Process of change for people with knee osteoarthritis undergoing cognitive functional therapy: a replicated single-case experimental design study

Abstract

Purpose

To examine the applicability and process of change of Cognitive Functional Therapy (CFT) in the management of pain and disability in people with knee osteoarthritis who were offered knee replacement surgery and had risk factors for poor response to surgery.

Methods

Single-case experimental design with a mixed-methods, repeated measures approach was used to investigate the process of change through CFT in four participants. Qualitative interviews investigated beliefs, behaviours and coping responses, and self-reported measures assessed pain, disability, psychological factors, and function at 25 timepoints. Study registered with the Australian and New Zealand Clinical Trials Registry (ACTRN12619001491156).

Results

Qualitative data indicate that CFT promoted helpful changes in all participants, with two responses observed. One reflected a clear shift to a biopsychosocial conceptualisation of osteoarthritis, behavioural re-engagement and the view that a knee replacement was no longer necessary. The other response reflected a mixed conceptualisation with dissonant beliefs about osteoarthritis and its management. Psychological and social factors were identified as potential treatment barriers. Overall, quantitative measures supported the qualitative findings.

Conclusion

The process of change varies between and within individuals over time. Psychological and social barriers to treatment have implications for future intervention studies for the management of knee osteoarthritis.

IMPLICATIONS FOR REHABILITATION

• Cognitive Functional Therapy is applicable in the management of knee osteoarthritis.

• Reconceptualisation of osteoarthritis reflected a helpful change.

• Psychological and social factors emerged as barriers to recovery.

Keywords:

• Osteoarthritis
• biopsychosocial
• behaviour change
• chronic pain
• case report

Introduction

The personal and societal impact of knee osteoarthritis places substantial burden on health systems worldwide [1,2]. All major clinical guidelines recommend non-surgical care for all people with knee osteoarthritis [3,4]. However, up to 25% of knee replacement surgeries are performed on inappropriate candidates who are at high risk of non-response to surgery [5,6]. This is partly due to misconceptions among patients and clinicians about osteoarthritis and its management [7].

Common misconceptions are that pain and functional limitation associated with knee osteoarthritis are due to irreversible structural changes caused by ‘wear and tear’ [8]; and that loading the joint causes damage [7]. These beliefs impact on a person’s confidence to be physically active often leading to disability and distress [7–9]. In contrast, evidence supports that osteoarthritis-related pain and disability is influenced by the interaction of multiple biopsychosocial factors, which modulate inflammatory processes and behavioural responses [10–13].

This contrasting view reinforces the critical role of non-surgical approaches to manage knee osteoarthritis [14]. Non-surgical core treatments for OA include exercise, weight-loss and education and support for self-management [15]. Despite their efficacy [16], many patients do not respond, and may also be at risk of not responding to surgery [6]. These patients may require a personalised non-surgical approach that addresses multi-dimensional drivers of pain and disability [13].

Cognitive Functional Therapy (CFT) is a physiotherapist-led behavioural approach that targets modifiable multidimensional factors underlying a person’s pain and disability [17]. CFT aims to coach people towards self-management, by providing a personalized biopsychosocial understanding of their condition, building confidence to engage in feared, painful or avoided movements and activities, coaching a healthy lifestyle and self-management of flare-ups [17]. Efficacy trials demonstrated CFT to be more efficacious than usual care [18], traditional physiotherapy [19,20] and group education and exercise [21] for people with disabling back pain. While knee osteoarthritis and chronic back pain share similar multidimensional drivers of pain and disability [11,13], it is not known whether CFT is also applicable to people with knee osteoarthritis.

To understand the applicability of CFT in the management of pain and disability in people with knee osteoarthritis, we conducted a multiple single-case series. This study involved people with osteoarthritis who had been offered knee replacement surgery, but had risk factors of poor response to surgery including, higher body mass index (BMI), better function, poorer mental health and less severe radiographic knee osteoarthritis Kelgren-Lawrence grading [6]. The specific aim of this study was to understand the process of change through CFT, using a mixed-methods, repeated measures approach. Qualitative interviews investigated beliefs, behaviours and coping responses and self-reported measures assessed pain, disability and psychological factors, and function.

Methods

This study complies with the Single-Case Reporting guideline In BEhavioural interventions (SCRIBE) 2016 [22].

Study design

A triangulation convergent mixed-method design embedded in an A-B-A′ non-randomised, single-case experimental design study (SCEDs) with repeated measures (Figure 1). Qualitative and quantitative data were collected concurrently and integrated during analysis to enable rich interpretation of the process of change [23].

Figure 1. Schematic representation of the study design and procedures. Footnote to Figure 1. Short questionnaire measured the following items: patient-specific function, pain intensity, and psychological factors (bothersomeness, fear beliefs, knee confidence, pain control and pain catastrophising). Long questionnaires measured the following items: disability and pain (Knee injury and Osteoarthritis Outcome Scale using the daily living and pain sub-scales), and psychological factors including, fear beliefs (Knee Osteoarthritis Fears and Beliefs Questionnaire), Pain control (Brief Pain Coping Inventory-2), and Pain catastrophising (pain catastrophising scale). Functional tests included: 40m fast-paced walk test (Wright et al 2011), and a 30 second chair stand test (Jones et al 1999). Both functional tests are OARSI guideline recommended (Dobson et al 2013).

Phase A consisted of a 10-week baseline phase during which participants received no intervention. This phase acted as a comparator to determine within-person change [24]. During Phase A, two qualitative interviews (see Data collection section below) were conducted at weeks 1 and 10 (pre-intervention), and weekly quantitative measures were collected on up to 10 occasions. Phase B was a 12-week intervention phase where participants received up to 8 sessions of CFT [17]. During Phase B two interviews were conducted, one after the first CFT session (early-intervention) and one at the end of the intervention (post-intervention); with weekly quantitative measures collected on up to 12 occasions. As the effect of behavioural interventions are expected to carry over after the intervention ceased, the subsequent Phase A’ was used as a follow-up monitoring period during which participants received no intervention. This phase consisted of the same weekly measures over 1 week at three different timepoints (6, 9 and 12-month) after commencing treatment. No qualitative data was collected during Phase A′.

The study was registered with the Australian and New Zealand Clinical Trials Registry (ACTRN12619001491156). There were no protocol deviations, but this paper reports non-obese participants only. Obese participants were offered an additional treatment phase of psychologically-based weight management and will be reported separately to adequately capture the extra complexity.

Participants

Eligible people were (i) aged ≥30 years with knee osteoarthritis presenting to orthopaedic surgeons or general practitioners in Perth, Australia; (ii) had been offered knee replacement surgery; (iii) had >15% risk for non-response to surgery according to a prognostic nomogram [6], and (iv) were willing to undergo CFT. Exclusion criteria were (i) previous major ipsilateral knee surgery; (ii) pregnancy; (iii) compromised cognitive state (e.g., psychiatric disorder); (iv) comorbidities causing severe mobility impairment (e.g., Parkinson’s disease).

Seventeen people were screened for eligibility. Sequential candidates were purposively selected to capture different profiles of factors assessed by the nomogram, resulting in four non-obese participants (BMI <30 kg/m²).

This study was approved by the Health Research Ethics Committee at Melbourne University (1646845); and reciprocal approval was granted at Curtin University (HRE2018-0097). Written informed consent was obtained from each participant.

Intervention

The CFT intervention was delivered face-to-face in a clinical setting by three specifically-trained physiotherapists who had reached an acceptable level of competency and had at least 100 h of experience in delivering CFT. The CFT intervention aimed to coach people to self-manage their pain symptoms using three principle processes: (1) making sense of pain; (2) exposure with control; (3) lifestyle coaching (see Table 1 for description of the principles, and a detailed CFT management for each case).

Table 1. Multidimensional factors and cognitive functional therapy (CFT) management for the four clinical cases.

Examination findings	Case 1	Case 2	Case 3	Case 4
Interview
History	60-Year-old female, property manager and cleaner developed right antero-lateral knee pain insidiously over the last 3 years. She continued to work and be active but the pain started interfering with daily activities. Progressively reported worsening of function. She was scanned and told by a Surgeon that she had degeneration in her knee and would need a total knee replacement in the near future. She had undergone massage, knee and hip strengthening exercises, orthotics, pain medication, and a hyaluronic acid injection, without long term benefits. She had no long-term self-management plan.	50-Year-old male, manual worker with right knee pain for the last 30 years, and worsening symptoms over the last 10 years. At age 16, injured his knee playing football and had a meniscectomy and several subsequent arthroscopic ‘clean ups’ with short-term benefit, but progressive functional deterioration. He was scanned and told by the surgeon that he had the “knee of a 65-year-old” and that a knee replacement is warranted. Rehabilitation over the years has included gym exercises, but it has been inconsistent. Currently, he has no active management plan. His pain condition is worsening in severity over the past 12 months.	73-Year-old male, farmer with bilateral knee pain worsening over the last four years. Significant illness 4 years ago. Currently impacting on working on the farm climbing silos, getting in and out of the truck, carrying, etc. He had several arthroscopic ‘clean ups’ and bilateral lateral release surgery followed by rehabilitation with limited success. Stopped playing tennis which he described ‘kept him fit and in touch with friends’. Orthopedic surgeon reviewed latest scans and said that a total knee replacement is inevitable. However, he should try to hold off due to medical history including bowel cancer and recent infection. Condition worsening. He has no current rehabilitation plan.	66-Year-old female, business manager with life-long history of right knee pain. Pain gradually worsened over the last 15 years, impacting on her ability to walk long distances, play tennis, deal with stairs, and get up from sitting. Has seen several GPs and physiotherapists. Tried different exercise programs, including Pilates, GLA:D and an individualized gym program with limited success. Orthopedic surgeon said the cause of the pain was cartilage degeneration due to having bowed legs, and recommended knee replacement surgery. Currently attending rehabilitation class but not noticing any improvement.
Patho-anatomical factors	Grade 1 Kellgren-Lawrence score on X-ray and MRI.	Grade 2 Kellgren-Lawrence score on X-ray and MRI.	Grade 3 Kellgren-Lawrence score on X-ray and MRI.	Grade 3 Kellgren-Lawrence score on X-ray and MRI.
Pain characteristics	Antero-lateral knee (patellofemoral joint) pain with loading activities such as during sit to standing, stepping up, and stepping down.	Postero-medial knee pain along the tibiofemoral joint with knee flexion such as sitting and kneeling, and with load in flexion such as climbing up and down ladders.	Antero-lateral knee pain along the tibiofemoral joint with walking, squatting, knee flexion climbing stairs, and lateral movements or changing directions.	Antero-medial knee pain along the tibiofemoral and patellofemoral joints on palpation, and with activities such as walking, stepping down, and getting up and down from sitting.
Psychological factors: Cognitive	Beliefs – knee pain during activities was causing more damage to her knee. Low pain self-efficacy. Avoidant coping style relating to loading, and physical activities.	Beliefs – “I’m old and I used my knees too much…this (pain) is wear and tear…not much space in there”. Very careful with knee	Beliefs – pain related to damage in knee joints due to repetitive loading; not much can be done about it. Worried about causing more damage – protective of the knee, avoidant of painful activities.	Beliefs – pain related to wear and tear of the cartilage due to having bowed legs. Worried about further damage and the potential need for knee replacement surgery. Very protective of the knee.
Emotional	Fear of movement and loading Fear about her capacity in the future Keen to “try and get another 10yrs out of the knee”	High pain-related anxiety Low mood due to pain Fear of movement and loading Hopeful about the future, but unsure on what to do	Fear of movement and loading Fear about future and overall health Worried that mother had knee issues	Fear of movement and loading Fear about future and need for surgery High social stress
Social factors	Working Good social support. Physically demanding job	Self-employed – work is biggest worry Impacting on interaction with family Little social support Significant ongoing stressors (work; home)	High social stress Significant ongoing stressors Misses socializing through tennis	High social stress. Positive social interactions in exercise class
Lifestyle factors	Active but has progressively been avoiding exercise and stopped playing squash, and horse riding due to pain, but she was still able to ski.	Walks the dog daily (20–30 min) Inconsistent with the gym Weight loss due to stress last year Poor sleep Work is physically demanding	Work is physically demanding Poor sleep – averages 5–6h/night (diagnosed with obstructive sleep apnoea – wears air pressure machine)	Reduced physical activity significantly (walking and tennis) High demands at work and home Little time for herself Moving less and less
Health related factors	Good overall health Within normal weight range	Poor mental health affected by pain and stress overweight	Poor mental health affected by own health and family stress Within normal weight range	High stress Slightly above normal weight
Physical examination
Sensory profile	Localized hyperalgesia over the lateral aspect of the patellofemoral joint and lateral knee soft tissue with light palpation. Pain increases with knee loading.	Localized hyperalgesia over the postero-medial aspect of the tibiofemoral joint. Pain increases with knee flexion. No pain with loading, unless flexion loading.	Localized hyperalgesia over the antero-lateral aspect of the tibiofemoral and patellofemoral joints. Pain increases with knee loading.	Localized hyperalgesia over the antero-medial aspect of the tibiofemoral and joints. Thigh muscles – tense and sensitive on palpation. Pain increases with knee loading.
Observation of pain provocative functional behaviours	Pain reported during sit to stand and stepping up and down. Sit to stand: Tendency to avoid right leg loading during load transfer. Internal hip rotation of right hip was also noted. Stepping up and down: Upright trunk posture, tensing quadriceps muscle and stiffening knee was noted.	Keeps right leg straight in sitting; puts right leg forward and loads the left leg when getting up from sitting to standing (avoidance of R leg loading). Guarded during active knee flexion (limited to 80 degrees of flexion with report of pain), braces quadriceps and breath-holds. Avoidance of right leg loading during standing. Loss of knee flexion during step down, guarding the knee, hurrying the movement and reluctant to load it during stepping up and stepping down.	Keeps feet tucked under the chair in sitting Use hands to stand up from sitting Braces and breath-holds when loading the knee; moves slowly Weakness and pain pushing off on lateral movement and tennis specific movements	Use hands to get up from sitting; narrow base, breath-holding – slow, stiff and painful. Braced into extension in standing and walking – stiff, unsteady, slow with report of knee pain. Walking with a narrow base and minimal push off, keeping ankles ‘locked’ in neutral (stiff legs). Avoidance of right leg loading during stepping down.
Guided behavioural experiments	Sit to stand: forward bend of the trunk, aligning knee over foot and equalising weight though both legs reduced pain. Stepping up and down: Forward bend of the trunk with a focus on driving power into the foot with a relaxed knee and knee alignment over foot reduced pain. Repetition of these tasks with reduced pain allowed the physiotherapist to guide the patient to reflect on the experience where relaxing and loading the knee provided pain relief – and protecting and avoiding leg loading increased knee pain.	Active knee flexion with focus on engaging hamstrings while relaxing of the quadriceps resulted in reduced the knee pain and increased flexion. Sit to standing with knee flexion and loading of both legs did not increase his pain. Although the knee was pain sensitive during repetitive sit to standing, it did not result in an increase in knee pain. Repetition of knee flexion and loading resulted in greater movement without an increase in pain disconfirming the belief that bending and loading is harmful.	Relaxation of the trunk and diaphragm breathing while getting up from sitting – encouraged to use a wide stance and push through legs to get up (instead of trunk and arms) – led to improved ease of movement and reduced intensity of pain. Side stepping keeping body relaxed, diaphragm breathing, and pushing through the legs to move was less painful and more mobile. Repetition of these tasks increased his confidence and challenged his beliefs that loading equals damage and pain.	Breathing and relaxation of thigh muscles before getting up from chair, leaning body forward and transferring weight to centre of feet; wide base and pushing through the feet to stand up led to pain reduction and increased speed. Walking with a wider base, relaxing the ankles and engaging feet during push off led to less discomfort and ease of walking. Engaging foot when stepping down led to increased confidence and reduced discomfort. Repetition of these tasks increased her confidence and challenged her belief that protection was necessary.
Conditioning	Repeated loading during sit to stand – and stepping resulted in fatigue of the leg muscles.	Deconditioning of hamstrings and hip extensors.	Deconditioning of quadriceps, hamstrings and calf muscles.	Weak calf muscles (soleus). General deconditioning.
CFT intervention	Case 1 6 sessions over 3 months	Case 2 6 sessions over 3 months	Case 3 8 sessions over 3 months	Case 4 8 sessions over 3 months
Making sense of pain Education focused on reconceptualising pain within a biopsychosocial context related to the person’s story. This process addressed any misconceptions about knee osteoarthritis the participant held (e.g., about the role of radiological imaging, the meaning of pain, activity avoidance and movement). Participants were guided to reflect on new ways to conceptualise their condition, behavioural responses to it and their lifestyle, in order to reach their functional goals and improve their knee and general health.	Explain how negative beliefs, fear of movement, activity avoidance and protective muscle guarding sets up a vicious cycle of pain sensitivity and loss of functional capacity. Behavioural experiments disconfirmed the belief that pain equals harm, and led to an experience that that loading her knee is safe. Reinforce that her knee health is optimised by a positive mindset towards her knee, engaging in regular, graduated and confident movement and loading of the knee. Also educate that her pain originates from the patella-femoral joint and not the tibiofemoral joint. Goals – achieve pain control during functional tasks, return to physical activity including horse riding and potentially squash.	Explain that knee replacement surgery is good for knees that can’t sustain load, not for knees that can’t bend. Avoidance of bending and loading the knee is unhealthy for it and keeps it sensitive. Explain how stress, poor sleep, increased body weight, high physical demands and low physical capacity can interplay and lead to joint sensitivity. The loss of knee flexion is associated with quadriceps guarding and weakness/inhibition of the hamstrings. Behavioural experiments disconfirmed the belief that protecting the knee is helpful, and confirmed that the muscles that bend the knee are inhibited, but once engaged he can bend the knee more. Goals – return to normality and do things without having to worry, including playing with the kids, getting in and out of the car, kneeling and climb ladders.	Explain how his activity levels and lifestyle significantly changed after his illness leading to a loss of muscle strength and likely contributing to joint sensitivity. The adoption of movement habits to protect the knee is not helpful and makes the knee sore and less healthy, setting up a vicious cycle of disuse, protection and tissue sensitivity. In addition, the social stressors and poor sleep may contribute as pro-inflammatory factors and further sensitise the knee joint. Behavioural experiments challenged his belief that loading the knee causes damage and pain, building his confidence to use the knee more. Reassurance that despite the changes seen on the scan, the sensitivity of his knees is also influenced by other factors such as muscle weakness, protective guarding, stress and sleep that are modifiable. Goals – reduce pain to work without worrying about his knee; return to playing tennis (singles).	Explain that the adoption of protective habits such as avoidance of movement, activity and stiffening of the knee and ankle increased knee load, contribute to a vicious cycle of disuse, protection and tissue sensitivity. It was also discussed that her activity levels have reduced over the last couple of years due to husband’s health and how this impacted on her conditioning, confidence, and tissue sensitivity. Behavioural experiments disconfirmed the belief that protecting the knee is helpful, and provided confidence to engage her knee and leg in a more relaxed manner during walking, sit to standing and stepping down. It was also discussed how stress and body weight can influence tissue sensitivity, and the importance of regular physical activity on her overall health and well-being. Goals – reduce pain, avoid surgery, and stay active (gym, play tennis, walk).
Exposure with ‘control’ Guided behavioural experiments aimed to promote pain controllability, and confidence, by discouraging unhelpful protective behaviours (e.g., reducing protective guarding through body relaxation) while engaging in provocative, feared or avoided functional tasks. Discrepancies between pain expectations and actual pain experiences were addressed, providing experiential learning opportunities. Newly learnt movement behaviours were then incorporated into valued daily tasks in a graduated manner, while respecting the person’s level of pain and distress. Strengthening and conditioning exercises were introduced according to participant’s presentation and goals.	Sit-to-standing, stepping up, and stepping down with a focus on adopting a relaxed trunk and forward lean posture while driving the movement from her feet. Visual feedback with mirror, video and written instruction. Integrate new strategies into daily life. Graduated conditioning with graded progressions towards sustained and repeated squats, forward trunk lean – focus on engaging the hip and foot while loading both legs equally.	Replace guarding of knee flexion by engagement of knee flexion (hamstring) muscles. Sit with both knees flexed. Sit-to-stand with equal knee flexion and loading Squatting with equal weight to facilitate eccentric lengthen of his quadriceps. Graduated targeted exercises to retrain knee flexors. Stairs training, taking weight through the leg with knee flexion, relaxed and controlled (not hurrying). Encouraged to integrate new movement strategies at work and at home, including kneeling and stepping up and down a ladder.	Replace guarding of the knee and avoidance of pushing through the legs, by relaxing the body and engaging the legs through graded movement and loading of the knee. Sitting with knees relaxed. Sit-to-stand keeping the trunk and arms relaxed, and loading both legs equally. Graded exposure to leg loading. Graduated exercises to target conditioning quadriceps, hamstrings and calf muscles (e.g., squats, steps, ½ lunges). Progress program towards tennis-specific exercises and practice. i.e., Side-side step movements keeping the trunk and arms relaxed, and engaging the legs (mimic tennis).	Replace knee protection with relaxation of thigh muscles and engagement of knee and foot. Sit to stand with wide base, relaxing body forward, no hand support, and engaging the feet and legs to stand up. Walking at a faster and more relaxed pace, wider base, and engaging foot and ankle to move forward Taking weight into leg and controlling own weight with foot when stepping down. Graded exposure to leg loading. General and targeted conditioning of the lower leg and feet muscles
Lifestyle change Participants were encouraged to perform physical activity based on their goals, preference, and accessibility, if they were not doing so. Physical activity was gradually increased in a time-contingent manner. For instance, if the participant’s goal was to be able to walk for 60 min daily, this activity was introduced at the level of their ability and progressed on a time-contingent fashion (e.g., increase of 10% or initially progressing frequency, then time). While engaging in physical activity, emphasis was placed on relaxed, confident movement. Where a person reported high level of stress or poor sleep, relaxation techniques and healthy sleep habits were taught.	Physical activity – daily walking program progressed to gym based on person’s preference.	Physical activity – continue with walking; added graduated stationary cycling to train knee flexion. Weight loss – importance of healthy diet, good sleep habits, and consistent exercise to reduce body weight and decrease inflammatory load on knee.	Physical activity – encouraged to continue being active, and to relax and carry on in the presence of pain (rather than worry, protect and rest until pain subsides). Based on preference – bike, walk, tennis. Sleep – aim for regular sleep routine	Physical activity – encouraged to make time for herself to exercise daily. Encouraged to exercise every day for at least 30 min. Weight loss – importance of healthy diet, and regular exercise to reduce body weight, manage stress, and decrease inflammatory load on knee.

Treatment dosage

Up to 8 sessions in total over the 12-week intervention. One-hour initial session, and 30–45 min follow-ups. Participants were seen weekly for 2–3 sessions and progressed to one session every 2–3 weeks.

Treatment adherence

Participants were encouraged to keep diary on their exercise/activity. Physiotherapists kept a ‘therapist log’ including clinical notes and communication with participants over the study period.

Treatment fidelity

The developer of CFT (POS) was present as an observer in the first session, as well as at a follow up (WEEK 6 or 12), using a fidelity form previously used [19].

Data collection

Qualitative component

Participants completed four one-to-one semi-structured interviews (30–60 min) conducted in person in a private room in a physiotherapy clinic or over the phone by an academic physiotherapist and experienced qualitative researcher (NRK), not previously known to the participants or involved with their treatment. Interviews were structured on the Common Sense Model [25], a health belief theory comprised of four dimensions: (i) understanding of pain, its causes and consequences; (ii) perceptions of pain controllability; (iii) behavioural responses and (iv) emotional responses to pain. This model has previously been applied to understand the process of change through CFT in participants with chronic low back pain [10] (Supplementary A–D). Interviews were audio recorded and transcribed verbatim. At the end of the treatment (Phase B), a short interview with the participant’s partner was completed to support any observed changes during the intervention period (Supplementary E).

Quantitative component

Collected weekly via an electronic short questionnaire during baseline (10 weeks), treatment period (12 weeks) and follow-up period (once at 6, 9 and 12 months after commencing treatment):

1. Patient-Specific Function: Patient Specific Functional Scale (PSFS) [26]. Participants nominated three activities they found difficult or were unable to perform because of their knee problem. Participants rated how well they were able to perform the activity compared to before their problem on an 11-point numerical rating scale anchored by ‘0 = able’ and ‘10 = unable’. The PSFS score reflecting the most disabling activity was used for analysis.

2. Pain intensity: measured as average intensity in the last week rated on an 11-point numerical rating scale (NRS – 0–10).

3. Psychological factors: Fear measured by two items from The Knee Osteoarthritis Fears and Beliefs Questionnaire [27]; Knee confidence measured by a single-item from the Knee Injury and Osteoarthritis Outcome Scale (KOOS) [28]; Pain control measured by a single-item from Coping Strategy Questionnaire [29]; and Pain self-efficacy measured by the Pain Self-Efficacy Questionnaire-2 [30]; Pain catastrophising measured by 3-items from the Pain Catastrophising Scale (PCS) [31].

The use of standardised outcome measures with established psychometric properties provides information about whether participants have made a meaningful change [32,33]. The following long questionnaires and functional tests were collected twice during baseline (weeks 1 and 10), three times during treatment period (weeks 4, 8 and 12) and follow-up period (at 6, 9 and 12 months after commencing treatment):

1. Disability: The KOOS Function sub-scale (KOOS-daily living). Minimal clinically important change (MCIC) of 8–10 points in each subscale is considered appropriate for the KOOS [34].

2. Pain intensity: KOOS Pain Sub scale (KOOS-pain).

3. Psychological factors: The Knee Osteoarthritis Fears and Beliefs Questionnaire (MCIC not available); Pain control measured by a single-item from the Coping Strategy Questionnaire and the Two-item Brief Pain Coping Inventory-2 [35](MCIC not available); and Pain catastrophising measured by the PCS (MCIC of 6 points) [31].

4. Function: two OARSI guideline-recommended [36] physical tests, the 40m fast-paced walking test [37], and the 30s chair stand test [38]. The MCIC for these tests are respectively, a change of 0.2–0.3m/s in walking speed, and a change of 2–3 sits to stands [37].

Qualitative data analysis

Transcribed interview data were analyzed following a 5-step Framework Approach [39]. Each step involved two analysts experienced in framework analysis (NRK and BIRdO) who met regularly to agree on interpretations through discussion. Step 1: Data familiarization through reading and re-reading transcripts collected at each time point. Step 2: Coding of data using deductive codes based on the Common Sense Model. Coded data were uploaded into NVivo 12 (QSR International, Melbourne, Australia) to facilitate further analysis. Step 3: Charting of coded data onto a matrix created for each participant with the four common-sense model dimensions as row headings and each time point as column headings. Step 4: Interpretation of salient perceptions at each timepoint for each participant and how they change over time. Step 5. Description of process of change for each participant presented as themes. Steps 4 and 5 involved group discussion to challenge and agree on emerging interpretations, Group discussion involved members of the research team who were physiotherapists not involved in the CFT intervention.

In the results, a narrative description of the process of change is supported by a tabular summary, in which selected quotes illustrate changes over time for each participant (Table 2).

Table 2. Process of change for each participant.

	Phase A	Phase B		Phase A′
Case 1	Pre-intervention	Early-intervention	Post-intervention	Follow up
Understanding of pain (cause)	Biomedical conceptualisation of the cause of her pain “… I’ve been told by the doctors that there’s no cartilage and the bones are rubbing.”	An alternative understanding of the cause of her pain “… I need to exercise to use my muscles properly and get my knee mobile without surgery and less pain.”	A shift in her conceptualisation about osteoarthritis “… the pain is less, definitely… it must be the way I use it, and how much I use it … I’ve realized the value of getting strong and active without fear. Understanding it was healthy to use my knee was key”
Pain control	No control over pain, and avoidance was the only strategy to manage symptoms “… oh no, no I can’t control it. I can’t stop the pain from coming unless I stopped the work. Or stopped the exercising or whatever I’m doing with my knee.”	A new and helpful way of using her knee “… different to what I thought, using my knee is actually helpful.”	Pain control through active engagement. “… just being able to use the leg a lot more confidently and with less pain. I can do leg exercises now which is great”
Behavioural response	Exercise was ineffective. “I tried exercising but it hurt my knee more…. resting also doesn’t work……I’m just trying to think what else could deteriorate your bones so quickly you know? Apart from old age!… … sometimes my knee will lock up and I don’t trust it walking downstairs.”	Exercise may be helpful. “… the exercises, are all to strengthen glute muscles and to teach me to use my knee differently…and it helps.”	Exercising confidently to keep joint healthy. “It (this program) definitely made a difference that’s for sure… so I don’t feel like I need to go and have surgery now …just keep working at the exercises, you know. But confident.”
Emotional response	Low knee confidence and worried about causing damage. “I’m lacking confidence… and I worry about the pain and damage that (using the knee) it’ll cause because it does, it swells up.”	Hopeful about the approach. “… it’s quite exciting to think there is maybe another way…to get my knee better without surgery and less pain. That gives me hope”	Improved knee confidence and reduced worry “I’m able to use it a lot more and I don’t worry about it – I don’t feel so frightened about using it… it feels stronger so I’m more confident. It’s great, and I don’t feel that I need surgery or anything at this stage.”
Case 2	Pre-intervention	Early-intervention	Post-intervention	Follow up
Understanding of pain (Cause)	Pain caused by a worn-out knee that is bone on bone.. “…on the scan, it was just basically bone on bone sort of thing, you know, there was no sort of gap there where cartilage should be…”	Pain due to damage and the way he uses his knee.. “(…) it’s not being used properly you know like my muscles and stuff around it … I am not bending the knee like I should.”	Pain influenced by damage and behaviour. “… you go and see the specialist and he just basically said after you seen the x-ray, you can see how degenerated it was I suppose and that’s where I thought the pain was just from that … but obviously, the pain is also from me not using it correctly. The damage is done, but I’m finding that there’s things that can be done to help… it won’t fix the inside of the knee though.”
Pain control	Avoidance of painful or potentially damaging activities. “I try and watch what I’m doing a lot… if I can avoid getting up on the ladder or something I will. I may take some painkillers, but generally I just try and rest it up.”	Changing the way in which he uses the knee may help. “… the [physiotherapist] just went over a lot more things and just made a lot of sense of what’s wrong with me and what I need to do to make things a bit better….so, I’ll try to change some bad habits and use this leg a bit more now.”	Engaging with knee in a new way has been helpful. “Mostly it’s not hurting now so, I gotta continue to do that and eventually I will be able to bend it more and then I won’t have any pain at all…but kneeling is still sore…the inside is still damaged – hopefully this program will still help it.”
Behavioural response	Protection and avoidance. “Nothing really helps…I tried exercises but … not having cartilage in the knee and just the wear and tear on it over the years…you just don’t wanna make it worse.”	Protect less and use more. “the plan is good, yeah….I’m gonna probably just continue on, use my knee more, and hopefully strengthen around it and make that sort of like less painful in the knee joint …”	Active engagement with less protection.. “…I seem to be able to do more and it’s not hurting now. I mean just even just going up stairs – I can sort of do it easier, than what it used to be like… I hope it continues this way but with time I am not sure this will be enough I still have trouble kneeling and stuff like that…”
Emotional response	Fear about future ability to work and interact with kids. “I’ve got my kids playing footy now and I’d like to go and have a kick with them but I can’t… I got to be careful … it’s sad when you can’t play with your kids. Also, at work, “… you can’t say ‘I don’t want to do that job because it’s going to hurt my knee’… it’s scary and is making me sort of think about my future….my career.”	Happy about having a sensible explanation and plan.. “I was just a lot happier yesterday now that I’ve seen someone and found out that… I can sort of rectify this problem.”	Low confidence in key tasks; unsure about surgery. “I’m hoping, it’s improved that much that I wouldn’t even have to look at getting a knee replacement…it’s heading that way, I can mostly use it more, but, I’m still worried I will need one…. time will tell.”
Case 3	Pre-intervention	Early-intervention	Post-intervention	Follow up
Understanding of pain (cause)	Pain caused by joint degeneration. Further damage is inevitable. “You’ve got bone on bone… if that cushion is gone and it starts wearing away the bone structure and the contact areas I can’t expect improvement, I wouldn’t expect improvement I’d expect further deterioration and probably further pain and increased angle or – I’d like to think the reverse but I’m trying to be a realist.”	Pain caused by degeneration, protection and weakness. “My knee is worn out and I have been compensating for my knee, sort of protecting it. So, my muscles around the knee are weak and they are not supporting the knee…there is more pressure in the joint, and pain.”	Using the knee is good, but further damage is possible. “What I’ve learnt is that you shouldn’t be protecting it, it should be taking some of that pain, and making the joint work to feel better, which is what I wouldn’t have expected. I’d say there’s still some reservations about damaging a bit of bone… Well when you look at the – my x-rays like, because they’re so lop sided and um, you know out of alignment… and there’s no meniscus or cartilage there to cushion… I’m yet to be convinced.”
Pain control	Control of pain through avoidance. “Not doing things is the only way to control it really!”	Exercise may be helpful. “Some exercises that make me use my knee and make my muscles stronger. I shouldn’t be protecting my knee so much, I should be out there doing more with it.”	Exercise has been helpful. “I only get pain when I push the limits um, it’s not very restrictive at all anymore. But if I were to lunge or move quickly in tennis, I still get it…not much I can do about that.”
Behavioural response	Avoidance of important daily tasks. “Not much has helped the pain. I had deterioration in the last four years… I’ve started getting pain when I do the activities I enjoy, then going down stairs and putting socks on, you only got to bend a little fraction at a certain angle and you get that jabbing pain… quickly it (knee) just collapses and gives me a sharp and really painful short period of time and I back off – I can’t stand, it – the knee collapses. I try to avoid these activities but it’s hard.”	There are better ways to use the knee. “I learned some strategies to sort of make me use my knee in a better way, so I can do more with it”	Actively engaged but not confident “My function is less and less restricted, still some restriction… bending down to pick stuff up off the floor and drying my feet like, there’s still pinching sometimes, but less pain and I don’t feel quite as clumsy. But the damage is still there, so you never know…”
Emotional response	Frustration with loss of function and social interaction.. “It’s frustrating, but at least I know if all else fails I can get a knee replacement and so that – that’ll be compromised ability to move but I’ll be free of pain and won’t have to worry about the knee as far as that (…) we are now the old people so trying to get another social group where you enjoy socialising is as important as the activity.”	Hopeful with plan and the potential to prevent surgery. “I am hopeful that there is something I can do to improve my pain and maybe prevent surgery.”	Fearful that may still need surgery. “I fear I may not be able to avoid surgery… because of my previous experience with surgery (compromised drainage) there’s an extra incentive to avoid surgery, but you ought to be careful with it… As I said, I’m yet to be convinced this will work.”
Case 4	Pre-intervention	Early-intervention	Follow up
Understanding of pain (cause)	Pain caused by degeneration and damage in the knee. “… it’s obviously damaged so it’s swelling up … it is a hindrance…. I guess it’s a degeneration of some sort and I think spurs grow and all sorts of things happen that um, and I don’t think I have much cartilage left in the knee so – in fact there’s probably none… I had an MRI on my right and that told me that. The x-rays I had on both… said it was bone on bone.”	Broader understanding about pain and its contributors. “I know there’s inflammation….it all comes to influence it you know, what stresses are in your life and how they affect you… I understood where the physiotherapists were coming from, and where they were heading. How much things outside of your knee that you don’t imagine actually affect you.”	Biopsychosocial understanding about pain. “… now I know stress can cause problems in your body, wherever; obviously, my problem is my knees.”
Pain control	Protection and avoidance to control pain. “Being careful with my knee is the only way to have a bit less pain. I don’t want it [the knee] to give way. If there’s a nerve pain there and it suddenly grabs then I don’t know what would happen.”	Relaxation and active engagement may control pain. “New ways of using my body, protecting less and using it more. They are saying that pain is not necessarily a sign to stop, it’s a sign that you need to do more I guess to, to right the problem.”	Reframing pain helped active engagement. “Understanding it is not all about damage helped a lot, and enabled me to use it more, do more exercise, walk and even play tennis… I guess all the exercise helps inflammation and that has helped the pain too.”
Behavioural response	Exercise is provocative. “I have to be careful with my knee; it catches…. when it catches, I can’t walk or I could fall I suppose. I have to watch, I do watch it…I tried exercising and being active, but when I do a bit more with the knee, it swells up and it aches…it’s the damage in there.”	An active plan is helpful. “[The physiotherapist has] given me a plan with all the things I have to do…all very sensible, and I’ve got it on video so I can see what I’m supposed to do and all of that. I have been doing more and, it’s been good so far”	Being active is key for joint and overall health. “So generally they’re heaps better. I can do more things than I could before. I used to not be able to walk very far, now I can walk further, I can do steps and I can actually go down steps, it’s much better”
Emotional response	Fear of surgery as the only option. “Well I’ve been told there’s no option but a replacement…that scares me… so I don’t really want that so I will do as much as I can to stop that but in 5 years’ time I don’t know… and it can be quite a dehumanising process going to hospital. Especially publically…”	Relief about having a plan to self-manage her condition. “I was feeling quite lost really…. It is relieving to have a clear and sensible plan. There is hope!”	Confident about a future without surgery. “Some of the things I can do now, I couldn’t do before. It still hurts at times but I don’t worry about it anymore; I am confident it is ok and that I can keep going without the surgery…Well, personally I think it’s now up to me, to keep doing what I’m doing.”

Footnote: The Radar graphs displaying the participant’s change in patient-specific function, pain and psychological factors across the three phases of the study (Phase A – pre-intervention; Phase B – early after commencement of the intervention and post-intervention; Phase A’ – long-term follow up after commencement of the intervention). Outcomes have been adjusted so that a reduction in value indicates improvement for all outcomes. The graph ‘PRE’ graph displays average data from weeks 1 and 10 of the baseline period. The graph ‘EARLY’ displays data for week 2 after commencement of the intervention. The graph ‘POST’ displays average data from weeks 6 and 12 of the intervention period. The graph in Phase A’ displays data for one week at months 6, 9, and 12 after the commencement of the intervention.

Qualitative description using the Common Sense Model constructs across three timepoints of the study, and supporting objective data across the three phases of the study.

Quantitative data analysis

Weekly measures of disability, pain and psychological factors were graphed for visual inspection of baseline stability and change during Phase B. The academic physiotherapist who collected the data (JST) was not involved in the statistical analysis. Simulation Modelling Analysis (SMA Version 8.3.3, http://clinicalresearcher.org) was used as a statistical test for significant change in each variable between phases A and B after adjusting for autocorrelation [40]. SMA is a nonparametric bootstrapping approach effective at handling short time series data-streams, whilst both adjusting for autocorrelation and reducing the risk of Type I and II errors [40].

Data from the long questionnaires and functional tests were tabulated for all eight timepoints over all three phases. Change scores from Phase A to B were calculated by subtracting the average score from weeks 1 and 10 in Phase A, from the week 12 score in Phase B; and Change scores from phase B to A’ were calculated by subtracting the week 12 score in Phase B, from the average scores from months 6, 9 and 12 in Phase A’. Change scores were referenced against consensus MCIC change where available.

Data integration

To facilitate interpretation of the process of change within and between participants, qualitative data was tabulated alongside radar graphs displaying changes in the weekly measures for each participant (Table 2). To support attribution of any systematic change between Phase A and B to the intervention, the interview content (participant and their partner/significant other – Supplementary E) and therapist log were examined to identify explanations for change other than the intervention.

Role of funding source

The funder played no role in the design, conduct, or reporting of this study.

Results

Participants’ profiles and nomogram score

Table 1 outlines a detailed description of the four cases. Appendix Table A1 describes the participants’ profiles on the risk factors assessed by the nomogram. Participants had different profiles of risk relating to mental health and disability. All participants completed the treatment and all assessments across the three phases of the study. There were no adverse events, and all participants attended all sessions as scheduled. The clinicians reported in the therapist log that for all participants, their functional movement behaviours were modifiable and pain was controllable via guided behavioural experiments used in the first consultation.

Qualitative findings

Two themes describing different responses to the approach were observed. One (illustrated by Cases 1 and 4) reflected a clear shift to a biopsychosocial conceptualisation of osteoarthritis, behavioural re-engagement and the view that a total knee replacement was no longer necessary. The other response (illustrated by Cases 2 and 3) reflected a mixed conceptualisation (biomedical and biopsychosocial) with dissonant beliefs about osteoarthritis and its management. While these two participants were more behaviourally engaged (physical, work and social) after CFT, they were uncertain about their future need for knee replacement surgery. See Table 2 for a detailed description of the process of change for each participant.

Significant others

All significant others reported that their respective partner had a similar positive response to CFT that impacted not only on pain, but also on their partner’s ability to engage with meaningful activities be it physical, social or work-related (Supplementary F).

Quantitative findings

On visual inspection, all participants showed a stable baseline (Phase A) for all weekly measures of patient-specific function and pain, and pain-related psychological factors. Findings from the weekly measures and corresponding statistical tests for change from Phase A to B for each participant are presented in Figure 2 and Appendix Table A2. These, provide evidence for change for case 1 for patient-specific function and four of five psychological measures, and for case 4 for patient-specific function and pain, and all four psychological measures. In contrast, there was no evidence for change in any of the weekly measures for case 3, and only evidence for change in confidence for case 2.

Figure 2. Patient-specific function and pain intensity. Simulation Modeling Analysis for significant change between baseline and treatment period over the course of the intervention (correlation of data stream with phase vector and respective p-values for patient-specific function and pain intensity). Significant values are displayed in bold. Footnote to Figure 2. NRS: Numerical Rating Scale.

Radar graphs in Table 2 provide an overview of each participants’ profile and how they varied over the course of the study, including long-term timepoints of months 6, 9 and 12 post-intervention.

Findings from the long questionnaires and functional tests over Phases A and B indicate that cases 1 and 4 reported greater changes in disability (KOOS), pain psychological factors and functional tests (Table 3) when compared to cases 2 and 3.

Table 3. Patient self-reported outcome measures and functional tests across the three phases of the study.

Measures	Phase	Case 1	Case 2	Case 3	Case 4
Activity limitation (KOOS Daily Living Scale)	A. (week 1)	83.8	48.5	58.8	60.3
	A. (week 10)	80.9	45.6	57.3	57.4
	B. (week 12)	92.6	67.7	80.9	77.9
	Change Score (A – B)	−10.25	−20.65	−22.85	−19.05
	A’. (month 6)	94.1	61.8	79.4	76.5
	A’. (month 9)	94.1	64.7	82.3	77.9
	A’. (month 12)	97.1	75	86.8	79.4
	Change Score (B – A')	2.50	−0.53	1.93	0.03
	MCIC	8–10 points
Pain (KOOS Pain scale)	A. (week 1)	69.4	41.7	58.3	61.1
	A. (week 10)	66.7	36.1	55.5	55.6
	B. (week 12)	80.6	50	69.4	72.2
	Change Score (A – B)	−12.55	−11.1	−12.5	−13.85
	A’. (month 6)	83.3	50.80	66.7	72.2
	A’. (month 9)	86.1	53.90	69.4	72.2
	A’. (month 12)	94.4	49.40	66.7	75
	Change ScoRE (B – A')	7.33	1.37	−1.80	0.93
	MCIC	8–10 points
Pain catastrophising (PCS)	A. (week 1)	8	26	29	11
	A. (week 10)	8	28	27	12
	B. (week 12)	2	25	26	2
	Change score (A – B)	6	2	2	9.5
	A’. (month 6)	1	17	14	3
	A’. (month 9)	0	17	12	3
	A’. (month 12)	1	17	11	2
	Change score (B – A')	−1.33	−8.00	−13.67	0.67
	MCIC	6 points
Knee osteoarthritis fear & beliefs (KOFBeQ)	A. (week 1)	50	75	76	29
	A. (week 10)	51	69	74	31
	B. (week 12)	110	62	74	87
	Change score (A – B)	−59.5	10	1	−57
	A’. (month 6)	107	55	74	85
	A’. (month 9)	109	63	87	79
	A’. (month 12)	109	61	82	89
	Change score (B – A')	−1.67	−2.33	7.00	−2.67
	MCIC	Not available
30s Chair stand test (number of repetitions)	A. (week 1)	7.0	6.3	11.0	11.0
	A. (week 10)	6.5	6.0	10.5	10.5
	B. (week 12)	10.5	8.0	15.5	17.0
	Change Score (A – B)	−3.8	−1.9	−4.8	−6.3
	A’. (month 6)	10	9	15	*
	A’. (month 9)	10	8	15	15
	A’. (month 12)	10	9	16	17
	Change Score (B – A')	−0.50	0.67	−0.17	−1.00
	MCIC	2 − 3 Repetitions
	40 Metre fast-paced walk test (speed in metres/s)	A. (week 1)	1.44	1.18	1.88	1.40
A. (week 10)		1.48	1.61	2.08	1.42
B. (week 12)		1.8	1.7	2.2	20.62
Change score (A – B)		−0.34	−0.31	−0.22	−19.21
A’. (month 6)		1.932	1.739	2.198	*
A’. (month 9)		1.702	1.778	2.235	1.990
A’. (month 12)		2.010	1.810	2.198	1.896
Change score (B – A')		0.08	0.08	0.01	−18.68
MCIC		0.2–0.3m/s

KOOS (score range 0–100. High scores indicate lower knee osteoarthritis -related problems); PCS (score range 0–52. High scores indicate higher pain-related catastrophic thoughts); KOFBeQ (score range 0–99. High scores indicate lower fear and negative beliefs about osteoarthritis). *Due to COVID-19 restrictions, we could not collect data for Case 4 at this timepoint.

A similar pattern is again observed in the long-term at months 6, 9 and 12 post-intervention, with cases 1 and 4 maintaining the improvements in disability (KOOS), pain, psychological factors and functional tests (Table 3). However, cases 2 and 3 maintained the improvements in disability, pain, and functional tests, and they improved further in pain catastrophising and in fear beliefs about knee osteoarthritis.

Findings from integration of qualitative and quantitative data

Overall, the changes observed in the quantitative measures support the qualitative findings, where cases 1 and 4 demonstrated a greater shift in all factors as measured by both weekly short questionnaires, long questionnaires, and functional tests, when compared to cases 2 and 3 (Figure 2 and Appendix Table A2, Table 3).

Specifically, cases 1 and 4 showed a very similar pattern of change. Both cases showed improvement in most quantitative measures in line with their qualitative reports of change in understanding of the condition, improved knee confidence and active engagement in their lives (Figure 3).

Figure 3. Process of change based on the Common Sense Model, for people with knee osteoarthritis undergoing CFT.

In comparison, cases 2 and 3 presented a mixed pattern of change. Cases 2 and 3 showed no improvement in most weekly quantitative measures, no meaningful improvement in pain catastrophising, and little improvement in fear beliefs about knee osteoarthritis after the intervention (Table 3). In contrast, they showed meaningful change in the long questionnaires for disability, pain, pain catastrophising and functional tests, as well as some improvement in fear beliefs (particularly case 3) at long term follow up. The quantitative findings are in line with the experience of these two participants reported by the qualitative findings, where despite reported improvements in function, pain and social interaction, they still reported a lack of confidence in key activities and concern about the long-term consequences to their joint (Figure 3).

Discussion

Best-practice guidelines strongly recommend non-surgical core treatment for osteoarthritis that include education, exercise and lifestyle changes that support self-management [16]. However, societal misperceptions that osteoarthritis is part of ageing, caused by wear and tear and that surgical replacement of the ‘worn out’ joint is inevitable, affect engagement with these recommended non-surgical treatments to manage the condition [41]. This had led to calls for the management of osteoarthritis to consider the individual needs and context of the person with knee pain [15].

This study aimed to understand the process of change and applicability of CFT, a person-centred behavioural approach for the management of knee osteoarthritis in people at risk of poor response to knee replacement surgery. CFT promoted helpful changes in the four participants. Two different responses to the approach were observed which may be explained by differences in individual psychological and social profiles as discussed below.

A shift towards a biopsychosocial conceptualisation of osteoarthritis

A biopsychosocial conceptualisation about knee osteoarthritis accommodates an understanding of joint pathology within a multidimensional framework. This view supports that CFT can help a person reconceptualise their condition, and actively engage in valued life activities with reduced pain, greater confidence, increased physical activity levels and overall healthier lifestyle [10,42]. Across different musculoskeletal [43] and non-musculoskeletal conditions [44], studies suggest a positive relationship between a helpful understanding of the condition and perceived control over the management of the condition [45].

Cases 1 and 4 had a clear shift in their understanding about knee osteoarthritis. This was evidenced by a reconceptualization of pain as a sign of tissue sensitivity caused by multiple factors rather than a sign of damage. These participants adopted a new understanding in which movement and loading of the knee was viewed as necessary to optimise the health of their joint. This conceptualisation was associated with them becoming more active despite some level of pain, increasing confidence about their future, and adopting a view that a knee replacement was not inevitable. Focusing on what one can do despite their condition, rather than what one cannot do because of their condition, is consistent with the International Classification of Functioning, Disability and Health (ICF), established by the WHO to promote a common, evidence-based approach to healthcare. The ICF shifts attention away from (only) curing diseased/damaged body parts; towards empowering people to participate in life situations regardless of their health state [46,47].

This study adds to a growing body of evidence[10,48] that experiential learning (a central component of CFT) to de-threaten pain, enhance pain control and build self-efficacy is an important potential mechanism underpinning positive change for people with disabling musculoskeletal pain [49,50].

A shift towards a mixed conceptualisation of osteoarthritis

Misconceptions about osteoarthritis are common among patients, clinicians, and the public [7–9]. These misconceptions appear deeply rooted in society and commonly reinforced by clinicians and may be perpetuated by dominant ways of thinking about the body and pain (i.e., that pain is a sign of damage and it is unsafe to use painful body parts) [42]. This may potentially undermine a person’s confidence to accept an alternative biopsychosocial narrative [9,42].

By the end of the intervention, cases 2 and 3 shifted aspects of their understanding about their condition. They appeared to adopt a mixed conceptualisation whereby they still perceived that pain was caused by tissue damage, but at the same time, believed that targeting other factors could help avoid or delay knee replacement surgery. While this change allowed them to become more functionally active, their focus on the impairment was still present, which may undermine their confidence about the future of their knees and the need for a joint replacement [42].

Interestingly, follow up data indicates sustained changes in the long-term, with cases 2 and 3 reporting reductions in fear and catastrophic thoughts by month 12. It is possible that continued engagement with their knee without experiencing the feared outcome (i.e., more pain and less function) played a role in shifting their cognitive and emotional responses [45,50,51]. The adoption of positive coping strategies has been demonstrated to mediate a shift in beliefs towards a healthier and less threatening understanding of their condition [52]. This highlights the potential need for additional support/booster sessions beyond the 12 weeks’ intervention.

Why did these four cases respond differently to CFT?

Key differences in the social context, work status, activity levels and mental health profile of these participants may have influenced the different patterns of change. Case 1 was still working and somewhat active, with a good social network of support. While Case 4 reported high levels of social stress, she was still attending group exercise classes, which provided a positive environment for interaction. In contrast, Cases 2 and 3 reported greater psychological and social distress, and less social support.

There is growing evidence that health outcomes are explained less by the quality and availability of medical care than they are by the context of people’s lives [53]. For instance, people with lower health literacy, lower income, no life partner and poorer mental health are more likely to develop osteoarthritis and co-morbid health conditions [54]. Community context [55], social support (e.g., family, caregivers) and participation are key for people with osteoarthritis to maintain a healthy lifestyle [56] and stay engaged with meaningful activities [57]. This highlights the potential need for integrated co-care.

How do these findings inform the management of knee osteoarthritis?

Osteoarthritis has a dominant pervasive ‘impairment’ narrative that is difficult to change, [10,58]. Some people may need more time and/or a more salient experience of safety to shift their conceptualisation about pain as was observed by cases 2 and 3. While CFT may promote a helpful shift, people presenting with more complex psychosocial profiles may require additional co-care to effectively address social determinants of health [59,60]. For instance, co-care with psychologists may provide targeted strategies for management of contextual stress and co-morbid mental health [17,50,61]. Ongoing monitoring of the person’s profile using repeated outcome assessments (e.g., Örebro multidimensional screening [62]) prior to each treatment session and at 3-6 month intervals post-intervention may highlight barriers for recovery, manage flare-ups and trigger appropriate referrals. How this process could be implemented, and its effectiveness are important questions to be addressed by future research.

Nevertheless, for some people a positive and safe experience in which pain is controlled or reduced may not be achievable. Their meaningful activity may always be associated with pain, irrespective of how much other contributing factors are able to be downregulated. In that scenario, reconceptualising the meaning of pain (i.e., accepting that pain may still be present but it is not dangerous to move) and reframing their valued activities (i.e., shifting priorities to align with current functional state), may enable people to participate in their lives despite pain [63].

Design considerations

SCEDs are vulnerable to plausible rival hypotheses that may explain the outcomes such as, maturation, regression to the mean and external factors. Although the authors strategized to limit this problem, by using a stable baseline with ten data points (above the recommended five) [22,64], a therapist log and interview content with the participant and their partner, such possibilities cannot be fully excluded. While the study aimed to include different profiles of people with knee osteoarthritis, some profiles were not included, most evidently, people who are obese. The authors recognise this is a common and important co-morbid issue for people with knee osteoarthritis, and this will be addressed in a separate study. Finally, the authors acknowledge that our backgrounds as researchers with an interest in cognitive and behavioural approaches to the management of musculoskeletal pain and disability, and our clinical backgrounds as physiotherapists caring for people with knee osteoarthritis informed the analysis and interpretations of the data from this study. The presentation of qualitative data serves to render the authors’ interpretations visible to readers.

The strength of the current study lies on the richness of data provided by the SCED. This design enabled an in-depth analysis of the process of change in a challenging group of patients with knee osteoarthritis undergoing a novel individualised behavioural approach.

Conclusions

This study demonstrated the applicability of CFT for the management of knee osteoarthritis. While CFT promoted helpful changes in four people with knee osteoarthritis, two different responses to the approach are documented. One reflected a clear shift to a biopsychosocial conceptualisation of osteoarthritis and its management. The other response reflected a mixed conceptualisation with dissonant beliefs about the condition, rendering them uncertain regarding knee replacement surgery in their future. Individual psychological and social factors emerged as potential barriers to treatment, and have been discussed as considerations for future intervention studies for the management of knee osteoarthritis.

Acknowledgements

The authors would like to acknowledge the participants for devoting their time and efforts during the process of this study.

Disclosure statement

No potential conflict of interest was reported by the author(s). We confirm that the manuscript has been read and approved by all authors and that there are no other persons who satisfied the criteria for authorship but are not listed. We further confirm that the order of authors listed in the manuscript has been approved by all of us. The institution of one or more of the authors (PFC, AS, PO, MMD) has received, during the study period, funding from the Australian National Health and Medical Research Council Centre of Research Excellence in Joint Replacement Surgery (APP1116325)

Data availability statement

Data for the four participants of this study is embedded in the manuscript and individually presented either in a graphic or table format.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Appendix

Table A1. Participants’ characteristics.

Participants	Body mass index (BMI)	Pain and disability (WOMAC)	Mental health (SF-12)	Osteoarthritis grade (KL)	Risk of poor-response (%)
Case 1	24	33	61.1	KL < 4	24.7
Case 2	29.8	77	53.6	KL < 4	19
Case 3	25.9	57	37.8	KL < 4	33.1
Case 4	27.05	57	61.9	KL < 4	23

BMI (body mass index); WOMAC (The Western Ontario and McMaster Universities Osteoarthritis Index – ranges 0–96 – higher scores indicate worse pain, stiffness, and functional limitations) – The WOMAC score was calculated from the Knee Injury and Osteoarthritis Outcome Score (KOOS); SF-12 (lower scores indicate worse mental health); KL (Kelgren-Lawrence radiographic osteoarthritis grading scale – KL < 4 indicates low to moderate joint changes).

Table A2. Pain-related psychological factors.

Simulation modeling analysis for significant change between baseline and treatment period over the course of the intervention (correlation of data stream with phase vector and respective p-values for pain-related psychological factors). Significant values displayed in bold.