Treatment interventions and assessment of fear-avoidance for chronic musculoskeletal pain: a systematic review protocol

Rationale

The Fear-Avoidance (FA) model was developed by Lethem et al in 1983 (26) to explain the pathways by which acute pain transitions to chronic using a biopsychosocial perspective (1, 2). According to the FA model, a self-reinforcing cycle of catastrophising, fear, hypervigilance, and further avoidance of movement can lead towards dysfunction, deconditioning, depression, and increased pain (3). The FA model is made up of four main components: kinesiophobia, catastrophising, pain-related fear, and FA beliefs (4, 5). Some researchers have suggested adding victimisation, disability, self-efficacy, and input from the interpersonal social environment, suggesting that the model should extend beyond pain-related fear and highlights the significance of concurrent, and often competing, goals (2, 6-8) (Fig 1).

• Download figure
• Open in new tab

Fig 1: The components of the Fear Avoidance model and potential implications.

In the assessment and treatment of chronic musculoskeletal pain, kinesiophobia may be described and understood by a range of synonyms (e.g., “fear of movement”, “fear of (re)injury”, “avoidance behaviour”, “avoidance hypervigilance”, “behavioural performance”, “pain-related fear with impaired physical performance”, and “avoidance of activity”) which tend to dominate the research landscape (9-11). Kinesiophobia is an important mediator of new and prolonged chronic pain and disability, and as such is an important element of the FA paradigm; rates are estimated at 51-72% of people with chronic pain (9, 12). Kinesiophobia presents a major challenge to successful physical rehabilitation; unchecked, it may contribute to the development of disuse syndrome, which is associated with additional physical and psychosocial impacts such as depression, muscle atrophy, and medication misuse (13). These fundamental effects risk drawing practitioners focus away from less-reported aspects of the FA model.

Several systematic reviews have focused on the results of interventions to treat kinesiophobia (14-17); a focus which risks overlooking the multifactorial origins and drivers of chronic musculoskeletal pain. Bordeleau et al (9) recommends further research into the multi-faceted nature of kinesiophobia, encompassing physical, psychological, and social interventions. These concerns are further compounded by the nature and use of tools available to measure kinesiophobia; methods such as the Tampa Scale of Kinesiophobia (TSK) that is widespread and accessible (18). Liu et al conducted a review of related measures and found that The Fear-Avoidance Beliefs Questionnaire (FABQ) and TSK were the most often used questionnaires (19,20). Indeed, many systematic reviews have specified use of the TSK as part of their study inclusion criteria (9, 15). This further exacerbates the focus on kinesiophobia, and simultaneously dampens potential for assessment of the full FA model on chronic musculoskeletal pain. Moreover, current reviews have primarily included randomised controlled trials (RCT) in their search strategy, which restricts the diversity of interventions explored.

In addition to the TSK and FABQ, there are other FA model-adjacent measures including the Kinesiophobia Causes Scale (KCS) (21), Athlete Fear-avoidance Questionnaire (AFAQ) (5), and the Fear-Avoidance Components Scale (FACS) (7). Some researchers have started to use a combination of the aforementioned measures, but it is unknown whether this is appropriate given the potential for discordance between each measure and the FA model construct (9, 22-24).

The current literature on assessing FA and its components, beyond just kinesiophobia, is lacking. There is a need for a systematic review that includes intervention studies assessing and/or treating FA by appreciating its components on chronic musculoskeletal pain, as it is more patient-tailored and clinically useful. As such, utilising this study’s interpretation of the FA model will influence the search criteria to include all relevant study designs and the current measures of FA components. Finally, the voice of community collaborators, such as patients, carers, and allied health professionals is vital in ensuring patient-centred conclusions and recommendations for assessment and interventions of FA in musculoskeletal health. Therefore, the objectives of this systematic review are as follows:

1. Identify and characterise interventions used to treat FA in people with chronic musculoskeletal pain by considering all study designs, relevant measures, and FA synonyms;

2. evaluate the use of common FA model-related measures that assess treatment interventions in people with chronic musculoskeletal pain;

3. engage with relevant community collaborators in the development of the systematic review search methodology and, at completion, to provide reflections on the findings and ensure further research to be clinically relevant.

Patient and Public Involvement Statement (PPI)

A panel of patients with chronic pain who attended the Royal National Orthopaedic Hospital, UK attended an online focus group about chronic pain and were reimbursed for their time. The nature of this PPI is an involvement activity, where the opinions of community collaborators were collected on the topic of FA, and so ethical approval is not required. Regarding kinesiophobia, patients spoke about the importance of early intervention, education, and rehabilitation from encouraging, knowledgeable, and positive health professionals. Within the discussion, participants mentioned other aspects of fear avoidance, such as catastrophisation, and were concerned that their fear of movement would never fully disappear but felt hopeful that it could be dampened with correct management. Participants all agreed that peer support was vital to their personal ‘journey’ in managing fear of movement. Moreover, a concern was raised about how a participant felt that they were unable to take part in research studies due to their comorbidities, and so consequently believed that their views and experiences were not reflected in the research.

The authors felt that the term ‘kinesiophobia’ was quite uncommon among allied health professionals, musculoskeletal researchers, and patients, and that research may exist on this topic without naming it in such a way, and consequently be missed by systematic review searches. To combat this, the aim was to reach multiple health professionals to consider their understanding of kinesiophobia so that they could offer the terminology they use with other clinicians or their patients. The authors circulated the following question on Xⁱ (formerly Twitter),

Calling ALL health professionals! What do you mean when you say *kinesiophobia*? Please comment below + include a description of your job role + share with your network.

This resulted in 20,001 impressions, 727 engagements, 23 retweets and 37 comments. Direct responses came from 36 individuals, representing physiotherapists, musculoskeletal researchers, orthopaedic surgeons, pain management clinicians, chiropractors, and sports trainers. Examples of responses include,

“…being scared to move too much for fear of triggering a flare up of pain.” “…rationale of irrational fear of causing or increasing tissue damage…”

“when an individual doesn’t have the knowledge, skills and/or confidence to move their body…”

“I don’t tend to use the word clinically however, in children+young[sic] people I would describe kinesiophobia as the loss of spontaneous, fluent and playful movement due to a memory or thought that links movement to a negative and feared consequence…”

“…no patient has ever used the term unless learnt from a clinician.”

“…I don’t like when it’s describe[sic] as “irrational” fear it makes me feel like we’re blaming the person.”

“…delineating terms may help us work more effectively with patients…”

These interactions subsequently informed the design of our search terms and inclusion/exclusion criteria.

In addition to this protocol being informed by PPI, results of this systematic review will be shared with another focus group including patients, carers, and health professionals to discuss their impression of the findings and potential for impact on them, to gauge how well their experience is represented by the findings and generate ideas for future research.

Eligibility criteria

Eligible studies include primary interventional research articles in English or with an available English translation. The FA model was initially outlined by Lethem et al (26) and studies that explored psychological factors in chronic pain were increasingly prevalent in the 1970s, so the date range for included reports will be from 1970 onwards.

Information sources

Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, Embase, The Allied and Complementary Medicine Database (AMED), PEDro, Web of Science, and grey literature (e.g., medRxiv, ISRCTN protocol registry) will be searched using database-specific syntax detailed below. Reference lists of resulting articles will be searched for articles previously missed. The final list of selected articles will be entered into Connected Papersⁱⁱ, which will also search for associated papers that could have been missed.

Search strategy

The search strategy will be formulated using the Peer Review of Electronic Search Strategies (PRESS) Evidence-Based Checklist (27). The searches will be divided into three themes: [1] musculoskeletal conditions, [2] FA components and their synonyms, and [3] relevant outcome measures. These themes will be combined with the Boolean operator “AND”. Any restrictions outlined in the exclusion criteria will be combined cautiously, as to not remove relevant articles, with “NOT”.

Ovidⁱⁱⁱ will be used for the AMED, Embase, and MEDLINE databases whilst CENTRAL and PEDro will be searched on their own respective search engines. Key search terms will be used for every database and logically matched with their respective subjecting headings or Medical Subject Headings (MeSH) term system. For instance, Ovid has a “Map Term to Subject Heading” option, whilst CENTRAL uses MeSH terms. The relevant subject headings are database-specific and will be exploded where necessary. If a subject heading system is unavailable, then a multi-purpose (mp) search would be used for the free-text key terms, i.e. several fields including title, abstract, and keywords (27).

The key search terms for the search strategy are:

1. Search 1 (musculoskeletal conditions): “pain”, any MSK “-algia” conditions (e.g. causalgia, neuralgia, epicondyalgia, etc.), MSK pain, physical suffering, MSK pain (any body part, e.g. elbow, knee, neck, back, etc.), “phantom limb”, “complex regional pain syndrome”, “sciatica”, “neuropath*”, arthritis (all forms, e.g. osteoarthritis, rheumatoid arthritis, etc.), “sacroilitis”, “Failed back surgery syndrome”, “persistent spinal pain syndrome”, “whiplash”

2. Search 2 (FA components): kinesiophobi*, “fear of mov*”, “fear of injur*”, “fear of reinjur*”, “fear of re-injur*”, “fear” AND “movement”, “fear” AND “injur*”, “fear” AND “reinjur*”, “fear” AND “re-injur*”, “fear avoidance”, “behavioural performance”, “behavioral performance”, “avoidance behaviour”, “avoidance behavior”, “pain-related fear”, “avoidance of physical exertion”, “avoidance hypervigilance”, “impaired physical performance”, “scared of movement”, “mov* anxiety”, “avoidance learning”, “catatastrophi?*”, “maladaptive pain belief”, “pain belief”, “victimi?ation”

3. Search 3 (relevant outcome measures): “ Tampa Scale of Kinesiophobia”, “Kinesiophobia Causes Scale”, “Fear-Avoidance Beliefs Questionnaire”, “Fear-Avoidance Components Scale”, “Athlete Fear-Avoidance Questionnaire”, “Pain Anxiety Symptoms Scale”, “Pain Catastrophi?ing Scale”, “Patient-Reported Outcomes Measurement Information System Pain Interference”

4. Search 4: Search 1 AND Search 2 AND Search 3

5. Search 5: Search 4 NOT exclusion criteria (e.g. postoperative populations)

Table 1 shows an example search of the MEDLINE database using the Ovid platform:

Study records

Data items

Study identifiers, study design, cohort size, cohort location, intervention, participant gender, participant age, cohort diagnosis/disorder, cohort symptom characteristics, cohort comorbid diagnoses, treatment period (e.g., 6-months), outcome measure (e.g., TSK), treatment outcome, comparative group (where applicable), power (where applicable and achievable), effect size (where applicable and achievable), level of evidence for therapeutic studies (28).

Outcomes and prioritisation

The primary outcome will be a list of the identified and characterised treatment interventions used to treat FA components and the questionnaires used to assess outcome. Secondary outcomes will be the opinion of community collaborators through PPI to provide reflections and develop a list of future research priorities.

Risk of bias of individual studies

For RCTs, the Risk of Bias (ROB) 2.0 tool (29) will be used to assess the risk of bias. For non-randomised studies, the Risk of Bias In Non-randomised Studies – of Interventions (ROBINS-I) tool (30) will be used. This tool considers that each study is attempting to mimic an RCT and judges them on similarly thorough criteria in sevens domains that could affect bias. Analysis of bias will be carried out for all included studies twice between two reviewers (ST, FW) independently. Where disagreements occur, discussion led by a mediator (AT) will aim to resolve differences and result in agreement. Both tools utilise relevant cut-off points that categorises the risk of bias as either “low risk of bias”, “some concerns”, or “high risk of bias”.

Data synthesis

The identification and extraction of core themes from included papers will be performed using a combined reflexive thematic analysis and realist synthesis approach (32-35). Reflexive thematic analyses and associated thematic networks are established qualitative analytical tools designed to reveal and organise themes in a hierarchical manner through a structured approach. The themes generated during reflexive thematic analysis will be tested and interpreted by key community collaborators in the PPI groups; these interpretations will be integrated with the reflexive thematic analysis data through and adapted realist synthesis methodology. Realist synthesis has been applied to systematic reviews to uncover the mechanisms underpinning interventions positioned within contextual factors. This combination of established qualitative approaches will enable the robust integration of the systematic review and PPI reflexive data to better appreciate multiple perspectives on interventions.

Meta-biases

Tabulation and visual display of the individual risk of bias assessments will be created. Due to the assumed heterogenous nature of the gathered studies, it will not be feasible to conduct a meta-analysis.

Confidence in cumulative evidence

The Grading of Recommendations Assessment, Development and Evaluation (GRADE) method will be used to assess the quality of evidence for all outcomes (35). This includes risk of bias, consistency, directness, precision, and publication bias. This will result in allocation of a high, moderate, low, or very low-quality study label.