Better post-operative prediction and management of chronic pain in adults after total knee replacement: the multidisciplinary STAR research programme including RCT

Article history

The research reported in this issue of the journal was funded by PGfAR as project number RP-PG-0613-20001. The contractual start date was in October 2015. The final report began editorial review in May 2021 and was accepted for publication in June 2022. 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 Gooberman-Hill et al. This work was produced by Gooberman-Hill 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 adaption 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 Gooberman-Hill et al.

Work package 1: systematic reviews

Our systematic review of post-operative risk factors for chronic pain after total knee replacement included 14 studies published up to October 2016, with data from 1168 people. Studies focused on acute pain, function and psychological factors. Risk factor measures and outcomes were heterogeneous. In a narrative synthesis we were unable to draw firm conclusions on potential interventions. The need for further prospective studies in representative populations was clear.

Research published up to December 2018 into pre-operative interventions mainly focused on exercise and education. In the eight trials, with a total of 960 people randomised, there was no association with these interventions and long-term pain outcomes. In the peri-operative setting, we identified 44 trials published up to February 2018, with a range of 10 to 280 people randomised. Unifactorial interventions including some forms of analgesia, early rehabilitation, electrical muscle stimulation and anabolic steroids were associated with improved long-term pain outcomes. However, studies were small and merit further evaluation. There was reassurance that some common peri-operative treatments are not associated with chronic pain. Post-operative interventions evaluated in 17 trials published up to November 2016, with a total of 2485 people randomised, mainly focused on physiotherapy. There was no strong evidence favouring one format of therapy over another.

There has been little research into treatments for chronic pain after total knee replacement. Considering interventions for general chronic post-surgical pain, we identified 66 randomised trials with a total of 3149 participants in our systematic review with searches up to March 2016. A more focused updated search including treatments for chronic pain after arthroplasty of the large joints was conducted in October 2020. Many unifactorial interventions have been evaluated, and specific nerve-focused treatments deserve further research.

Work packages 1 and 2: analysis of national databases

We undertook two analyses of linked databases to identify pre-, peri- and post-operative risk factors for chronic pain outcome. In the first analysis with NJR and HES data, the pre- and 6-month-post-operative Oxford Knee Scores (OKS) was available for 258,386 patients, 43,702 (16.9%) of whom were identified as having chronic pain at 6 months post surgery. Post-surgical predictors of chronic pain were mechanical complication of prosthesis, surgical site infection, readmission, reoperation, revision and an extended hospital stay. However, these post-surgical predictors explained only a limited amount of variability in chronic pain outcome.

In the second analysis, we analysed primary care data from CPRD and secondary care data from the HES–PROMs database and included 4570 patients. At 6 months after surgery, 10.4% of patients were classified as non-responders to surgery regarding their knee pain. Expressing the effects as absolute risk differences allowed us to quantify the relative importance of individual risk factors in terms of the absolute proportions of patients achieving poor pain outcomes. Pre-operative risk factors were having only mild knee pain symptoms, currently smoking, living in the most deprived areas, having a body mass index between 35 and 40 kg/m² and having had previous knee arthroscopy surgery. Post-operative risk factors were revision surgery and manipulation under anaesthetic within 3 months after the operation, and use of opioids and antidepressants within 3 months after surgery.

Work package 2: long-term follow-up of COASt cohort and analysis of national databases

We characterised the long-term trajectory of chronic pain, including pain characteristics and resource use, through the 5-year follow-up of the COASt cohort of 1581 patients with total knee replacement, and analysis of the linked CPRD and HES databases.

We applied cluster analysis to data on 128,145 patients with primary total knee replacement included in the English PROMs programme to derive a cut-off point on the pain subscale of the OKS. A high-pain group was identified, defined as those with a score of ≤ 14 points on the OKS pain subscale 6 months after total knee replacement. About one in eight people experienced chronic pain 1 year after total knee replacement. Of these patients with chronic pain after surgery, after imputing significant missing data assumed to be missing at random, 65% experienced no-chronic-pain by year 5, 31% fluctuated and 4% remained in chronic pain. People with chronic pain in year 1 had worse quality of life to start with; this improved, but less rapidly than for those not in chronic pain. People with chronic pain reported slightly higher primary care consultation costs than those not in chronic pain but their prescriptions for analgesia were much more frequent, more costly to the health-care system and continued to grow even after surgery, especially prescriptions for opioids.

Work package 3: finalisation of an assessment protocol and care pathway

We refined and finalised the novel STAR care pathway and associated training materials. The STAR care pathway involves a clinic appointment for patients who have troublesome pain at 3 months after surgery. A specially trained extended scope practitioner (ESP) conducts a clinic assessment with the patient, comprising history, examination, radiography and questionnaire completion. Based on this assessment, which focuses on understanding the reasons for and impact of the pain, the patient is referred to the appropriate existing services for treatment, such as a surgeon, general practitioner (GP) or specialist, or receives ongoing monitoring. The ESP follows up with patients by telephone for up to 12 months.

Work package 4: randomised controlled trial

In a multicentre pragmatic, open randomised controlled trial, we evaluated the STAR care pathway. We screened 5036 people, randomised 363 patients with pain at 3 months after knee replacement from eight NHS Trusts in England and Wales and collected 12-month outcomes from 313 (85%) randomised participants. The sample had a mean age of 67 years, was 60% female and 94% white. Our analysis of clinical effectiveness indicated that at 12 months the intervention arm had lower mean pain severity and lower mean pain interference than the usual care arm. For pain interference at 12 months, the adjusted difference in means was –0.68 points on the Brief Pain Inventory pain interference scale [95% CI –1.29, –0.08; p = 0.026]. For pain severity at 12 months, the adjusted difference in means was –0.65 points on the Brief Pain Inventory pain severity scale [95% CI –1.17, –0.13; p = 0.014]. Our analysis of cost-effectiveness indicated that people receiving the STAR pathway from an NHS and personal social services perspective had lower costs (–£724, 95% CI -£1500 to £51) and more quality-adjusted life-years (QALYs) (0.03, 95% CI –0.008 to 0.06) than those receiving usual care. The STAR pathway was the cost-effective option: the incremental net monetary benefit at the £20,000-per-QALY threshold was £1256 (95% CI £164 to £2348). This was also the case from a patient perspective. Embedded qualitative research found that patients thought that the STAR pathway was acceptable, and patients described how it provided an opportunity for them to discuss their concerns and to receive more information about their condition while ensuring they received further treatment and ongoing support.

Work package 5: qualitative study

In semistructured interviews with 34 people, we found that people with chronic pain after total knee replacement who made little or no use of services did so because they became stuck in a cycle of appraisal of the validity of their need for help and concern that treatment may not be of benefit. Some were concerned that further treatment may even worsen their pain or cause further harm. When describing chronic post-surgical pain, some participants described sensations of discomfort including heaviness, numbness, pressure and tightness associated with the prosthesis, and some also reported a lack of felt connection with their knee as their movement was no longer natural and required deliberate attention, and that they had a lack of confidence in it.

Work package 6: implementation and dissemination

We found that health-care professionals involved in the delivery and implementation of the STAR care pathway valued its focus on the identification of neuropathic pain and psychosocial issues, enhanced patient care, formalisation and validation of referral practices and an increased knowledge of pain management. Stakeholders supported formal implementation of the STAR pathway. Whether or not this would be supported by hospital management was felt to be dependent on whether or not it was shown to be cost-effective.

Conclusions: implications for health care

After knee replacement, screening for pain with the OKS pain subscale beginning at 2 months after surgery can facilitate the delivery of targeted care from 3 months. Our findings indicate that the STAR care pathway can provide improved care and outcomes for people who have pain after knee replacement. To our knowledge, the STAR care pathway is the first multifactorial intervention for the treatment of post-surgical pain to have been evaluated in a randomised controlled trial. In database analyses and systematic reviews, we identified risk factors for and univariable interventions to prevent or treat chronic pain. After further research these may provide additional components to the care pathway.

Our work also indicates that people with pain could be empowered to seek health care and that health-care professionals can be encouraged provide support. This could include information for people living with chronic pain to inform them that health care may provide benefit and that seeking care is not futile. Informing patients of the likely outcomes after surgery may be a key part of pre- and post-surgical care.

Recommendations for research

We recommend that further research addresses the following points, numbered in descending order of priority:

1. How to implement the STAR care pathway into the NHS.

2. How to improve communication between patients and professionals before surgery.

3. Whether or not patient education and supportive care can enable earlier recognition of chronic pain.

4. The STAR care pathway showed benefit to patients for both pain and interference at 6 and 12 months. Further follow-up would describe the longer-term outcomes of this intervention and the health-care resources utilised by participants.

5. How to reshape the STAR pathway for other surgeries.

6. The STAR programme focused on care after surgery. Future research could make use of the recently developing evidence base about the time before surgery as an opportunity for intervention. Specifically, we now have a greater understanding of risk factors for poor outcome and using this understanding to design and evaluate pre-surgical intervention may prove of long-term benefit to patients and health-care systems.

7. How to better manage patient’s feelings of disconnectedness from the new knee and sensations of otherness to improve incorporation of the prosthesis.

8. Promising interventions, identified in systematic reviews and suggested by our risk factor studies, should be evaluated in appropriately powered high-quality randomised controlled trials.

9. New interventions with evidence of effectiveness in the treatment of chronic pain after knee replacement should be considered as new components of multifaceted personalised care as delivered in the STAR intervention.

Study registration

All systematic reviews were registered on PROSPERO (CRD42015015957, CRD42016041374 and CRD42017041382). The STAR randomised trial was registered as ISRCTN92545361.

Funding

This project was funded by the National Institute for Health and Care Research (NIHR) Programme Grants for Applied Research programme and will be published in full in Programme Grants for Applied Research; Vol. 11, No. 3. See the NIHR Journals Library website for further project information.

Total knee replacement

Osteoarthritis is the most common joint disease, affecting nearly 10% of adults in the UK1 and about 23% of adults in the USA. 2 The prevalence of knee osteoarthritis depends on its definition: international estimates vary from 8.2% (presence of symptoms) to 9.3% (self reported) to 31.7% (radiographic changes). 3

The primary reason that people choose to undergo total knee replacement is the expectancy of pain relief. 4 In 2019, over 100,000 primary total knee replacements were performed by the NHS,5,6 and it is estimated that about 11% of women and 8% of men will receive a knee replacement during their lifetime. 7

For many people with advanced osteoarthritis, total knee replacement is an effective treatment to relieve pain and improve function. However, some people experience continuing pain in the months and years following surgery.

The STAR programme

The overall aim of the STAR programme was to generate high-quality evidence about how to improve health care and outcomes for people with chronic pain after total knee replacement.

This programme provided the opportunity to conduct a major programme of work addressing multiple important themes. The programme comprised six interconnected work packages, which aimed to improve health care and outcomes for patients with chronic pain after total knee replacement. All work packages were underpinned by collaborative working with patients and full details of patient and public involvement (PPI) are reported in Patient and public involvement.

Specifically, the programme aims were as follows:

1. Synthesise evidence on the prevention of chronic pain after knee replacement and the treatment of chronic pain after diverse surgeries through systematic reviews of published studies. Identify post-surgical predictors of chronic pain after knee replacement through a systematic review and analysis of data from the National Joint Registry (NJR), linked to the Hospital Episode Statistics (HES) and Patient Reported Outcome Measures (PROMs) databases (work package 1).

2. Characterise the long-term trajectory of chronic pain, including an examination of pain characteristics and resource use, through extended follow-up of an existing cohort of patients with total knee replacement up to 5 years after surgery and analysis of the Clinical Practice Research Datalink (CPRD), linked to the HES database (work package 2).

3. Refine and finalise a novel care pathway for patients with chronic pain after total knee replacement through consensus work with health-care professionals. Test intervention delivery and acceptability to patients and evaluate views about implementation of the intervention at future trial centres using the Normalisation Measure Development (NoMAD) tool (work package 3).

4. Evaluate the clinical effectiveness and cost-effectiveness of the new care pathway for patients with chronic pain after total knee replacement in a pragmatic, open-label, parallel group, multicentre superiority randomised controlled trial with a 2: 1 allocation ratio and embedded economic evaluation and qualitative studies (work package 4).

5. Identify reasons for the non-use of services and how to improve access through a qualitative study with patients living with chronic pain after total knee replacement who make little or no use of formal health-care services (work package 5).

6. Conduct a process evaluation of the implementation of findings into clinical practice and distribute evidence-based information about the identification, assessment and management of chronic pain after total knee replacement (work package 6).

A research pathway diagram is provided in Figure 1.

FIGURE 1.

The STAR programme.

Background patient and public involvement work leading to the programme

The STAR programme was developed in collaboration with the University of Bristol’s Musculoskeletal Research Unit patient involvement group, called the Patient Experience Partnership in Research (PEP-R). 49 PEP-R comprises people with experience of musculoskeletal conditions, including pain after surgery. STAR was also developed with input from a representative of Versus Arthritis, the UK’s largest charity dedicated to supporting people with arthritis. Ongoing collaboration with Versus Arthritis shaped the research and provided input into the design of the key implications for research and practice. The STAR programme was preceded by a 1-year NIHR Programme Development Grant that included PPI work. PPI informed the overall design of the programme, with specific input into the study of post-operative predictors and the trial design, including the acceptability of randomisation, timing of data collection, primary outcomes and questionnaire length. The PEP-R group also discussed and approved the PPI plans, specifically requesting a model of ‘taking the research to the patient’ rather than ‘taking the patient to the research’.

The approach to PPI was based on principles of coworking and partnership, in which we aimed to empower the individual patient partners by considering them part of the research team. We worked together to design PPI that resulted in patients having a considerable input into the relevance and quality of our research. 83,84 Our reporting of PPI is in keeping with the recommendations of the Guidance for Reporting Involvement of Patients and the Public 2 (GRIPP-2) short form. 85

Patient and public involvement during the programme

To complement PEP-R’s ongoing engagement, we established a dedicated forum for the programme, comprising four patients with experience of chronic pain after knee replacement. Forum members were offered support and training by an experienced PPI co-ordinator. They received training built into forum sessions, including overviews of qualitative research, systematic reviews, statistics and trial management. Patients were also given information about local PPI events.

During the programme, there were a total of 23 meetings of the STAR patient forum and two one-to-one discussions between the PPI co-ordinator and individual members. STAR was also discussed at six meetings of the PEP-R patient group. Patients were provided with reimbursement for their time in the form of shopping vouchers, as this was their preference. Travel was either reimbursed or arranged for patients to attend meetings. From April 2020 onwards, face-to-face STAR and PEP-R forum meetings were cancelled owing to COVID-19 and we continued to work with the PPI groups remotely. A member of PEP-R attended Programme Steering Committee meetings with the PPI coordinator. STAR PPI activity and its impact in each work package is summarised in Appendix 1 and Box 1.

Summary of the value of patient and public involvement in the programme

Includes information published in Bertram et al. 86 and Bradshaw et al. 87

Patient and public involvement was an essential part of the programme. We involved patient partners at all stages and this had a significant impact on the study design, such as improvements to patient documents, recruitment and retention methods,86 the communication of results and the planning of next steps (Box 1). Their involvement ensured that patients’ voices were included in the design and delivery of this research and that the outputs were relevant and meaningful to them.

Our patient partners also had an impact on the design of the PPI, as we followed their advice to adopt an approach of ‘taking the research to the patient’, rather than having two PPI members simply attending less frequent and more formal management or advisory group meetings. This had a positive impact on PPI group members as they believed this resulted in improved equality of power and decision-making, an atmosphere of strong collaboration and respect, and the opportunity for them to have a real impact on research. In addition, the research benefited through allowing a greater number of patients with differing experiences to be involved, providing a wider insight into issues of greatest importance to people in chronic pain. The design of the PPI was an ongoing discussion between group members and the PPI coordinator. The STAR forum group members told us that they felt they were working in partnership with the research team and equality was achieved by having the researchers come to them. This approach also had an impact on the research team as they all had the opportunity to attend the regular PPI group meetings and work in collaboration with patients. For some researchers, this was their first opportunity to hear about the experiences of patients with long-term pain and explain their research in plain English face to face. We recommend this approach and continue to use it as our preferred approach to PPI within the Musculoskeletal Research Unit (Translational Health Sciences, Bristol Medical School, University of Bristol, Bristol, UK).

Background

Although pre-operative risk factors for chronic pain after total knee replacement have been explored extensively, post-operative risk factors have not.

Treatments in the pathway through total knee replacement may potentially modify risk factors for poor patient outcomes and adverse events. These have been reviewed previously, but not with an emphasis on chronic pain.

Little research has focused on the treatment of chronic pain after knee replacement. 48 Interventions for the management of chronic pain after other surgeries may have value in the context of total knee replacement.

Risk factors for chronic pain after total knee replacement: systematic review

This section has been published as Wylde et al. 88

Aims

This systematic review aimed to identify early post-operative patient-related risk factors for chronic pain after total knee replacement.

Methods

A prospectively registered systematic review (PROSPERO CRD42016041374),89 was conducted and reported as recommended by PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. 90 We searched MEDLINE, EMBASE and PsycINFO from inception to October 2016 with no language restrictions (see Appendix 2, Systematic review search strategy as applied in MEDLINE on Ovid).

Eligible studies were longitudinal in design and met the following inclusion criteria:

1. patients observed within 3 months of knee replacement

2. intervention group – people with a risk factor

3. control group – non-exposed people

4. outcome – chronic pain at ≥ 6 months after knee replacement

Risk of bias was assessed with a non-summative checklist (Box 2),87 based on components of the methodological index for non-randomized studies (MINORS)91 and Newcastle-Ottawa Quality Assessment Scale. 92 Results were reported as a descriptive narrative analysis.

BOX 2 - Four-point system for assessing risk of bias in cohort studies

The following checklist components were rated as adequate, inadequate or not reported:

1. inclusion of consecutive patients (consecutive = adequate)

2. representativeness (multicentre = adequate)

3. per cent follow-up (> 80% = adequate)

4. minimisation of potential confounding (multivariate analysis = adequate).

Results

Searches identified 14 cohort studies (1168 patients) evaluating the association between patient-related factors in the first 3 months post operation and pain at ≥ 6 months after primary total knee replacement (see Appendix 2, Figure 2).

FIGURE 2.

Risk factors for chronic pain after total knee replacement: systematic review flow diagram. TKR, total knee replacement.

Post-operative patient-related factors evaluated included acute pain (eight studies), function (five studies) and psychosocial factors (four studies). In studies with no risk of bias other than patient selection, there was a suggestion that acute post-operative pain during the hospital stay was associated with chronic pain. 52,93 However, in one of these studies, the association was largely explained by pre-operative pain. 52 For all other post-operative patient factors, there was insufficient evidence to draw firm conclusions about an association with chronic pain after total knee replacement.

Effectiveness of interventions to prevent chronic pain after total knee replacement: systematic reviews

These reviews have been published as Wylde et al.,100 Beswick et al.,101 and Dennis et al. 102 A comprehensive overview has been published as Wylde et al. 103

Aim

These systematic reviews aimed to assess the effectiveness of pre-, peri- and post-operative interventions in preventing chronic pain in patients receiving total knee replacement.

Workshop

In March 2016, 57 invited experts and colleagues met with the STAR team at a workshop to discuss interventions for the prevention of chronic pain after knee replacement. Those attending included surgeons, anaesthetists, physiotherapists, nurses and former patients, as well as researchers with interests in randomised trials, health economics, qualitative studies, PPI, systematic reviews and cohort studies.

The systematic review work package lead explained that the aim of the systematic reviews was to identify interventions that may improve long-term pain outcomes after knee replacement. The following definition of an intervention was presented:

Any clinical treatment or public health measure designed to reduce the incidence or modify the effects of particular diseases.

Yarnell and O’Reilly 104

Workshop participants considered that aspects of pre-operative care were potential targets for an intervention to prevent long-term pain. Many patients receive pre-operative education at a single class and from information booklets. Changes to content and interventions to improve uptake of classes were seen as potentially valuable. Pre-operative interventions might include cognitive behavioural therapy, self-management and peer support, weight management, exercise, treatment of comorbidities, nutritional guidance, and intra-articular injections. A multimodal approach might include education, psychological support, management of comorbidities and other components.

In the peri-operative period, workshop participants considered effective pain management important, possibly using gabapentin, nerve blocks and multimodal approaches. Certain anticoagulants are associated with micro-bleeds, which may lead to long-term pain. The use of tourniquets may also be associated with long-term pain.

An enhanced recovery protocol was considered a possible intervention to limit long-term pain. After hospital discharge, workshop participants noted the potential value of mid-term rehabilitation, peer support groups, provision of contact points, an introduction to community services and online resources, and psychological support including cognitive behavioural and mindfulness therapies. Provision of physiotherapy in different formats was also recognised as worth evaluating. Other interventions suggested were intra-articular injections, weight management, podiatry and realignment.

Methods

The systematic reviews were registered prospectively as PROSPERO CRD42017041382. 105 We established a database of all randomised controlled trials and systematic reviews in total knee replacement. These were identified through searches of The Cochrane Library, MEDLINE, Embase, CINAHL (Cumulative Index to Nursing and Allied Health Literature) and PsycINFO^® (American Psychological Association, Washington, DC, USA) in November 2016 (updated February 2018 for peri-operative interventions and December 2018 for pre-operative interventions) (see Appendix 3, Systematic review search strategy as applied in MEDLINE on Ovid).

Eligible studies were randomised controlled trials that met the following inclusion criteria:

1. patients with osteoarthritis awaiting or who have received total knee replacement

2. intervention – treatment in the pre-, peri- or post-operative setting

3. control – usual care or alternative treatment

4. outcomes – chronic pain at ≥ 6 months after knee replacement (≥ 12 months for post-operative interventions), adverse events.

The database was screened, and interventions divided into pre-, peri- and post-operative contexts, and into intervention groups. Risk of bias was assessed using the Cochrane tool. 106

Results: pre-operative interventions

Eight randomised controlled trials with nine comparisons (960 patients) were eligible (see Appendix 3, Figure 3). There was moderate-quality evidence of no effect of exercise programmes on chronic pain after total knee replacement, based on a meta-analysis of six interventions with 229 participants (standardised mean difference 0.20, 95% CI –0.06 to 0.47; I² = 0%). A sensitivity analysis restricted to studies at low risk of bias confirmed these findings. Studies evaluating a combined exercise and education intervention (one study) and education alone (one study) suggested similar findings.

FIGURE 3.

Effectiveness of pre-operative interventions: systematic review flow diagram. RCT, randomised controlled trial.

Results: peri-operative interventions

Forty-four randomised controlled trials at low risk of bias evaluated interventions in the peri-operative setting with a pain outcome or score with a pain component at ≥ 6 months’ follow-up (see Appendix 3, Figure 4). Intervention heterogeneity precluded meta-analysis. There was weak evidence for small reductions in chronic pain after total knee replacement in people who received peri-operative local infiltration analgesia (three studies), ketamine infusion (one study) or pregabalin (one study). Supported early discharge (one study) showed weak evidence of a small reduction in chronic pain. More clinically important benefits were seen for electric muscle stimulation (two studies), gait training (one study) and a course of anabolic steroids (one small pilot study).

FIGURE 4.

Effectiveness of peri-operative interventions: systematic review flow diagram.

For a range of peri-operative treatments there was no evidence linking them with unfavourable pain outcomes. For example, blood conservation with tranexamic acid during knee replacement was not associated with chronic pain. However, otherwise extensively researched interventions including venous thromboembolism prevention and tourniquet use have not been evaluated in relation to chronic pain.

Results: post-operative interventions

Randomised controlled trials of post-discharge interventions commencing in the first 3 months after total knee replacement were included (see Appendix 3, Figure 5). Seventeen trials with 2485 participants were included. All studies were at risk of bias because participants were not blinded to arm allocation, and five because of incomplete outcome data. Twelve trials evaluated physiotherapy interventions. Other interventions were nurse-led telephone follow-up, neuromuscular electrical stimulation and a multidisciplinary intervention. One study showed benefit for home-based exercises aimed at managing kinesophobia in reducing pain severity compared with no intervention. Otherwise, narrative synthesis found no evidence that one type of physiotherapy intervention is more effective than another. A 10-day multidisciplinary outpatient rehabilitation programme provided between 2 and 4 months after surgery showed no long-term benefit for pain or function. 107 For other interventions, there was insufficient evidence to draw conclusions about effectiveness.

FIGURE 5.

Effectiveness of post-operative interventions: systematic review flow diagram. RCT, randomised controlled trial; TKR, total knee replacement.

Effectiveness of interventions to manage chronic pain after surgery: systematic review

This review has been published as Wylde et al. 2017. 108

Aims

This systematic review aimed to assess the efficacy of interventions to treat chronic pain after non-cancer surgeries.

Methods

The systematic review was registered prospectively as PROSPERO CRD42015015957. 109 We searched MEDLINE, Embase, PsycINFO, CINAHL and The Cochrane Library from inception to March 2016 (see Appendix 4, Systematic review search strategy as applied in MEDLINE on Ovid). An update in October 2020 was timed to contextualise the STAR intervention.

Eligible studies were randomised controlled trials that met the following inclusion criteria:

1. patients with chronic pain after non-cancer surgery

2. an intervention for pain received by patients at a minimum of 3 months after surgery

3. a comparator of no treatment, placebo, usual care or alternative treatment

4. an outcome relating to pain.

Risk of bias was assessed using the Cochrane tool. 106

Results

As shown in Appendix 4, Figure 6, 66 trials with data from 3149 participants were included. Most trials included patients with chronic pain after spinal surgery (n = 25) or amputation (n = 21). Interventions were antiepileptics, capsaicin, epidural steroid injections, local anaesthetic, neurotoxins, opioids, acupuncture, exercise, spinal cord stimulation, further surgery, laser therapy, magnetic stimulation, mindfulness-based stress reduction, mirror therapy and sensory discrimination training. Opportunities for meta-analysis were limited by heterogeneity. For all interventions, there was insufficient evidence to draw conclusions on effectiveness but the review provided clear suggestions for future research.

FIGURE 6.

Effectiveness of interventions to manage chronic pain after surgery: systematic review flow diagram. RCT, randomised controlled trial.

Review update

To examine the STAR trial in a contemporary context, we updated searches with terms relating to arthroplasty of the large joints, post-surgical pain, and randomised controlled trials (see Appendix 5, Systematic review search strategy as applied in MEDLINE on Ovid). Systematic reviews and trial registries were checked for studies.

Eligible studies were randomised controlled trials that met the following inclusion criteria:

1. patients with chronic pain after arthroplasty of the large joints

2. intervention – treatment for chronic pain

3. control – no treatment, usual care or alternative treatment

4. outcome – pain.

Of 3901 articles identified and screened by one reviewer, 69 were potentially relevant (see Appendix 5, Figure 7). After detailed evaluation by two reviewers, four published randomised evaluations of treatments for chronic pain after knee replacement were identified in the original review and update (see Appendix 5, Table 1). No study was judged to be at high risk of bias. For people with general chronic pain, there were encouraging findings warranting further research into intra-articular botulinum toxin110 and denervation therapy. 111 Radiofrequency genicular nerve treatment showed similar outcomes to treatment with anaesthetic and corticosteroid. 112 No studies were found that evaluated a multifaceted intervention, but one study focusing specifically on treatment of neuropathic pain with topical lidocaine suggested that further research is merited for this personalised treatment. 113

TABLE 1 - Details of included studies

Study, country, date of recruitment	Inclusion criteria	Number randomised (intervention : control)	Intervention	Comparator	Risk of bias issues	Key results
Published studies
Singh et al. 2010,110 USA, 2006–9	Pain after total knee replacement for > 3 months, NRS pain intensity ≥ 6/10	n = 54; 60 knees (30 : 30)	Single intra-articular botulinum toxin A injection	Single intra-articular injection of saline	Low risk of bias	Reduced pain intensity in botulinum A group after 3 months. Pain relief to around 40 days
Ma et al. 2016,111 China, 2014–15	Intractable pain of knee joint after total knee replacement	n = 100 (50 : 50)	Denervation therapy	Drug treatment	No losses to follow-up	Denervation therapy associated with improved symptoms
Pickering et al. 2019,113 France, 2016	Localised neuropathic pain after knee surgery	n = 36 (24 : 12)	5% lidocaine-medicated plaster for 3 months	5% plaster with no drug for 3 months	No losses to follow-up	Lidocaine plaster reduced localised neuropathic pain
Qudsi-Sinclair et al. 2017,112 Spain, 2012–14	Pain after total knee replacement	n = 33 (15 : 18; 14 : 14 received treatment)	Single radiofrequency genicular nerve block	Single analgesic block with corticosteroid	Some concerns: uneven follow-up in small study	Similar pain outcomes in both groups
Wylde et al. 2022120	Chronic pain 2 months after primary total knee replacement	n = 363 (242 : 121)	STAR intervention	Usual care	Low risk of bias	STAR was clinically effective and cost-effective in improving pain outcomes at 1 year
From trial registries and published protocols
NCT02211534119	Persistent post-operative pain following total knee replacement	NA	Pulsed electromagnetic energy field therapy	Sham pulsed electromagnetic field	NA	NA
NCT02931435117	Chronic knee pain despite total knee replacement at ≥ 6 months	NA	Nerve block with radiofrequency ablation	Sham radiofrequency ablation	NA	NA
NCT03825965115	Persistent post-surgical pain following total knee replacement	NA	Cannabinoids	Placebo	NA	NA
NCT04100707118	Knee pain 3 months after total knee replacement	NA	Genicular nerve blocks	Sham comparator	NA	NA
NCT03973177116 Crossover	Refractory chronic knee pain for more than 6 months after total knee replacement	NA	Phenol injection: neurolysis of genicular nerves	Methylprednisolone injection	NA	NA
Larsen et al. 2020114	Chronic pain after primary total knee replacement	NA	Neuromuscular exercise and pain neuroscience education	Pain neuroscience education	NA	NA
NA, not applicable; NRS, numeric rating scale.

FIGURE 7.

Interventions to manage chronic post-surgical pain update: systematic review flow diagram. RCT, randomised controlled trial.

In addition to the STAR trial, six studies were ongoing or not published (see Appendix 5, Table 1). These focus on exercise and education,114 cannabinoids,115 phenol neurolysis of genicular nerves,116 genicular nerve blocks117,118 and pulsed electromagnetic field therapy,119 all in people with general chronic pain after knee replacement.

Background

Although chronic pain places considerable burden on health-care systems and individuals, little research has characterised chronic pain after total knee replacement and its impact on health-care resource use.

Aims

We aimed to characterise the natural history and course of chronic pain after total knee replacement, including health-care resource use, through additional follow-up of the Clinical Outcomes in Arthroplasty Study (COASt) cohort and analysis of the linked CPRD and HES databases.

Background

As part of our Programme Development Grant, we conducted preliminary studies to design an intervention to improve the management of chronic pain after total knee replacement. Studies included a systematic review, a survey of NHS practice, qualitative work with health-care professionals, an expert group meeting and PPI activities. 48,124–126 We found a lack of clear pathways and referral processes for patients with chronic pain after total knee replacement,124,125 hindering patients’ access to targeted and individualised care and highlighting the need for improved access to appropriate pain management interventions for this population.

Aim

The aim of this work package was to refine and finalise a new care pathway for patients with chronic pain after total knee replacement. Specific objectives comprised the following:

1. the refinement of intervention content

2. the testing of intervention delivery and acceptability

3. the evaluation of views about intervention implementation within a trial.

Key findings

Consensus questionnaires and meetings with health-care professionals ensured that the intervention components were appropriate and informed substantive changes. Testing of intervention delivery identified that the intervention was acceptable to patients after small changes to assessment clinic processes. Engagement with stakeholders indicated that the intervention could be successfully implemented for evaluation within a trial setting. Based on this work, our novel intervention was refined and prepared for evaluation. Findings also informed the design and development of a comprehensive intervention training manual and training event.

An overview of the intervention is shown in Figure 8. The intervention involves patients reporting moderate-to-severe pain at 3 months after total knee replacement attending an assessment clinic with a trained Extended Scope Practitioner (ESP). The clinic uses a standardised procedure involving (1) clinical history; (2) a review of patient-reported outcome measures including measures of pain, depression and neuropathic pain; (3) a knee examination to evaluate knee tenderness, surgical wound healing, range of motion, alignment, stability, patellofemoral joint function, infection and signs of CRPS; (4) an evaluation of radiographs for evidence of fracture or concerns with alignment, fixation, sizing or implant position; and (5) a blood test for markers of infection. Based on the findings of the assessment clinic, patients are referred to the appropriate existing services for further treatment. This may include one or more of the following: an orthopaedic surgeon for surgical review, a physiotherapist for muscle strengthening and exercise, a GP for treatment of depression or anxiety, and pain specialists for treatment of neuropathic pain or CRPS (via GPs). The care pathway is individualised and flexible, and the number of referrals reflects the needs of the patient. Monitoring is also available if appropriate. As part of the intervention, patients receive telephone follow-up from the ESP up to six times over 12 months to ensure any referrals are being undertaken and discuss any further referrals based on clinical need. The intervention aims to enable appropriate onwards referral to existing services to ensure that underlying reasons for chronic pain are considered and that treatment is targeted at these to improve pain management and to reduce the impact of pain.

FIGURE 8.

The STAR care pathway. PFJ, patellofemoral joint. CRPS, chronic regional pain syndrome.

Background

After developing the STAR care pathway, evidence was required on its effectiveness, cost-effectiveness and acceptability.

Aim

This work package aimed to evaluate the clinical effectiveness and cost-effectiveness of the STAR care pathway compared with usual care for chronic pain after total knee replacement.

Background

As with other chronic pain problems,75 people with chronic pain after total knee replacement may make limited or no use of formal services, and may believe that pain is inevitable and that little can be done. 78,149

Background

Implementation science provides models to help ensure maximal uptake of new interventions and processes by health-care professionals and organisations. 152–154 NPT is an approach from implementation science that we used to assist optimal implementation of the care pathway in the trial and to understand how the pathway could be best put into practice if clinically effective and cost-effective. We also conducted dissemination activities, including conference presentations, journal articles, and communication for professionals, patients and the public.

Theme 1: how and when people with pain after total knee replacement should be identified

Our work recommends early post-operative screening, beginning at 2 months after surgery, to identify patients with troublesome pain after knee replacement. This can then facilitate targeted care delivery at 3–4 months post surgery when acute pain is transitioning to chronic pain, with the aim of preventing longer-term chronicity. To identify patients with troublesome pain after total knee replacement who would likely benefit from intervention, we have, to the best of our knowledge, developed and applied the first robust and standardised approach using a derived cut-off score on the OKS pain subscale.

Theme 2: predicting who will develop chronic pain after total knee replacement

Our database analyses confirmed the importance of pre-operative pain as a predictor of chronic pain after knee replacement. Those with the mildest pre-operative knee pain were more likely to move to a worse post-operative pain state. Patients taking opioids and antidepressant medications were more likely to have worse pain outcomes. Other risk factors, including smoking, obesity and comorbidities are potentially modifiable before surgery or may identify groups requiring additional care and monitoring.

Theme 3: prevention of chronic pain after total knee replacement

Randomised trials to assess long-term outcomes after pre-, peri- and post-operative interventions are feasible and necessary to ensure that patients receive care with reduced or no risk of chronic pain. Some unifactorial interventions merit further study and may have a role in the knee replacement pathway.

Theme 4: trajectory of chronic pain after total knee replacement

Although two-thirds of people with chronic pain 1 year after surgery appear to improve and no longer experience pain as the years pass, our estimates (after imputing significant missing data assumed to be missing at random) indicate that one in three still experience chronic pain 5 years after their operation or fluctuate in and out of pain. When describing chronic post-surgical pain, some people describe sensations of discomfort including heaviness, numbness, pressure and tightness associated with the prosthesis. Some people report a lack of felt connection with their knee and a lack of confidence in it.

Theme 5: how people manage chronic pain after total knee replacement

Chronic pain after total knee replacement is linked with slightly more visits to the GP, more physiotherapy and a greater consumption of stronger analgesia even years after the operation. People with chronic pain after knee replacement who make little or no use of services often feel nothing more can be done, or that further treatments may not be of benefit or cause further harm.

Theme 6: economic impact of chronic pain after total knee replacement

We found that chronic pain after total knee replacement is associated with an additional average expenditure of about £100 per patient in primary care consultations during the first year after surgery. This extra cost is sustained over time, for at least up to 8 years after the operation. Analgesia prescription is also more costly for patients with chronic pain and the associated extra cost increases after surgery primarily owing to the increased prescription of opioids. During the 8 years following knee replacement, pain analgesia prescription costs the NHS on average an additional £35 per patient per year for those in chronic pain. Additional costs associated with chronic pain, including primary and hospital care as well as private physiotherapy (but excluding prescription medication), can be as high as £1300 in the first year after surgery. This decreases over time, reaching £55 by the fourth year. Furthermore, chronic pain can be expected to be associated with increased formal and informal care as well as productivity losses, which would significantly increase its socioeconomic impact.

Theme 7: how to optimise the management of chronic pain after total knee replacement

Our trial showed that the STAR care pathway is a clinically effective and cost-effective intervention for reducing pain severity and pain interference over 1 year for patients with troublesome pain at 3 months after knee replacement. The novel STAR care pathway was designed to provide personalised care through a multifaceted treatment approach. The intervention addresses the need for clear pathways and referral processes to facilitate access to targeted care matched to individual patients’ pain characteristics.

The STAR care pathway provides patients with reassurance and confidence in their recovery and ensures that they receive the treatments they require. Patients valued the opportunities to discuss their concerns with a health-care professional and derived a sense of reassurance and encouragement from the STAR clinic and follow-up calls.

To the best of our knowledge, our trial is the first robust evaluation of a multifaceted and personalised intervention for patients with pain at three months after knee replacement.

For people with general chronic pain after knee replacement, there were encouraging findings from our systematic reviews that warrant further research into unifactorial interventions. However, there is currently insufficient evidence to support their use in clinical practice. This was also apparent in the more general literature reporting evaluations of interventions for chronic pain after diverse surgeries. The design of the STAR pathway will permit the incorporation of new evidence-based interventions into patient care.

Challenges in the programme as a whole

A strength of the STAR programme was the inclusion of work packages that delivered findings to build a broader picture of chronic pain after total knee replacement. The programme comprised inter-related studies and it was important that these were connected to one another.

The main challenges to the programme related to the implementation of strategies to improve trial recruitment and ensure the success of the trial. Differences between anticipated and actual recruitment related to the availability of patients at trial sites, including during periods when fewer patients had surgery such as elective surgery closures during winter bed pressures. Feasibility assessments and recruitment projections accounting for patient availability were developed to ensure realistic recruitment targets at each site. These methods of optimising recruitment were successful in producing a powered sample for analysis.

An opportunity presented by the programme was the multidisciplinary nature of the team and the research delivered. Over 5 years, the programme team worked together closely and shared progress, details of research methodologies and their clinical experience. As such, the programme served to develop UK health research capacity in knee replacement and pain research. The learning from one another’s approaches is intangible and difficult to define but is likely to deliver long-term benefit to team members’ individual capacity, future research leadership and potential to support high-impact research in the future.

Recommendations for future research

Based on the programme, we recommend the following areas for future research. These recommendations focus on moving from work on risk factors to future interventions, improvements to the STAR care pathway and enabling fuller engagement between people living with or at risk of pain after knee replacement and health care:

1. Research that identifies new interventions with evidence of effectiveness may need to be considered in future refinements of the STAR intervention to ensure that it remains current, evidence-based and able to deliver a personalised approach to the management of chronic pain after knee replacement.

2. Research to pre-operatively identify those patients who are most likely to have poor outcomes and to develop a tool for the prediction of poor outcomes after surgery.

3. Development of evidence-based information for patients to inform, empower and manage expectations – including in the pre-operative phase of care – and to enable people with chronic pain after knee replacement to seek health care, to provide information about forms of management that are available in the NHS and to encourage professionals to provide access and encouragement.

4. Research to explore and develop an intervention to identify how to support people to manage any feelings of disconnectedness from their replaced knee.

5. Longer-term follow-up of patients who received the STAR care pathway to describe longer-term outcomes.

6. Given the complexity of pain that extends or emerges after surgery, individualised multimodal interventions matched to pain characteristics after other surgeries – as evaluated in the STAR trial – merit development and evaluation.

7. The STAR programme focused on care after surgery. Future research could focus on the time before surgery as an opportunity for intervention. We have a greater understanding of risk factors for poor outcome and using this understanding to design and evaluate pre-surgical interventions may prove of long-term benefit to patients and health-care systems. For example, research could develop and evaluate an intervention to address opioid medication prescribing. There may be need for research into clearer expectation setting and communication between patients and professionals before surgery.

Reflections on work packages

Working with our PPI group has been valuable across all work packages. The group gave support to the STAR randomised controlled trial that helped with the inclusion, recruitment, retention and engagement of patients.

Trial recruitment originally aimed to recruit equal numbers of participants from four sites with a variety of patient characteristics. In an ‘internal pilot phase’, we increased the number of study sites from four to eight to ensure that we met the recruitment target.

Without a clear mechanism linking pre-operative and long-term post-operative pain, it proved hard to justify to peer reviewers that systematic reviews of trials of a broad range of pre-, peri- and post-operative interventions should include a focus on long-term pain as an outcome. However, pain should be included with other adverse events in long-term follow-up when changes are made to patient care in the knee replacement patient pathway. Direct effects may be clear through nerve damage and other biological mechanisms. However, associations with long-term pain may be indirect, mediated through other adverse events and, as we have shown, extended hospital stay and readmissions.

Challenges and successes

Successes for the programme included high-quality and robust results relating to patient care and outcomes after knee replacement. Although the research focused on the time after surgery, the programme provides findings are programmatic in nature and relate to the whole patient journey through knee replacement. These results are directly relevant to patients, health-care professionals and providers. We now have new important knowledge on impact, risk factors and management that will inform care.

Challenges included the need to improve recruitment and retention methods in the trial. Extending trial recruitment to additional sites proved a successful strategy. In addition, collaboration with the PPI group assisted in the improvement of recruitment and retention methods. This collaboration and the input provided by the PPI group was crucial to the success of the trial and the programme.

The COVID-19 pandemic fell during the follow-up phase of the main trial. Although this did not affect our follow-up response rate, concerns were raised over how responses may be affected by the national lockdown and limited availability of health-care services. Sensitivity analyses were implemented to assess this and did not find any significant effects.

Service developments

During the programme we developed, delivered and evaluated a new care pathway for people with chronic pain after knee replacement. This represents a service development that is acceptable, clinically effective and cost-effective and ready for implementation in health care.

Implications for practice and lessons learned

This programme was designed to inform and evaluate improvements to services and patient well-being. We discussed these with stakeholders towards the end of the programme, who expressed a desire to see the delivery of the STAR care pathway in the NHS.

The following key implications from the work relate to a pressing need to provide support and care for people experiencing chronic pain after knee replacement:

1. Access to an evidence-based care pathway, such as the STAR care pathway, can improve outcomes and is cost-effective. Delivery of the pathway should provide benefit.

2. People with chronic pain after knee replacement may benefit from clearer information about the likelihood that their pain may change over time, including that it might improve, and that engagement with health-care professionals provides an opportunity to address concerns about their replacement.

3. For people who do not make use of health care but who have chronic pain after knee replacement, providing them with information about the potential benefit of health care and encouraging professionals to enable them to seek care may have benefit by enabling these people to access evidence-based forms of management for their pain.

Taken together, the programme has provided evidence about what provides benefit and what is not known, and has identified directions of future research need. The programme has generated a large body of evidence about pain after total knee replacement and serves to raise awareness of the problem, pushing the issue higher up the agenda of research, practice and policy. The impact of the programme as a whole is beyond the sum of its parts: we have generated evidence and awareness of this important issue, and there is more work to be done.

Programme and Trial Steering Committees

We thank the members of the Programme and Trial Steering Committees for their support in the development, conduct and delivery of the programme as a whole:

1. Joy Adamson, chair, Programme Steering Committee; member, Trial Steering Committee

2. Paul Ewings, member, Programme Steering Committee; chair, Trial Steering Committee

3. Lizzy Betts, member, Programme Steering Committee; member, Trial Steering Committee

4. George Peat, member, Programme Steering Committee; member, Trial Steering Committee

5. Mark Rockett, member, Programme Steering Committee; member, Trial Steering Committee.

Researchers

Sophie Cole, Antonella Delmestri, Kirsty Garfield, Sara Khalid, Spyros Kolovos, Fiona MacKichan, Maria T Sanchez-Santos and Anushka Soni.

STAR site personnel

Shannon Biggs, Tim Board, Emma Bruce, Philip Buckley, Ben Burston, Kate Button, Joanna Clarke, Jean Denton, Vikram Desai, Andrew Emms, Colin Esler, Anne Evans, Jessica Falatoori, Jason Fletcher, Giles Head, Emma Hoskins, Natalie Jackson, Heather Jarvis, Phillips Jon, Debbie Jones, Dileep Kumar, Tracey Lee, Leigh Morrison, Gemma Munkenbeck, Adam Nelson, Claire Nicholas, Valerie Parkinson, Michael Parry, Charlotte Perkins, Kathryn Robinson, Paul Stanton, Rowenna Stroud, Joanne Taylor, John-Paul Whittaker, Matthew Williams, Deborah Wilson and Lucy Young.

Collaborators

Jo Adams, Aideen Ahern, Nicholas Ambler, Kate Button, Paul Dieppe, Peter Gladwell, Beverley Hopwood, Athene Lane, Helen Lewis-White, James Murray, Donna Noonan and Nick Smith.

Support

Elizabeth Bradshaw, David Carmichael, Beverley Evanson and Christine Hobson.

Contributions of authors

Rachael Gooberman-Hill (https://orcid.org/0000-0003-3353-2882) (Professor of Health and Anthropology) was chief investigator with responsibility for delivery of the overall programme and is the corresponding author.

Vikki Wylde (https://orcid.org/0000-0002-8460-1529) (Associate Professor in Musculoskeletal Health) was a co-applicant and led and conducted work in the systematic reviews, intervention design and randomised trial.

Wendy Bertram (https://orcid.org/0000-0001-8234-2052) (Senior Research Associate, Clinical Trials Manager) was involved in the intervention design, co-ordinated the pilot trial and was trial manager for the randomised trial.

Andrew J Moore (https://orcid.org/0000-0003-3185-1599) (Research Fellow in Qualitative Health Service Research) led the qualitative research throughout the programme and led the design and delivery of the qualitative implementation evaluation.

Rafael Pinedo-Villanueva (https://orcid.org/0000-0002-4723-5128) (Associate Professor in Health Economics) led the health economics analysis based on observational data.

Emily J Sanderson (https://orcid.org/0000-0003-2268-4194) (Research Associate in Medical Statistics) was the trial statistician and planned and conducted the statistical analysis in the trial.

Jane Dennis (https://orcid.org/0000-0001-9718-2653) (Senior Research Associate in Evidence Synthesis) undertook the systematic reviews and meta-analysis.

Shaun Harris (https://orcid.org/0000-0003-0698-1543) (Senior Research Associate in Health Economics) undertook the STAR randomised trial health economic analyses.

Andrew Judge (https://orcid.org/0000-0003-3015-0432) (Professor of Translational Statistics) was a co-applicant and led the analyses of routine data from the NJR and Clinical Practice Research databases, and from the COASt study.

Sian Noble (https://orcid.org/0000-0002-8011-0722) (Senior Lecturer in Health Economics) was a co-applicant and led the health economics analysis in the STAR randomised trial.

Andrew D Beswick (https://orcid.org/0000-0002-7032-7514) (Research Fellow, Systematic reviews) was a co-applicant, led the systematic reviews and co-ordinated the writing of the final report.

Amanda Burston (https://orcid.org/0000-0003-3480-4210) (Senior Research Associate, Public and Patient Involvement) was a co-applicant and led and coordinated PPI in the programme.

Tim J Peters (https://orcid.org/0000-0003-2881-4180) (Professor of Primary Care Health Services Research) was a co-applicant and led the statistical design and analysis of the STAR randomised trial.

Julie Bruce (https://orcid.org/0000-0002-8462-7999) (Professor of Clinical Trials) was a co-applicant and advised on the conduct and reporting of all aspects of the research.

Christopher Eccleston (https://orcid.org/0000-0003-0698-1543) (Professor of Pain Science) was a co-applicant and advised on the conduct and reporting of all aspects of the research.

Stewart Long (Versus Arthritis Director of Involvement, Influencing and Support) was a co-applicant and advised on the conduct and reporting of all aspects of the research.

David Walsh (https://orcid.org/0000-0002-6316-5692) (Professor of Rheumatology) advised on the conduct and reporting of all aspects of the research.

Nicholas Howells (https://orcid.org/0000-0002-8514-0322) (Consultant Orthopaedic Surgeon) was a named collaborator and principal investigator for the trial site at North Bristol NHS Trust.

Simon White (https://orcid.org/0000-0002-6264-2020) (Consultant Orthopaedic Surgeon) was a named collaborator and principal investigator for the trial site at Cardiff and Vale University Health Board.

Andrew Price (https://orcid.org/0000-0002-4258-5866) (Professor of Orthopaedic Surgery) was a co-applicant and provided specialist surgical input into the programme.

Nigel Arden (https://orcid.org/0000-0002-3452-3382) (Professor in Rheumatic Diseases) was a co-applicant and provided specialist clinical input into the programme.

Andrew Toms (https://orcid.org/0000-0002-8035-0494) (Professor of Orthopaedic Surgery) was a named collaborator and principal investigator for the trial site at Royal Devon and Exeter NHS Foundation Trust.

Candida McCabe (https://orcid.org/0000-0002-3593-0027) (Florence Nightingale Foundation Clinical Professor in Nursing) was a co-applicant and provided specialist clinical input into the programme.

Ashley W Blom (https://orcid.org/0000-0002-9940-1095) (Professor of Orthopaedic Surgery) was a co-applicant and provided specialist surgical expertise into the programme.

Publications

Conference posters and presentations

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.

Systematic review search strategy as applied in MEDLINE on Ovid

1. Epidemiologic Studies/

2. exp Case-Control Studies/

3. exp Cohort Studies/

4. Cross-Sectional Studies/

5. (epidemiologic adj (study or studies)).ab,ti.

6. case control.ab,ti.

7. (cohort adj (study or studies)).ab,ti.

8. cross sectional.ab,ti.

9. cohort analy$.ab,ti.

10. (follow up adj (study or studies)).ab,ti.

11. longitudinal.ab,ti.

12. retrospective$.ab,ti.

13. prospective$.ab,ti.

14. (observ$ adj3 (study or studies)).ab,ti.

15. exp clinical study/

16. randomized controlled trial/

17. 15 not 16

18. adverse effect?.ab,ti.

19. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14 or 17 or 18

20. Arthroplasty, Replacement, Knee/

21. Knee Prosthesis/

22. (arthoplast$ adj3 knee$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier]

23. (knee$ adj3 replac$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier]

24. (knee adj3 implant$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier]

25. 20 or 21 or 22 or 23 or 24

26. 19 and 25

Systematic review search strategy as applied in MEDLINE on Ovid

1. randomized controlled trial/or randomized controlled trial.pt.

2. controlled clinical trial.pt.

3. randomized.ab.

4. placebo.ab.

5. randomly.ab

6. trial.ab

7. randomised.tw

8. 1 or 2 or 3 or 4 or 5 or 6 or 7

9. review/

10. ‘systematic review$’.mp

11. 9 or 10

12. 8 or 11

13. Arthroplasty, Replacement, Knee/

14. Knee Prosthesis/

15. (arthoplast$ adj3 knee$).mp. [mp = title, abstract, original title, name of substance word,

16. subject heading word, keyword heading word, protocol supplementary concept word, rare disease

17. supplementary concept word, unique identifier]

18. (knee$ adj3 replac$).mp. [mp = title, abstract, original title, name of substance word, subject

19. heading word, keyword heading word, protocol supplementary concept word, rare disease

20. supplementary concept word, unique identifier]

21. (knee adj3 implant$).mp. [mp = title, abstract, original title, name of substance word, subject

22. heading word, keyword heading word, protocol supplementary concept word, rare disease

23. supplementary concept word, unique identifier]

24. 13 or 14 or 15 or 16 or 17

25. 12 and 18

Systematic review search strategy as applied in MEDLINE on Ovid

1. randomized controlled trial/or randomized controlled trial.pt.

2. controlled clinical trial.pt.

3. randomized.ab.

4. placebo.ab.

5. randomly.ab.

6. trial.ab.

7. randomised.tw.

8. 1 or 2 or 3 or 4 or 5 or 6 or 7

9. Pain, Postoperative/

10. ((postoperative adj6 pain*) or (post-operative adj6 pain*) or postoperative- pain*).mp.

11. ((post-operative adj6 analg*) or (postoperative adj6 analg*)).mp.

12. ((post-surgical adj6 pain*) or (post surgical adj6 pain*) or (postsurgery adj6 pain*) or (post adj surg* adj pain*)).mp.

13. ((post* adj pain*) or pain relief after or pain following surg*).mp.

14. ((posttreatment adj6 pain*) or (pain control after adj6 surg*) or ((post-extraction or postextraction or post-surg*) and (pain* or discomfort))).mp.

15. ((analg* adj6 postoperat*) or (analg* adj6 post-operat*) or (pain* adj6 after surg*) or (pain* adj6 after operat*) or (analgesi* adj6 after operat*)).mp.

16. ((pain* or analg*) adj6 (“follow* operat*” or “follow* surg*”)).mp.

17. 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16

18. exp Neuralgia/

19. (neuralgia* or neurodynia).tw.

20. ((neuropathic or nerve*) adj3 pain*).tw.

21. Amputation/

22. Phantom Limb/

23. Failed Back Surgery Syndrome/

24. 18 or 19 or 20 or 21 or 22 or 23

25. (chronic* or constant* or continu* or persist* or longterm or long-term or longstanding or long-standing or long lasting or longlasting or phantom).mp.

26. exp Pain, Intractable/or exp Chronic Pain/

27. exp pain/and (chronic* adj5 pain*).mp.

28. 25 or 26 or 27

29. 17 or 24

30. 24 or 28

31. 8 and 29 and 30

Systematic review search strategy as applied in MEDLINE on Ovid

1. controlled clinical trial.pt.

2. randomized controlled trial.pt.

3. clinical trials as topic/

4. (randomi#ed or randomi#ation or randomi#ing).ti,ab,kf.

5. (RCT or “at random” or (random* adj3 (administ* or allocat* or assign* or class* or cluster or crossover or cross-over or control* or determine* or divide* or division or distribut* or expose* or fashion or number* or place* or pragmatic or quasi or recruit* or split or subsitut* or treat*))).ti,ab,kf.

6. placebo.ab,ti,kf.

7. trial.ti.

8. (control* adj3 group*).ab.

9. (control* and (trial or study or group*) and (waitlist* or wait* list* or ((treatment or care) adj2 usual))).ti,ab,kf.

10. ((single or double or triple or treble) adj2 (blind* or mask* or dummy)).ti,ab,kf.

11. double-blind method/or random allocation/or single-blind method/

12. or/1-11

13. (systematic or structured or evidence or trials or studies).ti. and ((review or overview or look or examination or update* or summary).ti. or review.pt.)

14. (0266-4623 or 1469-493X or 1366-5278 or 1530-440X or 2046-4053).is.

15. meta-analysis.pt. or (meta-analys* or meta analys* or metaanalys* or meta synth* or meta-synth* or metasynth*).ti,ab,kf,hw.

16. ((systematic or meta) adj2 (analys* or review)).ti,kf. or ((systematic* or quantitativ* or methodologic*) adj5 (review* or overview*)).ti,ab,kf,sh. or (quantitativ$ adj5 synthesis$).ti,ab,kf,hw.

17. (integrative research review* or research integration).tw. or scoping review?.ti,kf. or (review.ti,kf,pt. and (trials as topic or studies as topic).hw.) or (evidence adj3 review*).ti,ab,kf.

18. review.pt. and ((medline or medlars or embase or pubmed or scisearch or psychinfo or psycinfo or psychlit or psyclit or cinahl or electronic database* or bibliographic database* or computeri#ed database* or online database* or pooling or pooled or mantel haenszel or peto or dersimonian or der simonian or fixed effect or ((hand adj2 search*) or (manual* adj2 search*))).tw,hw. or (retraction of publication or retracted publication).pt.)

19. or/13-18

20. Arthroplasty, Replacement, Knee/

21. Knee Prosthesis/

22. (arthoplast$ adj3 knee$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

23. (knee$ adj3 replac$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

24. (knee$ adj3 implant$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

25. (knee$ adj3 prosthe$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

26. (knee$ adj3 endoprosthe$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

27. Arthroplasty, Replacement, Hip/

28. Hip Prosthesis/

29. (arthoplast$ adj3 hip$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

30. (hip$ adj3 replac$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

31. (hip$ adj3 implant$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

32. (hip$ adj3 prosthe$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

33. (hip$ adj3 endoprosthe$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

34. Arthroplasty, Replacement, Shoulder/

35. Shoulder Prosthesis/

36. (arthoplast$ adj3 shoulder$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

37. (shoulder$ adj3 replac$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

38. (shoulder$ adj3 implant$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

39. (shoulder$ adj3 prosthe$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

40. (shoulder$ adj3 endoprosthe$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

41. Arthroplasty, Replacement, elbow/

42. Elbow Prosthesis/

43. (arthoplast$ adj3 elbow$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

44. (elbow$ adj3 replac$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

45. (elbow$ adj3 implant$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

46. (elbow$ adj3 prosthe$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

47. (elbow$ adj3 endoprosthe$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

48. Arthroplasty, Replacement/

49. Wrist/or Wrist Joint/

50. 48 and 49

51. (arthoplast$ adj3 wrist$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

52. (wrist$ adj3 replac$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

53. (wrist$ adj3 implant$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

54. (wrist$ adj3 prosthe$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

55. (wrist$ adj3 endoprosthe$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

56. Arthroplasty, Replacement, Ankle/

57. (arthoplast$ adj3 ankle$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

58. (ankle$ adj3 replac$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

59. (ankle$ adj3 implant$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

60. (ankle$ adj3 prosthe$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

61. (ankle$ adj3 endoprosthe$).mp. [mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

62. or/20-47

63. or/50-61

64. Pain, Postoperative/

65. chronic pain/

66. Pain, intractable/

67. (chronic* or constant* or continu* or persist* or longterm or long-term or longstanding or long-standing or long lasting or longlasting).mp. adj3 pain/[mp = title, abstract, original title, name of substance word, subject heading word, floating sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms]

68. ((postoperative adj6 pain*) or (post-operative adj6 pain*) or postoperative- pain*).mp.

69. ((post-operative adj6 analg*) or (postoperative adj6 analg*)).mp.

70. ((post-surgical adj6 pain*) or (post surgical adj6 pain*) or (postsurgery adj6 pain*) or (post adj surg* adj pain*)).mp.

71. ((post* adj pain*) or pain relief after or pain following surg*).mp.

72. ((posttreatment adj6 pain*) or (pain control after adj6 surg*) or ((post-extraction or postextraction or post-surg*) and (pain* or discomfort))).mp.

73. ((analg* adj6 postoperat*) or (analg* adj6 post-operat*) or (pain* adj6 after surg*) or (pain* adj6 after operat*) or (analgesi* adj6 after operat*)).mp.

74. ((pain* or analg*) adj6 (“follow* operat*” or “follow* surg*”)).mp.

75. exp Neuralgia/

76. (neuralgia* or neurodynia).tw.

77. ((neuropathic or nerve*) adj3 pain*).tw.

78. or/64-77

79. 12 or 19

80. 62 or 63

81. 78 and 79 and 80

82. limit 81 to humans

83. limit 82 to yr = “2016 -Current”

84. This search strategy includes terms from Marques et al. 156 and Blom et al. 157

Model conceptualisation and structure

The target population was individuals classified as having chronic pain 3 months after a total knee replacement surgery. The model was designed as a cohort Markov model with time-dependent annual transition probabilities and a time horizon of 5 years. The cycle length was 1year. The setting of the economic evaluation was patients treated by NHS hospitals in England and the study perspective that of the NHS.

Simulated individuals in the model received either the pain management intervention (STAR) during the first year only or usual care. After the first year, simulated patients can either remain in chronic pain (CP) or move to a non-chronic pain (NCP) health state (Figure 9). Patients move between CP and NCP states at any point. Mortality was not included in the model given its short time horizon and because the intervention is assumed not to have any impact on the risk of death.

FIGURE 9.

Model structure. TKR, total knee replacement.

We discounted health utility and costs after the first year at 3.5% per year following National Institute for Health and Care Excellence guidelines. 158

Data sources

Model inputs were parameterised based on evidence from the STAR trial120 for the intervention comparator and real-world data from both a population-based cohort study (COASt)121 and the CPRD for the usual care comparator (Table 2).

Transition probabilities

Using evidence from the STAR trial, we estimated the probabilities of transitioning from baseline CP to CP or NCP in year 1 (the cycle during which the treatment, i.e. pain management, was given). Based upon the evidence from the STAR trial, these transition probabilities differed between the two arms, (i.e. usual care and pain management). Subsequent transition probability estimates (years 2–5) were calculated using data from COASt and were the same for both comparators.

Costs

The estimation of costs included medication prescriptions, primary care consultations, hospital admissions, and the cost of the STAR intervention. Costs were taken from the intervention and usual care trial arms for year 1. For years 2–5, we extracted the annual per cent changes observed in (1) CPRD for prescriptions and consultations, and (2) COASt for hospital admissions for both the CP and NCP groups. We then applied per cent changes to the respective year 1 parameter estimates for each of the pain groups. Costs from the trial and COASt cohort considered only those associated with the pain of the intervention and treatment of the knee, whereas CPRD costs considered all reported primary care consultations and the cost of prescriptions for pain. Medication prescriptions, consultations and hospital admission costs differed between arms (i.e. usual care and pain management) as identified in the trial; however, the difference in hospital costs identified in the trial was only applied to year 1 and assumed to be different for the CP and NCP health states only thereafter. The cost of pain management was applied during the first cycle only and to the pain management arm only. We estimated costs separately for each health state (i.e. CP and NCP).

Health utility estimates

For the first cycle, health utility was estimated using reported EQ-5D-5L (mapped to the three-level valuation set)145 from the STAR trial. We used the baseline, 6-month and 1-year health utility estimates from the trial to calculate the mean QALYs for the health states (i.e. CP and NCP) for each of the comparators using the area under the curve method. We then adjusted these QALY estimates by controlling for baseline health utility using multiple regression. 159 Estimates of QALYs for years 2–5 were generated by applying the per cent of potential change160 (PoPC) observed over time in those with CP/NCP in COASt (identified at the 12-month point) to the QALY estimates for year 1 from the trial (also identified at 12 months). Under the base-case analysis, QALYs differed between comparators (usual care and pain management) and between health states CP and NCP.

Scenario analysis

For the base-case analysis, patients were classified as being in CP at baseline (i.e. 3 months post surgery). QALY estimates for both health states (CP and NCP) are different between usual care and pain management; in both cases CP QALY estimates are lower than NCP. Some the assumptions made about findings in the trial and especially for the years beyond the trial were varied and their impact on results examined via scenario analysis.

Probabilistic sensitivity analysis

To assess the impact of model parameter uncertainty on model results, we applied appropriate distributions to the transition probabilities, costs and health utilities. We ran 10,000 independent simulations of all input values. We used the results from the simulations to report the impact of parameter uncertainty on model results through a cost-effectiveness acceptability curve.

Results

Deterministic results indicate that the STAR pain management intervention would dominate usual care by leading to an expected 0.086 more QALYs over the 5 years (3.177 QALYs under the STAR pain management intervention compared with 3.091 for usual care) and £375 lower costs (£3189 and £3563, respectively). Assuming a cost-effectiveness threshold of £20,000 per QALY gained, this result would equate to an iNMB in favour of the STAR intervention of £2086 (95% confidence interval –£14,234 to £19,644]). The probabilistic sensitivity analysis suggests that, again at a cost-effectiveness threshold of £20,000 per QALY gained, the STAR intervention would have a 0.62 probability of being cost-effective (Figure 10).

FIGURE 10.

Base-case cost-effectiveness plane. PSA, probabilistic sensitivity analysis; WTP, willingness to pay.

If QALY estimates for health states were the same regardless of comparator, the STAR pain management intervention would still be dominant under the deterministic analysis, with an iNMB of £782 and 0.64 probability of being cost-effective. If the CP status of modelled patients was classified earlier at 10 weeks (instead of 12) post surgery, the deterministic dominance of the STAR intervention would still prevail with an iNMB of £1,925 and a 0.61 probability of being cost-effective. Finally, if the STAR pain management intervention was assumed not to have any impact on hospital admissions, then it would still lead to higher QALY estimates (3.18 vs. 3.09 QALYs) but would cost £243 more than usual care over the 5 years, leading to a deterministic incremental cost-effectiveness ratio of £2839 per QALY gained, an iNMB of £1469 and a 0.59 probability of being cost-effective according to the probabilistic sensitivity analysis.

Discussion

The model results suggest that the STAR intervention is likely to be both cost-saving and more effective than usual care during the first 5 years of implementation. Expected gains in quality of life are small however (0.086 QALYs), well under the reported minimally clinically important difference for patients having a knee replacement (0.182, based on PROMs data from the English NHS PROMS). 161 The probabilistic analysis shows that it is most likely that the STAR intervention would be cost-effective, but also that there is an important level of uncertainty around these results. The 95% CI ellipse in the cost-effectiveness reaches all four quadrants, meaning that it is possible that both extra costs and QALY gains could favour either comparator.

The expected dominance of the STAR intervention appeared robust to different assumptions about quality of life and the starting point of the intervention. Identifying the CP status of individuals at 10 weeks furthermore suggests that the time of classification is an influencing factor in the cost-effectiveness of the STAR intervention. Though the difference was small, our scenario results suggest that it may be beneficial to wait until at least 3 months post surgery to classify the CP status of individuals to improve both the expected iNMB and the probability of the cost-effectiveness of the intervention. Expected dominance was however sensitive to hospital admissions being less costly under the STAR intervention than for usual care, as we found hospital costs to be a key driver of the STAR intervention being cost-saving. However, eliminating any effect on hospital admissions still showed the STAR intervention likely to be cost-effective.

Limitations of the economic model included sourcing the model input data from three different sources as well as differences in the time points at which chronic pain categorisation occurred in each. The trial showed that those who receive the STAR intervention move out of chronic pain at a faster rate and sooner than those in usual care; however, this was not incorporated into the model as it used a cycle length of 1 year and so changes in CP status could not be considered earlier. Model results are hence likely to underestimate the benefits of the STAR intervention.

In conclusion, modelling the 5-year costs and quality of life of the STAR pain management intervention compared with NHS usual care based on trial findings and long-term observational data suggests that the intervention is likely to be cost-effective and possibly dominant. Gains in both costs and quality of life are, however, both small and largely uncertain. Long-term follow-up of the participants involved in the trial would help refine assumptions and potentially reduce the uncertainty around these results.

Background

The STAR care pathway for people with pain at 3 months after total knee replacement comprised an assessment appointment with an ESP and up to six telephone follow-up calls over 12 months. 120 The pathway aims to identify the potential underlying causes for chronic pain and enable onward referral to appropriate treatment. The clinical effectiveness and cost-effectiveness of the STAR care pathway was evaluated in a multicentre randomised controlled trial. Here, we report on an evaluation of the experiences of the health-care professionals who mobilised the new pathway at different trial sites.

Normalisation Process Theory (NPT)162 can be used as a framework to explore the work involved in implementing the new pathway. Understanding implementation processes can provide an understanding of how and why some new interventions might become normalised and embedded within routine practice, while others do not. 162 NPT provides four key constructs to understand the different kinds of work that people do when implementing a new intervention:162,163

1. coherence – how people individually and collectively make sense of a new intervention

2. cognitive participation – how people build a community of practice around a new intervention

3. collective action – the operational work that people do to enact a new intervention

4. reflexive monitoring – the appraisal work that people do to understand how a new intervention affects them.

It is possible to assess the potential of a new intervention to become normalised as part of routine practice by identifying and assessing factors that are known to affect implementation processes. 164 These factors can be identified using the NoMAD instrument, which is a 23-item measure of normalisation based on the four core constructs of NPT. NoMAD can be used to identify and understand implementation processes from the viewpoints of those directly involved in implementing interventions in health care. 81 The aim of this study is to understand how the STAR care pathway was implemented within a multicentre randomised controlled trial.

Reflexive monitoring

Participants described how the intervention impacted on them. They valued many elements of the care pathway, including a focus on neuropathic pain, psychosocial issues, an increased knowledge of pain management, and the formalisation and validation of referral practices. Some ESPs noted that involvement in the trial and delivery of the care pathway had improved their own practice and management of patients and had increased their awareness of long-term pain.