little was known about the relationship between BP and different patterns of leg pain and these age-related adverse health states. 5–7,20,21 Back and leg pain had been associated with falls and functional difficulties, but studies were limited. 22 Different patterns of back and leg pain had not been taken into consideration, and it was unknown whether certain presentations were more strongly associated with adverse health states.

There was also little information about the natural history of symptoms among community-dwelling older adults reporting back and leg pain within the UK. Cohorts of older people with BP have reported that around 50–60% of participants will continue to experience pain and disability at 12 months’ follow-up. 23,24 There were some data on the natural history of symptoms due to spinal stenosis, with the majority of people (around 55%) reporting no change in pain over a 3-year period while around one-third improved and 10–13% reported a worsening of symptoms. 25 It was unclear how recovery may differ among those with different back and leg pain presentations.

In this section, we will examine:

1. the presentation of back and leg pain symptoms among OPAL cohort participants at baseline and the association with adverse health outcomes

2. for those with back and leg pain at baseline, the persistence of symptoms at 2 years’ follow-up.

Methods

The OPAL protocol and a description of the participants at baseline26 can be accessed here: http://doi.org/10.1136/bmjopen-2020-037516

A more detailed description of the baseline cross-sectional analyses can be accessed here: https://ora.ox.ac.uk/objects/uuid:911cf725-135f-4951-a1b2-28eb2032d9ab

Briefly, there are 5409 community-dwelling older adults enrolled in the OPAL cohort study. Participants are over 65 years of age, registered with one of 35 participating primary care practices, and were recruited using electronic record searches. Older adults living in residential care or nursing homes were ineligible. Potential participants received an information leaflet, consent form and baseline questionnaire. Those who returned a completed consent form and questionnaire to the study office were enrolled in the study. Of the individuals invited, 42% (5409/12,839) were enrolled in the study. Follow-up was by yearly postal questionnaire. A summary of the data collected is available in Appendix 1, Table 5. Participants were included in these analyses if they completed the questions about back and leg pain.

Results

Baseline presentation

The mean age of participants included in these analyses was 75 ± 6.8 years. Over half of the participants reported BP, with 33.7% (1786/5304) reporting BP only, 11.2% (594/5304) reporting BP and NC, and 8.3% (441/5304) reporting BP and non-NC leg pain. Only the prevalence of BP and NC increased with age. BP was more common in females and those reporting more comorbidities. Those with leg pain reported poorer health behaviours (smoked more, higher BMI), lived in more deprived areas and reported lower education levels. Participants with BP had lower quality of life compared to participants without BP. Quality of life was poorest in those with NC. The age of onset was similar across BP groups. Around 30% of participants reported that their BP started 25 years ago or more. Those with leg pain reported more frequent BP. A report of pain on most days/every day was most common among those with NC. Participants with NC reported the highest ratings of pain troublesomeness. Very few reported improving symptoms since onset (3.7%) and, most commonly, symptoms fluctuated over time (75%). The proportion of participants reporting persistent pain or pain getting worse since onset was highest in those with NC (75%). See Appendix 1, Table 6 for a full description.

Just over half of the participants (55%; 2937/5304) reported at least one adverse health state. The most common adverse health states reported were a fall in the last year (29%; 1534/5304) and being frail (27%; 433/5304) (Figure 1). Confidence to walk half a mile was generally high among the cohort except for those reporting NC (see Appendix 1, Table 6). Across the three BP groups, there was a greater prevalence of all adverse health states compared to those with no BP; this increased further among those who reported leg pain, and it was highest among those with NC.

FIGURE 1.

Prevalence of adverse health outcomes by BP groups. Notes: Poor sleep quality = fairly bad or very bad sleep quality on Pittsburgh Sleep Quality Index. Urinary incontinence = never or less than once per week was considered continent and the remaining responses were considered incontinent. Mobility decline: ‘Compared to one year ago, how would you rate your walking in general?’ scored on a five-point scale constructed for the study. Participants reporting worsening of walking were classified as having mobility decline. Participants who rated their confidence to walk at 9–10/10 were categorised as having low walking confidence (reverse scored). Frail = ≥ 5 on Tilburg Frailty Index.

After adjusting for demographics, lifestyle factors, comorbidities and multisite pain, all BP groups were associated with the adverse health states studied (Table 1). For all the adverse health states, there was an increase in the strength of association with the addition of leg pain. This was particularly noticeable in participants with NC, where we observed the strongest associations with frailty, mobility decline, low walking confidence and falls compared to no BP.

TABLE 1 - Cross-sectional association between back and leg pain and adverse health states compared to no BP; relative risk ratios (95% CI); no BP is the reference category

Range: 1–10 where higher score means lower level of confidence walking.

Note

Models adjusted for demographic (age, sex, occupational physical demands, education, deprivation) and lifestyle factors (BMI and smoking status), comorbidities and multisite pain.

Age-related adverse health outcomes	BP only	BP + non-NC leg pain	BP + NC
Confidence to walka	1.23 (1.06 to 1.43)	1.89 (1.52 to 2.36)	3.11 (2.56 to 3.78)
Mobility decline	1.13 (1.00 to 1.26)	1.42 (1.22 to 1.64)	1.74 (1.54 to 1.97)
Falls in previous year	1.14 (1.03 to 1.26)	1.40 (1.22 to 1.61)	1.42 (1.25 to 1.61)
Frail	1.38 (1.24 to 1.54)	1.77 (1.55 to 2.02)	1.88 (1.67 to 2.11)
Poor sleep quality	1.17 (1.02 to 1.35)	1.68 (1.41 to 1.99)	1.67 (1.42 to 1.97)
Urinary incontinence	1.17 (0.95 to 1.43)	1.59 (1.22 to 2.07)	1.43 (1.12 to 1.83)

Limitations

Data were collected using postal questions, so we were reliant on self-reporting of symptoms. Some participants may have been misclassified or symptoms misdiagnosed as spinal-related leg pain (e.g. vascular claudication can have similar symptoms).

Discussion

Back and leg pain is a common problem for community-dwelling older adults, and for many it is a long-standing problem, with few reporting an improving clinical picture. At 2 years’ follow-up, only 23% of participants no longer reported symptoms. Symptom persistence was highest among those with NC at baseline, with 90% still reporting symptoms.

At baseline, older adults with back and leg pain reported lower HRQoL, with those with NC most affected. The report of back and leg pain was associated with adverse health states, including frailty, falls and mobility decline, compared to no report of BP. Participants reporting symptoms of NC were most affected and the strongest associations were with mobility-related adverse outcomes. At follow-up, those with NC at baseline remained the most severely affected, with a quarter of this group reporting severe activity limitations due to their symptoms. These findings confirmed our hypothesis that NC is particularly problematic for older people and highlighted the need to develop effective interventions to reduce the burden of this condition. However, BP alone is also a widespread persistent problem associated with age-related adverse health states, so ways to reduce the burden of BP are needed.

Starting point

As we have described, NC is a common, debilitating spinal condition affecting older people. 28 At the start of this programme of work, we had identified a strong theoretic underpinning and limited evidence from clinical practice, cohort studies and small RCTs for a physiotherapy intervention for NC, providing us with a firm starting point to refine and test an intervention. Experts hypothesised that stretching and mobilising the spine relieved pressure on spinal nerves and blood vessels and that aerobic exercises improve circulation, alleviating ischaemic changes. 46

However, recommendations were not substantiated by high-quality evidence46 despite being used commonly in clinical practice. 47,48 Systematic literature reviews reported that the current evidence from non-operative care was of very low to low quality, thus prohibiting recommendations to guide clinical practice. 16,17,49–51 Trials to date were of small sample sizes and often included only short-term follow-up. Interventions focused on the mechanics of spinal stenosis with little regard to the psychological impact of pain or ageing. Notably, interventions did not include strengthening exercises to counter the loss of muscle mass associated with ageing that contributes to loss of mobility in older people;52 nor had the authors considered addressing negative beliefs about pain or ageing that were potential barriers to engaging with and adhering to rehabilitation. 53,54

We explored the rehabilitation needs of older people with NC during a preparatory qualitative study. 55 Participants sought rehabilitation in the hope of re-engaging with meaningful activities and reducing pain. A combination of one-to-one and group sessions was preferred to deliver rehabilitation. One-to-one time with a physiotherapist was considered an important prerequisite for treatment to allow the therapist to understand the participant’s condition and circumstances. Participants endorsed exercising in a group for peer-to-peer support and socialising. Discourses around ageing influenced their experiences – especially regarding the use of walking aids, which they considered stigmatising. These views and preferences informed the refinement of the proposed intervention to ensure it would meet the needs of people with NC.

A full description of the BOOST intervention56 can be accessed here: https://doi.org/10.1016/j.physio.2019.01.019

The preparatory qualitative study55 can be accessed here: https://ora.ox.ac.uk/objects/uuid:92c9ee3e-30eb-4fc0-95fd-08db3bc4707b

Methods

We followed Medical Research Council guidance on developing a complex intervention. 57

Step 3: modelling processes and outcomes

A conceptual model of the change processes, active intervention elements and potential moderators of the BOOST intervention was produced (Figure 2).

FIGURE 2.

Conceptual model of the BOOST programme. (From Ward et al. 56)

The BOOST programme

The programme consisted of an individual assessment, 12 group sessions and two support telephone calls.

Prior to attending the group programme, each participant attended an individual appointment (1 hour) for an assessment and to set their individualised exercise and walking circuit targets for the group sessions.

Participants attended twelve 90-minute group sessions over a 12-week period. Each session followed the same format. Participants took part in an education and discussion session (30 minutes) which incorporated behavioural change strategies to encourage adherence with home exercises. The topics of discussion are included in Appendix 3, Table 14. This was followed by the exercise programme, lasting approximately 1 hour.

There was a short warm-up of seated exercises, which included arm raises, trunk rotation, pelvic tilting and knee lifts. Then participants undertook a circuit of strengthening which included the following exercises:

• sitting knee extension

• sit to stand

• standing hip abduction

• standing hip extension

• a combined hip flexor and calf stretch

• a progressive balance sequence (feet together static balance, semi-tandem stance static balance, full tandem stance static balance, forward tandem heel-to-toe walking, backward tandem heel-to-toe walking).

Each participant completed their individually tailored programme. Exercises were progressed over the 12 weeks. We used the Borg Rating Scale of Perceived Exertion to ensure participants achieved an adequate stimulus to promote strength gains and encouraged them to work at level 5–6 on this scale (the exercise feels hard) during the strengthening exercises (Figure 3).

FIGURE 3.

Borg scale.

Participants then undertook a supervised walking circuit, designed to improve walking ability and fitness, which also progressed over the 12 weeks by increasing the distance walked, increasing walking speed, adding balance challenges such as stairs or obstacles, or adding weights. The exercises carried out during the supervised sessions made up the home exercise programme (warm-up, exercise circuit and walking).

On completion of the 12 group sessions, participants were asked to carry out their home exercise programme at least twice per week. The physiotherapist reviewed each participant by telephone 1 and 2 months after they completed the supervised sessions, to promote long-term adherence with home exercises.

Control intervention

The control intervention was best practice advice (BPA) delivered in individual physiotherapy appointments, which reflected what we considered a ‘good’ standard of usual care physiotherapy. This intervention was informed by a survey of current physiotherapy practice47 and a recent trial,56 and through consultation with clinicians and patient representatives. We recommended that participants receive one session of advice and education. However, the clinicians highlighted situations where a review would be necessary (e.g. to review walking aids) so this was permissible.

Participants attended a 1-hour appointment consisting of an assessment followed by the provision of tailored advice and education and written information. Participants were prescribed up to four home exercises. Flexion and trunk stabilisation were recommended. The physiotherapist could prescribe a walking aid if indicated. A maximum of two half-hour review appointments were permitted. During review sessions, physiotherapists reinforced the verbal advice given, and reviewed walking aids or exercises.

Intervention delivery training

We developed a clinician manual and training programme. Physiotherapists delivering the BOOST programme attended a full day of training with the research team. They were also asked to complete some online training prior to attending. Those delivering the control intervention attended a separate half-day of training.

Pilot work

We piloted the training, which was refined following feedback. We conducted the first BOOST programme with two participants, and feedback indicated satisfaction with the programme content, duration and timing of sessions. Minor modifications were made. A researcher observed the control interventions, which were delivered as intended.

Summary of work undertaken

We focused on refining a complex intervention that was underpinned by existing knowledge, would meet the needs of older people with NC, and would be acceptable to clinicians and deliverable in the NHS. By combining group delivery with an individual assessment and follow-up calls, we ensured each participant underwent a thorough assessment so they were confident in the programme but could benefit from peer-to-peer support and socialising. This was also a way to deliver a longer-duration intervention for participants, which would increase the likelihood of achieving meaningful gains in muscle strength and mobility without placing excessive burden on physiotherapy departments.

Methods

The full protocol69 is available here: https://doi.org/10.1136/bmjopen-2020-037121

The statistical analysis plan70 is available here: https://doi.org/10.1186/s13063-020-04590-x

The paper reporting clinical effectiveness71 is available here: https://doi.org/10.1093/gerona/glac063

The paper reporting the cost-effectiveness72 is available here: https://doi.org/10.1186/s12962-022-00410-y

Results

Recruitment and follow-up

Recruitment was undertaken between 1 August 2016 and 29 August 2018 at 15 sites. We selected sites in different areas across England to maximise recruitment of participants from different ethnic and socioeconomic backgrounds. A list of trial sites and description of the local population is provided in Appendix 2, Table 9.

In total, 438 participants were found to be eligible, provided informed consent and were randomised (spinal clinics n = 394, OPAL n = 44). Three participants withdrew post randomisation and removed consent to use their data, so the final sample size was 435 participants (BPA n = 143, BOOST programme n = 292) (Figure 4).

FIGURE 4.

CONSORT diagram. ^a Not all participants who were eligible to be screened for BOOST were invited. ^b Numbers included in analysis are all participants with at least one follow-up ODI outcome and the baseline variables used in the model. This was due to lack of staff capacity at sites to accept additional referrals for screening or due to lack of a referral pathways between primary care centres and study sites.

The primary outcome was obtained for 88.0% (383/435) and 87.4% (380/435) of participants at 6 months and 12 months, respectively, with 93.0% (403/435) contributing data to the primary analysis.

Key secondary outcomes

See Williamson et al. 71 for a full report of the secondary outcomes (see also Appendix 2, Tables 12 and 13).

Walking and physical performance

The BOOST programme had a lasting impact on walking capacity (6MWT) (Figure 5) at 6 and 12 months’ follow-up, favouring the BOOST programme. The ODI walking item also reflected this finding. BPA participants showed very little change across the two follow-up time points. A similar response was observed for physical performance (SPPB) (see Appendix 2, Table 13).

FIGURE 5.

Six-minute walk test results. (From Williamson et al. 71)

Falls

Participants in the BOOST programme had a substantially reduced risk (by 40%) of reporting a fall over the 12-month period. The proportion of participants sustaining a fall-related fracture was very small but similar between groups.

Pain and other symptoms

Both groups reported a small reduction in Swiss Spinal Stenosis Scale (SSSS) symptoms subscale scores at 6 months, and these were larger for the BOOST programme. Small reductions were maintained at 12 months, and there was no longer a difference between the groups at 12 months. Similar findings were observed for troublesomeness, global rating of change and satisfaction with changes in back and leg problems.

Participant satisfaction

Participants in the BOOST programme were more likely to be satisfied with their treatment at 6 and 12 months compared to the control group.

Exercise adherence

Participants were asked how often they performed their home exercises. At 6 months, 190/257 (73.9%) BOOST programme participants reported performing their exercises at least twice per week, which reduced to 143/250 (57.2%) at 12 months. At 6 months, 102/125 (81.6%) BPA participants reported doing their exercises at least twice a week, which reduced to 89/125 (71.2%) at 12 months.

Adverse events

One serious adverse event (cardiac symptoms) occurred during a BOOST session which was deemed unrelated to the intervention. There were no serious adverse events reported for BPA. There were 12 adverse events reported for the BOOST programme. Four were assessed as definitely related to the programme, including aggravation of joint pains (n = 2), a fall during the walking circuit (no injuries) and skin irritation by an ankle weight. Two adverse events were reported for BPA, and neither was definitely related to the intervention.

Economic analyses

Complete QALY profiles were available for 357 (82%) participants based on the EQ-5D-5L. Completion of health resource use data for the base-case economic evaluation was similar at each time point between the BOOST and BPA groups. All data needed to calculate the cost components were available for 75–88% of health resource categories apart from hospital outpatient services, which ranged from 57% to 75%.

Intervention costs and resource utilisation

Mean total intervention costs varied between £242 and £911. The average cost per group session per participant varied from £11.80 to £67.00 depending on the number of participants per session. The mean cost per participant was generally lower across all sites if the target number of participants (n = 6) was achieved.

For health and PSS use, there were non-significant differences between the two groups in utilisation of hospital inpatient and outpatient care, community-based health care and social services, and days off work.

For the primary analysis, mean NHS and PSS costs, inclusive of intervention costs, over 12 months were £1974.06 for the BOOST programme versus £1826.64 for the BPA group. There was a non-significant cost difference in favour of the BPA group of £147.42 (95% CI £419 to £714). Mean total societal costs, inclusive of the intervention cost, were £2176.01 in the intervention group compared with £2140.54 in the BPA group. This generated a mean cost difference of £35.47 (95% CI £469.57 to £540.51) in favour of the BPA group. Societal costs (excluding NHS and PSS costs) were higher in the BPA group and primarily driven by economic valuation of time taken off work by a small number of patients and carers in the BPA group.

Health-related quality-of-life outcomes

The adjusted mean participant-reported QALY estimate for the base-case analysis over 12 months favoured the BOOST programme [0.621 (standard error 0.009) vs. 0.599 (standard error 0.006); between-group difference 0.021 (95% CI 0 to 0.044)].

Cost-effectiveness results: base-case analysis

NHS and PSS perspective

The baseline economic evaluation indicated that the BOOST programme was associated with marginally higher NHS and PSS costs (£147, 95% CI −419 to 714) and an increase in QALYs (0.020, 95% CI −0.003 to 0.045). The mean incremental cost-effectiveness ratio (ICER) for the BOOST programme was estimated at £7211 per QALY gained; that is, on average, the BOOST programme was associated with a higher cost and an increase in QALYs. The associated mean INMBs at cost-effectiveness thresholds of £15,000, £20,000 and £30,000 per QALY were £145, £244 and £464, respectively. The base-case mean INMB was >0, suggesting that the BOOST programme would result in an average net economic gain of approximately £244 (INMB = £244, 95% CI −£570 to £1058). The probability of cost-effectiveness for the BOOST programme was estimated as 67%, 78% and 83% at cost-effectiveness thresholds of £15,000, £20,000 and £30,000 per QALY, respectively (Figure 6). The joint distribution of costs and outcomes for the base-case analysis is presented graphically in Figure 7.

FIGURE 6.

Cost-effectiveness acceptability curve for the base-case imputed (covariate-adjusted) analysis. (From Maredza et al. 72)

FIGURE 7.

Cost-effectiveness plane for the base-case imputed (covariate-adjusted) analysis. (From Maredza et al. 72)

Societal perspective

The probability that the BOOST programme is cost-effective was higher from a societal perspective, ranging between 79% and 89% across cost-effectiveness thresholds.

Limitations

Five physiotherapists trained in delivery of the BOOST programme also treated 24/143 participants (16.8%) allocated to BPA due to physiotherapist availability. There was a risk that these physiotherapists could integrate elements of the BOOST programme such as behaviour change principles into their treatment sessions and provide treatment beyond the BPA protocol. However, the proportion of participants exposed to potential contamination is well below the 30% threshold considered a serious threat. 89 We carefully monitored intervention delivery, including fidelity assessments, and are confident that the risk of contamination between arms was minimised.

Resource use data for the health economic analyses were retrospectively recalled by participants, which may mean they were affected by recall bias. However, we would expect any bias to be similar in each randomised group. Also, our approaches to collecting resource use data did not disentangle resource use associated with NC from use associated with broader health factors.

Discussion

The BOOST programme improved walking capacity and physical performance and reduced walking disability and falls risk compared to a control intervention of BPA for older adults with NC at 12 months’ follow-up. There were also larger short-term improvements in symptoms and pain-related disability. We plan to optimise the programme to better target pain-related disability and symptoms for longer-term impact as part of the planned implementation. Exercise adherence decreased over time, and we will also focus on improving long-term adherence during optimisation.

The economic analyses suggest that the BOOST programme is a good use of NHS resources.

There were broadly similar clinical presentations between OPAL participants with NC and BOOST–OPAL participants (BP troublesomeness, comorbidities, multisite pain, change in walking and falls), meaning that we should be able to generalise findings from the BOOST RCT to the OPAL participants reporting NC. However, there were differences in socioeconomic factors and frailty, which may have been barriers to OPAL participants seeking treatment through the trial. This has implications for our implementation strategy to ensure the programme is available to all older people with NC.

With limited treatment options available to older people with NC, based on improvement in mobility and value for money, implementation of the BOOST programme should be considered.

Methods

A full description of methods90 is available here: https://doi.org/10.1136/bmjopen-2021-060128

We used a longitudinal qualitative design with semistructured interviews and open-ended questions. All BOOST RCT participants were eligible to take part. Consecutive participants were invited to take part. From those who agreed, we sampled participants to ensure that a range of age, gender, ethnicity, treatment allocation and recruitment sites were represented.

We conducted face-to-face interviews at participants’ homes after randomisation (T1), and telephone interviews (or face-to-face if requested) 1 month after the interventions (T2) and 12 months post randomisation (T3). Interviews were audio-recorded.

We aimed to recruit 60 participants, as we expected 50% attrition. We planned to analyse up to 30 participants interviewed at each time point. 91 Retention was better than expected, and 49/59 (82%) participants completed all three interviews (one withdrawal).

Results

Interviews from 16 men and 14 women were included in this analysis; 14 participants were allocated to BPA and 16 were allocated to the BOOST programme. The demographics of interview participants are included in Appendix 4, Table 19 but they were broadly similar with regard to age and gender. We interviewed a greater proportion of non-white participants compared to white participants in the trial to ensure we heard from participants from different backgrounds.

Limitations

We analysed data sets from a selected sample of 30 participants who provided a rich description of their experiences. The BOOST trial participants who did not take part in the interviews may have had different experiences from those described here. However, the strength of the study is that we used in-depth repeated interviews that focused on the participants’ current experience, so we were not reliant on their recollection of past experiences and can be confident that the data available reflected their experiences of the trial.

Discussion

Our interview findings resonate with the previous qualitative studies in NC,93,94 as pain was the main complaint impacting on their activities and leading to negative psychological responses. Expectations of trial participants were aligned well with the treatment targets of the BOOST programme. BOOST programme participants appeared more satisfied with their intervention compared to BPA. Dissatisfaction among BPA participants seemed to arise from lack of feedback and follow-up. BOOST participants benefited from peer support and discussions, which may have added to their satisfaction with the programme. Most participants felt the exercises were appropriate and helpful, although this did not necessarily translate to improvements in pain. Dissatisfaction with the BOOST programme was mostly related to lack of pain relief. Pain remained a substantial problem for some participants, and this highlights the importance of trying to better target pain. The BOOST programme participants talked about other improvements from the exercises, including improved posture, strength and confidence, highlighting the broader benefits of the BOOST programme. These types of changes were less evident in the narratives of BPA participants. BOOST participants also reported benefit from the cognitive–behavioural component of the programme, highlighting how the programme helped them to find their own solutions, manage flare-ups and understand the importance of long-term exercise.

Reasons for stopping the home exercises were similar between treatment arms. These included competing priorities, lack of motivation and aggravation of pain. Adaptations allowed some participants to continue at least some of the exercises. Some participants did not make adaptations when finding the exercises difficult or unhelpful, and stopped altogether. One participant had stopped the BOOST exercises to focus on participating in activities they enjoyed. These types of physical activity or exercise were not captured in the main trial follow-up questions, so it is not possible to know how many other participants may have gone on to other activities in lieu of the BOOST programme.

The barriers to long-term exercise identified in this study were not unique95,96 but highlight areas where we can optimise the BOOST programme. These include additional support to provide motivation for ongoing exercises, either by the physiotherapist or by linking to exercise opportunities in the community; to provide greater guidance on how to adapt exercises, rather than stopping altogether, if they are painful or if patients have a flare-up of symptoms; and to help plan integration into everyday life or transition to activities they enjoy for long-term adherence.

Methods

The full description of this study97 is available here: https://doi.org/10.1016/j.jclinepi.2022.09.002

The aim was to develop a tool to identify when older people were at risk of mobility decline. In particular, we wanted to understand whether BP was important, alongside other health conditions and symptoms, physical ability, psychological factors, and lifestyle and sociodemographic factors.

Results

There were 5358 participants who could walk at baseline and were eligible for this analysis. However, 174 (3.2%) died before the 2-year follow-up. Therefore, 5184 participants were included in the main model. Two-year follow-up data were provided by 4115/5184 (79.4%) participants. The mean age of participants was 74.7 years (SD 6.6), and 51.8% (n = 2683/5184) were women. At baseline, the majority of participants reported no mobility problems (3146/5184; 60.7%), and 5.4% (280/5184) reported severe mobility problems.

Among responders at 2 years, 18.6% (n = 765/4115) reported mobility decline (Figure 8). There were 375 (9.1%), 42 (1.0%) and 10 (0.2%) individuals who improved their mobility status by one, two and three points at 2 years, respectively. Participants who reported slight mobility problems (see lower panel, plot A) at baseline had the biggest decline over the 2 years of follow-up (35.6%).

FIGURE 8.

Absolute change in mobility status between baseline and year 2. Note: Participants who could walk at baseline (upper panel). Stratified by mobility status at baseline (lower panel): (A) no mobility problems; (B) slight mobility problems and (C) moderate or severe mobility problems. (From Sanchez et al. 97)

We examined the univariate relationship between the baseline variables and the outcome at 2 years (see Appendix 5, Table 23). Nearly all the selected baseline factors had a univariate relationship with the outcome. Among those with mobility decline at 2 years, there were more reports of back and leg pain and greater numbers reporting multisite pain and widespread pain at baseline. The biggest difference in pain presentation was lower limb pain at baseline, which was reported by 68% of participants with mobility decline at follow-up compared to 56% without. Baseline pain severity was also associated with mobility decline at follow-up.

The multivariable analyses found that in addition to mobility status (no problems or mild problems) at baseline, 13 variables were identified as predictors of mobility decline at 2 years in ≥ 80% of the multiple imputed data sets (Table 4). These included demographic and socioeconomic factors, multiple factors related to mobility, and general health-related factors. BP with and without leg pain was not associated with poor mobility in this multivariable model. However, lower limb pain and the presence of severe general pain were predictors of mobility decline.

The multivariable model revealed a median c-statistic of 0.740 (range 0.737–0.743), indicating moderately good discrimination. The median Brier score was 0.136 (range 0.135–0.137).

We developed a point scoring system, which produced the risk of mobility decline at 2 years (see Appendix 5, Table 24). The total score has a possible range from 0 to 39, with predicted risks ranging from 0.9% (0 points) to 90.1% (39 points). The median c-statistic of this scoring system was 0.734 (range 0.730–0.735), similar to that of the regression model, indicating no noticeable loss in performance. A hypothetical example is shown in Appendix 5, Tables 24–26.

After including participants who died during the 2 years’ follow-up, the model showed similar performance measures, but some differences in predictors (see Appendix 5, Table 27). Back and leg pain now featured in this model. Other factors included sex, occupation (very strenuous physical demand), hours moving (< 3 hours/day), fractures and unintentional weight loss as additional predictors. Pain severity and use of walking aid outside were no longer part of the model.

Limitations

The results are based on self-reported predictors, which may be prone to recall bias or not be as accurate as clinical assessment. Self-reported assessment of mobility using the walking item from the EQ-5D may not reflect the participants’ actual mobility status, but it was an indication of how they perceived their mobility. EQ-5D scores were consistent with other self-reported questions related to mobility in the OPAL cohort study, and its validity and reproducibility in older adults has been reported in several studies. 106,107 Self-reported predictors are particularly useful in clinical settings or in research when a detailed clinical assessment is impractical.

We combined participants with back and non-NC leg pain and those with NC leg pain into a single group to ensure we had adequate numbers in each of the BP groups for the analyses and to avoid including too many variables. This means we were unable to separate out the potential contribution of the different leg pain presentations in the analyses. We did not externally validate the models.

Discussion

We set out to understand the role of back and leg pain in mobility decline when considered alongside the broader perspective of health-related factors. We have identified a set of questions with a scoring system that could easily be applied in clinical practice to identify older people at risk of mobility decline.

Back and leg pain was associated with mobility decline in the univariable association but not in the multivariable model. Many of the factors associated with mobility decline were related to mobility. We saw in our previous analyses (see Objective 1) that back and leg pain was associated with mobility-related adverse outcomes, including confidence to walk and change in walking over the last year. The strongest association was with NC. Pain is likely to be a contributor to this mobility-related factor, and lower limb pain was identified as an independent predictor of mobility decline. General pain severity was also predictive, suggesting that the amount of pain experienced (regardless of the type of pain or area of pain) is important.

When considering possible interventions to reduce the risk of older people experiencing mobility decline, these findings suggest that pain needs to be addressed. While back and leg pain was not an independent predictor of mobility decline, it is important that it is not completely disregarded, as it may contribute to low walking confidence and mobility decline. We know from our earlier analyses of OPAL data that back and leg pain tend to be persistent and that pain management is often not prioritised by older people108 or by health professionals. Lower limb pain, such as that arising from hip and knee osteoarthritis, can be effectively managed through rehabilitation incorporating pain self-management, coping strategies and exercise. 109 The mobility-related factors identified can be addressed through rehabilitation. We have demonstrated in the BOOST RCT that walking ability can be improved through progressive tailored rehabilitation. These types of programmes should be made available to older people at risk of mobility decline.

Back and leg pain did feature in the sensitivity analyses, which included participants who had died. A possible explanation for this may be that the spine is a common site for metastatic cancer and so back and leg pain may be symptomatic of underlying disease. 110

Finally, a socioeconomic factor – a report of needing to be careful with money – was also identified in the risk assessment tool. The systematic literature review completed for this study found consistent evidence that low annual income and low number of financial assets were risk factors for mobility decline. 98 Our findings add to evidence of the impact of health inequalities due to socioeconomic factors on the health of older people. Income influences the choices that people make regarding their health, such as their diet or participation in physical activity and exercise, as well as their ability to access health care. 111 Addressing health inequalities has been identified as a key priority for the NHS,112 and ensuring adequate access to rehabilitation for older people needs to be part of this.

Implementation plan

We have developed a package of implementation for the BOOST programme which has been funded by the NIHR and is underway.

We consider the BOOST programme is worthwhile in its current format. We have presented the findings at a national conference (BritSpine) and held a seminar for 30 clinicians involved in the trial. Nearly all clinicians agreed it was worth implementing due to the associated clinical improvements and cost-effectiveness. Concerns about implementation were logistical as opposed to clinical; for example, colleagues working in small departments found it difficult to assemble sufficient patients to run the groups.

Implementation is guided by the i-PARIHS (Integrated Promoting Action on Research Implementation in Health Services) framework, a Knowledge Mobilisation theory (Figure 9). 113

FIGURE 9.

i-PARIHS framework.

Central to our implementation strategy is the development of a Massive Open Online Course (MOOC). On completion of the implementation study, the online course will be freely available here: https://learn.exeter.ac.uk/. We have used this approach previously. 114,115 The online training addresses implementation barriers as well as providing the content needed for successful implementation.

We are undertaking a two-stage implementation study. First will be an optimisation stage to improve the programme and launch a MOOC. The second stage will evaluate the MOOC learning outcomes and clinical outcomes of the BOOST programme delivered in routine NHS care.

The aims of the implementation study are to:

1. Optimise the content of the BOOST programme.

2. Design the online course and evaluate learning and clinical outcomes to ensure outcomes are the same or better than the original programme.

3. Establish the framework for a longer-term evaluation and implementation strategy.

The involvement of patients and/or the public

During the application process for this programme of work we assembled a PPI group, and we have continued to work closely with them. We advertised for PPI representatives via the INVOLVE website (https://involve.org.uk/) and approached organisations including the Rotary Club, the British Legion, Age UK, Men’s Sheds, BackCare and the British Orthopaedic Associations Patient Liaison Group. PPI representatives were also identified from participants in the preparatory interview study. 55 Ms Judith Fitch is the lead PPI representative and a co-applicant. We appointed a PPI representative to be an independent member of the Programme Steering Committee (Dr James Watson). Ms Fitch attended the co-applicants’ 6-monthly to yearly meetings to provide input into the progress of the programme of research. Dr Watson attended yearly Programme Steering Committee meetings.

Patient and public involvement engagement has been undertaken in face-to-face meetings, online meetings and via e-mails and phone calls. PPI representatives assisted with the development of the physiotherapy interventions, including attending the intervention refinement day, testing out the home exercises, assisting with development of the patient materials for the intervention, and posing as models for the exercise sheets.

Patient and public involvement representatives helped to develop the patient-facing materials. They provided feedback on layout and wording to make materials user-friendly. We piloted questionnaires with our PPI group, including completion of the OPAL baseline questionnaire by 20 older adults who spoke English as a second language to ensure suitability to participants from ethnic minorities. PPI representatives helped to develop the interview schedules and took part in practice interviews.

Patient and public involvement representatives helped to develop the prognostic tool. To widen the input we received, we advertised via the Oxford Biomedical Research Centre (BRC) and Oxford Collaborations in Leadership and Research in Health Care (CLARHC) PPI networks, the Patients Active in Research website (now known as ‘involve’: www.involve.org.uk/), and social media for more PPI representatives. We held a meeting in Oxford with seven PPI representatives and a meeting in Leeds with nine PPI representatives (aged 70–92 years). These discussions informed the final selection of the outcome (mobility question from EQ-5D-5L) and how we defined mobility decline. PPI representatives felt strongly that any decline (i.e. a single-step change downwards) needed to be considered important and included in the definition of decline. This is reflected in the analysis undertaken.

We met with our PPI contributors, who were very supportive of implementation of the BOOST programme, highlighting the importance of improving walking. Two PPI representatives are still involved in the BOOST implementation study. PPI representatives helped to write the plain English summary to disseminate findings.

Successes and challenges

Key findings

Back pain is a substantial and persistent problem for older people, associated with lower quality of life and age-related adverse health states including falls, frailty, low walking confidence and mobility decline. The impact is greatest when leg pain, especially NC, is reported. Although NC had the biggest impact, other presentations should not be ignored, due to the high proportion of participants who reported persistent symptoms.

The BOOST programme improved mobility and physical capacity, reduced falls in older people with NC and provided good value to the NHS. We have produced a manualised intervention that is ready for implementation. However, we only achieved short-term reduction in pain and pain-related disability. For a proportion of participants, pain remained problematic. On average, there were modest improvements in pain-related disability while participants were engaged with the programme and supported by the physiotherapist, but quantitative and qualitative data demonstrated the difficulty in maintaining long-term exercise adherence, which may have contributed to the reduced impact over time.

We have developed a prognostic tool to identify when older people are at risk of mobility decline. Back and leg pain had a univariable association with mobility decline but was not an independent predictor of mobility decline in the multivariable model. There were a variety of factors predictive of mobility decline in this model, including a range of domains (demographic and socioeconomic, mobility, general health and pain). We developed a scoring system for this tool so that it could be easily implemented in clinical practice, but it requires external validation.

Implications for practice and any lessons learnt

We have focused on NC, but there is also a need to consider how to better manage BP more generally in older adults. The majority of OPAL participants who reported BP at baseline had persistent symptoms. We are confident that the OPAL cohort is generalisable to community-dwelling older adults in England, so this represents a large proportion of the older population who are living with BP and not necessarily seeking health care. OPAL participants with symptoms suggestive of NC who did not take part in the BOOST trial reported lower educational attainment and lived in more deprived areas, but their clinical presentation was similar to those who did participate, suggesting they would have benefited from the BOOST programme. There is a need to develop interventions that could be delivered to the large number of older adults affected by BP, especially for those with more severe symptoms, and to understand how we can promote the uptake of interventions by people from lower socioeconomic groups to address health inequalities.

The improvement in clinical outcomes and cost-effectiveness of the BOOST programme indicates that implementation should be considered by clinicians and commissioners of services. However, we hope that we can better target pain during the planned implementation work. Treatment options are limited for this population, but there remains a need to help patients successfully manage their pain and to promote function and participation as part of healthy ageing. More consideration also needs to be given to how we support patients to transition from formal rehabilitation into lifelong habits of exercise and physical activity. There is a need for better linkage between rehabilitation services and community-based exercise provision and physical activity programmes.

We have focused on the development of a prognostic tool to identify mobility decline as this is a priority for older adults and key to maintaining independence in older age. This work highlighted a broader need to manage pain better in older people regardless of the type or presentation of pain. There were several mobility-related factors that were identified as predictors of mobility decline. These were present in some people who considered their mobility to be unaffected at baseline, so could be considered precursors to loss of mobility; they included slower walking speed, sometimes using a walking stick, low walking confidence and reduced balance. These are all modifiable factors that could be targeted through rehabilitation. The prognostic tool could be used not only by clinicians but also by older people and their families to spot signs of mobility decline and prompt them to seek rehabilitation rather than accepting it as part of old age. Matching treatments to risk factors may potentially slow mobility decline, and this requires evaluation.

Recommendations for future research

1. Develop and evaluate interventions that can be delivered at scale to the large number of older people who experience low BP, including consideration of health inequalities and how they impact on care-seeking by older people with BP.

2. Identify ways to achieve long-term improvements in pain and pain-related disability in older people with NC.

3. Externally validate the OPAL predictive tool.

4. Evaluate the clinical and cost-effectiveness of implementation of the OPAL predictive tool to reduce mobility decline among those at risk of mobility decline.

Conclusions

We have successfully undertaken a cohort study, a RCT and an embedded qualitative study. The assembled cohort and associated data remain a valuable asset for future trial recruitment and data analyses to further our knowledge about the role of pain, activity and comorbidities in the development of disability, frailty, falls and immobility in older adults. The prognostic tool will enable clinicians, their patients and families to identify when an older person is at risk of mobility decline. We have demonstrated the substantial impact that back and leg pain has on older people’s lives. We have produced an intervention for older adults with NC that improves mobility and physical capacity and reduces falls as well as being cost-effective, although we plan to further optimise the programme during implementation. Findings from this programme of research will inform future research focused on improving the lives of older people and the care they receive.

Contributions of authors

Esther Williamson (https://orcid.org/0000-0003-0638-0406) was a co-applicant and the programme lead for this project, oversaw its design and delivery and is lead author of this report.

Maria T Sanchez-Santos (https://orcid.org/0000-0003-1908-8623) undertook the analysis of the OPAL cohort data.

Ioana R Marian (https://orcid.org/0000-0002-0692-8112) conducted the statistical analysis of the BOOST RCT data.

Mandy Maredza (https://orcid.org/0000-0002-2030-3338) conducted the analysis of the health economic data.

Cynthia Srikesavan (https://orcid.org/0000-0002-3540-8052) conducted the analysis of the qualitative research data.

Angela Garrett (https://orcid.org/0000-0002-7271-6345) was the trial manager for the project and participated in the design, data collection and analysis of the study.

Alana Morris (https://orcid.org/0000-0002-3001-7888) was instrumental in the conduct of the OPAL cohort study including co-ordinating database searches and mail-outs and liaising with primary care practice staff.

Graham Boniface (https://orcid.org/0000-0003-1253-0060) was the lead research physical therapist who worked closely with BOOST study sites to ensure the BOOST interventions were correctly delivered, and supported the analysis of the qualitative research data.

Susan J Dutton (https://orcid.org/0000-0003-4573-5257) provided statistical oversight for the BOOST RCT.

Frances Griffiths (https://orcid.org/0000-0002-4173-1438) was a co-applicant and lead qualitative researcher on the study.

Gary S Collins (https://orcid.org/0000-0002-2772-2316) was a co-applicant who aided the statistical design and analysis of the study.

Stavros Petrou (https://orcid.org/0000-0003-3121-6050) was a co-applicant and lead health economist on the study.

Julie Bruce (https://orcid.org/0000-0002-8462-7999) was a co-applicant who aided the design and analysis of the study.

Jeremy Fairbank (https://orcid.org/0000-0002-0050-874X) was a co-applicant, who provided expertise on NC that aided the design and conduct of the study.

Zara Hansen (https://orcid.org/0000-0001-8915-7765) was a co-applicant, who aided the design and delivery of the study.

Karen Barker (https://orcid.org/0000-0001-9363-0383) was a co-applicant, who aided the design and delivery of the study.

Charles Hutchinson (https://orcid.org/0000-0003-3387-9229) was a co-applicant and lead radiologist on the study.

Christian Mallen (https://orcid.org/0000-0002-2677-1028) was a co-applicant who aided the design and analysis of the study.

Lesley Ward (https://orcid.org/0000-0001-7285-2032) contributed to developing the BOOST interventions and undertook the participant interviews for the qualitative study.

Richard Gagen (https://orcid.org/0000-0002-2750-1011) was instrumental in the collection and analysis of the BOOST MRI data.

Judith Fitch (https://orcid.org/0000-0002-6687-5200) was a co-applicant who aided the design and implementation of the study from the patients’ perspective.

David P French (https://orcid.org/0000-0002-7663-7804) was a co-applicant who aided the design and analysis of the study.

Sallie E Lamb (https://orcid.org/0000-0003-4349-7195) was the chief investigator. She led the team that secured funding, developed interventions and undertook the studies reported. She is guarantor for the data and approved the final version of the manuscript.

The BOOST programme was conducted as part of the portfolio of trials in the registered UKCRC OCTRU at the University of Oxford. It has followed their standard operating procedures, ensuring compliance with the principles of good clinical practice and the Declaration of Helsinki and any applicable regulatory requirements. We would therefore like to thank them for their support and guidance with this work.

Publications

Ethical statement

Ethical approval for the OPAL cohort study and the BOOST trial was provided by the London–Brent Research Ethics Committee (16/LO/0348) on 10 March 2016.

Data-sharing statement

All data requests should be submitted to the Chief Investigator of this programme of research for consideration. Access to anonymised data may be granted following review. Requests can also be made to Professor Sallie Lamb via e-mail s.e.lamb@exeter.ac.uk.

Presentations

Lyle SA, Williamson E, Darton FC, Griffiths FE, Lamb SE. From Data Collection to Intervention development: A qualitative study of older people’s experiences of living with neurogenic claudication. Oral Presentation: 14th International Forum for Back and Neck Pain Research in Primary Care. 2016.

Boniface G, Williamson E, Ward L, Hansen Z, French D, Rogers D, Lamb SE. Development of the BOOST Programme: a physiotherapy intervention addressing the physical and psychological impact of neurogenic claudication in older adults. Oral Presentation: 16th International Forum for Back and Neck Pain Research in Primary Care. July 2019.

Boniface G, Williamson E, Ward L, Hansen Z, Gandhi V, Nicolson P, Garrett A, Lamb SE. Ensuring treatment fidelity to a group based physical and psychological intervention for older adults with neurogenic claudication. Poster Presentation: 16th International Forum for Back and Neck Pain Research in Primary Care. July 2019.

Ward L, Griffiths F, Williamson E, Robinson R, Boniface G, Lamb SE. Developing a framework for the analysis of longitudinal qualitative data in clinical trials. Oral Presentation: 16th International Forum for Back and Neck Pain Research in Primary Care. July 2019.

Ward L, Griffiths F, Williamson E, Boniface G, Lamb SE. The physical and psychological impact of neurogenic claudication on older adults’ responses to physiotherapy. Oral Presentation: 16th International Forum for Back and Neck Pain Research in Primary Care. July 2019.

Ward L, Lamb SE, Williamson E, Robinson R, Griffiths FE. Drowning in data! Designing a novel approach to longitudinal qualitative analysis. Oral Presentation: 4th Qualitative Health Research Network (QHRN) Conference. March 2019.

Williamson E, Boniface G, Marian IR, Dutton SJ, Maredza M, Petrou S, Garrett A, Morris A, Hansen Z, Ward L, Nicolson P, Barker K, Fairbank J, Fitch J, Rogers D, Comer C, French D, Mallen C, Lamb SE. The clinical and cost-effectiveness of a physiotherapy delivered physical and psychological group intervention for older adults with neurogenic claudication: the BOOST randomised controlled trial. Oral Presentation: 17th International Forum for Back and Neck Pain Research in Primary Care. Nov 2021.

Williamson E, Garrett A, Marian IR, Dutton SJ, Maredza M, Petrou S, Boniface G, Hansen Z, Nicolson P, Barker K, Lamb SE. The prevalence of back pain with and without associated leg pain in older adults in England and the association with immobility, falls and frailty. Oral Presentation: Chartered Society of Physiotherapy (CSP) Conference. Nov 2021.

Disclaimers

This report presents independent research funded by the National Institute for Health and Care Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, CCF, PGfAR or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, the PGfAR programme or the Department of Health and Social Care.

The delivery of the BOOST programme

We randomised 438 participants, with 295 allocated to the BOOST programme. There were three complete study withdrawals (including use of any data) from the BOOST programme arm. Two were due to health problems and one because the individual would be unable to attend any treatment sessions. Therefore, 292 participants allocated to the BOOST programme are reported on here. Physiotherapists completed an intervention log for each participant. Each log acted as a structured record of the intervention and was used to monitor fidelity and adherence with the interventions.

We also carried out observation of intervention sessions. A structured checklist (see Report Supplementary Material 1) was used to assess the delivery of the core elements of interventions, which was scored as: not completed, partially completed or fully completed. Initial observations were used to provide feedback and support physiotherapists to deliver the interventions. Later in the trial, these visits were fidelity assessments to understand how the intervention would be implemented in a real-world clinical setting with no feedback to the physiotherapists.

Baseline setting

Baseline setting of the four strength exercises (sit to stand, knee extension, hip abduction, hip extension) was determined using the modified Borg scale (scored 5–6/10: ‘the exercise feels hard’), using weights to achieve the required intensity if indicated.

The most common prescription at baseline was two sets of 10 repetitions for each exercise.

For the sit to stand exercise, 165/279 (59.1%) participants had weight prescribed. The most common starting weight was 4.5 kg.

For the knee extension exercise, 222/279 (79.1%) participants had weights added. The most common starting weight was 1.5 kg.

Just under three-quarters of participants (196/279; 71.0%) had weight added for both hip abduction and hip extension. The most common starting weight was 1.5 kg.

In a very small number of participants (4/279; 1.4%) one or more of the strengthening exercises was not prescribed. One participant was deemed unsuitable for the groups so was not prescribed any of the exercises. This participant was withdrawn from the intervention by their physiotherapist due to health problems. They were experiencing dizziness and ‘funny turns’ and referred back to their general practitioner.

In the other three participants, the following were not given: sit to stand, n = 2; knee extension, n = 1; hip abduction, n = 3 and hip extension, n = 1. Reasons for not prescribing the exercises included recent surgery and pain.

The most frequently prescribed baseline balance exercise was the full tandem stance (n = 115.0; 41.4%). This was followed by semi-tandem stance (n = 83.0; 29.9%), parallel stance (n = 25.0; 9.0%) and forward tandem walking (n = 47.0; 16.9%). See Appendix 3, Figure 10.

FIGURE 10.

Progressive balance exercises.

The flexibility exercise (hip flexor stretch) had a recommended default baseline setting of three 10-second holds per leg. All participants were given this at baseline.

Baseline walking levels for the programme were based on participant-rated ability and confidence and the physiotherapist’s observations during the appointment. Tailoring of the walking component of the programme included walking duration (including rest stops), use of obstacles, walking speed, carrying of hand weights, or a combination of these. The majority of participants (214/276; 77%) were prescribed 20 minutes’ walking only to start the group programme (see Appendix 3, Figure 16). Two participants had the baseline walking missing from their treatment log, and one was withdrawn from the programme.

At the baseline appointment, walking aids were prescribed if indicated. Nine participants (3.2%) were prescribed a walking aid at the baseline appointment. Six participants (2.2%) were prescribed with a standard walking stick, two participants (0.7%) were prescribed with an outdoor four-wheeled rollator frame, and one participant (0.3%) was prescribed with two anatomical-fit (Fischer-type) walking sticks.

Exercise progression

Strengthening exercises

Over the duration of the group sessions (12 sessions over 12 weeks), participants’ exercises were progressed by manipulating sets, repetitions, and load. Physiotherapists also introduced power (perform the exercise faster) to some participants to make the exercises more challenging.

We calculated the dose completed at each session by multiplying sets × repetitions × load (kg). If the participant completed an exercise without load (e.g. two sets, 10 repetitions), an arbitrary load value of 0.1 kg was used. Where an exercise was performed on the right leg and then the left leg, if the load differed then we took the lowest number for this calculation.

This allowed us to plot the progressions over time. We have plotted the median dose (IQR represented by the line) in Appendix 3, Figures 11–14 for each of the strengthening exercises. The number of participants contributing data is displayed above each data point. We have also plotted participants classified as compliers (nine sessions or more) or non-compliers.

FIGURE 11.

Strengthening exercise: sit to stand prescription.

FIGURE 12.

Strengthening exercise: knee extension prescription.

FIGURE 13.

Strengthening exercise: hip abduction prescription.

FIGURE 14.

Strengthening exercise: hip extension prescription.

Overall, the BOOST programme participants were progressed across the 12 group sessions. Across the four exercises, compliers were prescribed a greater dose than non-compliers. Non-compliers still showed progression over time, but it was more variable than for those classified as compliers.

Speed (power) was also used to progress the exercises. Very few participants were asked to add speed to their movements at baseline; this tended to be added towards the later stages of the programme. At session 12, around a quarter of participants were adding speed to knee extension, hip abduction and hip extension exercises and one-fifth to sit to stand (see Appendix 3, Table 17).

TABLE 17 - Proportion prescribed a power element for strengthening [n (%)]

Exercise	Baseline	Session 1	Session 2	Session 3	Session 4	Session 5	Session 6
Exercise	Baseline	Session 7	Session 8	Session 9	Session 10	Session 11	Session 12
Sit to stand	3.0 (1.1)	3.0 (1.1)	5.0 (1.8)	10.0 (4.5)	11.0 (5.3)	23.0 (10.1)	16.0 (7.7)
Knee extension	6.0 (2.2)	1.0 (0.4)	4.0 (1.8)	18.0 (8.2)	12.0 (5.7)	25.0 (11.0)	16.0 (7.7)
Hip abduction	3.0 (1.1)	5.0 (2.2)	6.0 (2.7)	13.0 (5.9)	13.0 (6.2)	23.0 (10.1)	9.0 (4.3)
Hip extension	5.0 (1.8)	2.0 (0.9)	5.0 (2.3)	13.0 (5.9)	10.0 (4.8)	22.0 (9.7)	6.0 (2.9)
Sit to stand		20.0 (7.2)	16.0 (5.7)	35 (17.7)	25.0 (12.9)	39.0 (20.7)	40.0 (19.8)
Knee extension		21.0 (9.5)	17.0 (8.3)	35.0 (17.6)	27.0 (13.8)	47.0 (25.0)	54.0 (26.7)
Hip abduction		15.0 (6.8)	13.0 (6.4)	28.0 (14.1)	27.0 (9.7)	37.0 (19.7)	51.0 (25.4)
Hip extension		15.0 (6.8)	13.0 (6.4)	28.0 (14.1)	23.0 (11.8)	40.0 (21.2)	52.0 (25.7)

Balance and flexibility exercises

Over the 12 group sessions, the balance exercise was progressed with participants, with the majority of participants moving from static balance exercises to dynamic balance exercises (see Appendix 3, Figure 15).

FIGURE 15.

Balance exercise prescription.

Participants were also holding their stretches for longer towards the end of the 12 group sessions. At baseline, the median total stretch time was 30 seconds (IQR 30–30); at session 12, participants were holding the stretch for a median total of 90 seconds (IQR 60–90).

Walking programme

Over the 12 group sessions, participants progressed their walking during the supervised circuit and were undertaking a variety of added challenges by session 12 compared to baseline, where most were only walking (see Appendix 3, Figure 16).

FIGURE 16.

Walking programme prescription.

Walking aids prescribed in the group sessions

During the group sessions, seven participants (2.5%) were prescribed a walking aid (session 1: n = 3; session 8: n = 1; session 9: n = 1; session 11: n = 2). One participant (0.4%) who was prescribed a walking stick at their baseline appointment was also prescribed an outdoor three-wheeled rollator frame during the group sessions to use. Of the remaining six participants, three (1.1%) were prescribed standard walking sticks, one (0.3%) with an anatomical-fit (Fischer-type) walking stick and two (0.7%) with rollator frames (indoor use: n = 1; outdoor use: n = 1). One participant (0.3%) started to use their mother-in-law’s rollator frame from session 11 onwards.

Non-completion of exercises

There were occasions when participants attended a group but did not complete some or all of the exercises during the group session (see Appendix 3, Table 18). The most common reason for non-completion overall was participants staying only for the education component or having to leave early due to another commitment. We delivered 53 groups over a total of 636 sessions, so the proportion of sessions where participants did not complete exercises was very small.

TABLE 18 - Reasons for exercise non-completion. The number of participants reporting each reason (total number of occasions this exercise was not completed by a participant)

	Sit to stand	Knee ext	Hip abduction	Hip extension	Balance	Stretch	Walking circuit
Reason missing	4 (5)	3 (4)	5 (7)	2 (2)	4 (4)	8 (9)	12 (12)
Attended education session only or needed to leave early – had another appointment or commitment	5 (8)	5 (8)	7 (10)	8 (11)	6 (9)	6 (9)	14 (17)
Attended education session only – recent cataract surgery	1 (4)	1 (4)	1 (4)	1 (4)	1 (4)	1(4)	1 (4)
Discomfort from recent unrelated surgery/procedure	1 (1)						2(2)
Attended education session only due to feeling unwell	4 (5)	4 (5)	4 (5)	4 (5)	4 (5)	4 (5)	4 (5)
Pain	2 (2)	2 (2)	5 (6)	4 (4)	2 (2)	1 (1)	2(2)
Tired	1 (1)	1 (1)	1 (1)	1 (1)	2 (2)		2 (2)
Heart condition and shortness of breath				1(1)
Planned to do it at home					1 (1)
Could not manage it						1 (1)
Arrived late so did not do it							2 (2)
Dizzy							1 (1)
Had walked before class or planning to walk later							5 (8)
Too hot							4 (4)
Total	18 (26)	16 (24)	23 (33)	21 (28)	20 (27)		49 (59)

The sit to stand exercise was not completed by 18 participants on a total of 26 occasions. Knee extension was not completed by 16 participants on a total of 24 occasions. Hip abduction was not completed by 23 participants on a total of 33 occasions. Pain was not commonly reported as a reason for non-completion, but it was reported more often for hip abduction than for the other exercises. Hip extension was not completed by 21 participants on a total of 28 occasions. The balance exercises were not completed by 20 participants on a total of 27 occasions.

The walking circuit was not completed most frequently out of the exercises. It was not done by 49 participants on a total of 59 occasions. The most common reason was participants staying only for the education component or having to leave early due to another commitment. This probably affected the walking circuit the most as it was the final part of the programme. Some participants also preferred to do their walking before or after the groups (e.g. to walk their dog).