Clinical recommendations for chronic musculoskeletal pain in South African primary health care

Background Chronic musculoskeletal pain (CMSP) is prevalent globally and places a significant burden on individuals, healthcare systems and economies. Contextually appropriate clinical practice guidelines (CPGs) on CMSP are advocated to translate evidence into practice. Aim This study aimed to investigate the applicability and feasibility of evidence-based CPG recommendations for adults with CMSP in the primary health care (PHC) sector of South Africa (SA). Setting The PHC sector in South Africa (SA). Methods Consensus methodology was used, comprising two online Delphi rounds and a consensus meeting. A multidisciplinary panel of local healthcare professionals involved in CMSP management was purposefully sampled and invited to participate. The first Delphi survey considered 43 recommendations. In the consensus meeting, the results of the first Delphi round were discussed. The second Delphi round reconsidered the recommendations with no consensus. Results Seventeen experts participated in the first Delphi round, 13 in the consensus meeting and 14 in the second Delphi round. In Delphi round two, 40 recommendations were endorsed, three were not endorsed and an additional recommendation was added. Conclusion A multidisciplinary panel endorsed 41 multimodal clinical recommendations as applicable and feasible for the PHC of adults with CMSP, in SA. Although certain recommendations were endorsed, they may not be readily implementable in SA because of context factors. Contribution The study forms the basis of a model of care for contextually relevant PHC of CMSP. Future research should explore factors that could influence the uptake of the recommendations into practice to optimise chronic pain care in SA.


Introduction
Chronic pain and disability are a global and local South African (SA) concern. The pooled prevalence of chronic pain in developing countries is 18% (95% confidence interval: 10% -29%), (range 13% and 51%), 1 with South Africans having a similar prevalence of 20% in women and 18% in men. 2 In addition, one in three South Africans over the age of 65 years suffers from chronic pain. 2 Chronic pain is defined as persistent pain on most days, lasting longer than 3 months and persists beyond the expected healing time. Chronic musculoskeletal pain (CMSP) comprises pain associated with bone, joints, muscles and related soft tissue. 3,4 The World Health Organization (WHO) in collaboration with International Association for the Study of Pain (IASP) recently endorsed the International Classification of Diseases, 11th edition (ICD-11), which includes a classification of chronic pain. The classification comprises chronic primary pain, chronic cancer pain, chronic post-surgical pain, chronic neuropathic pain, chronic headache and orofacial pain, chronic visceral pain and CMSP. 4 The classification also acknowledges pain mechanisms within chronic pain classification as nociplastic, nociceptive and neuropathic pain. 3 management options, are essential for the holistic management of CMSP. While multiple therapies for CMSP are available and indicated, there exists uncertainty in the applicability of management options when considering different healthcare contexts. Clinical practice guidelines (CPGs) are particularly valuable where uncertainties about management options exist and where scientific evidence for an intervention is scarce. 6 Clinical practice guidelines for CMSP have the potential to optimise chronic pain management and health systems outcomes, which is important goals in a country such as SA, with a transforming healthcare system. South Africa is an upper-middle-income country, 7 which must address the complex needs of people living in uppermiddle-or lower-income settings in nine different provinces. Healthcare is provided through public and private healthcare systems, and there are many inequalities in healthcare access and resources between and within these systems. 8 Vulnerable and remote communities are particularly affected by these disparities. As part of healthcare transformation, the SA government developed the National Health Insurance (NHI), which aims to provide equitable healthcare access to all South Africans. The NHI of South Africa White Paper (DoH) 9 proposes CPGs to guide the delivery of evidence-based and cost-effective health services. The development of a new CPG can be resource and skill-intensive, and therefore existing good-quality CPGs can be adopted, adapted or contextualised to be implemented. The use of alternative approaches to CPG development enables resources to be channelled towards CPG implementation. Furthermore, CPGs from high-income countries may not be appropriate in resource-constrained environments because of differences in policy and legislation, organisational context, health system resources, human resources and capacity, access to healthcare, infrastructure, and patient resources and needs. 10 Contextually relevant CPGs may play an important role in the uptake of evidence-based practice as part of healthcare restructuring. However, according to a recent landscape analysis of CPGs for the SA context, no CPG for CMSP management in SA exists. 11 The existence of good quality CPGs on CMSP, and the range of contextual factors that may influence pain management in SA, indicated the need for alternative CPG development. The implementation of a contextually relevant CPG for CMSP that is acceptable and feasible for the SA context has the potential to support local decision-making and person-centred care in primary health care (PHC) because the majority of chronic pain conditions are managed in PHC. 12,13 This study therefore focusses on addressing the knowledge gap on the acceptability and feasibility of existing CPG recommendations for implementation in the local, South African PHC context. The existing CPG recommendations were extracted from high-quality CPGs. 14, 15 The information on acceptability and feasibility would provide information on which of the clinical recommendations could be adopted, and which would need to be adapted or contextualised. 10 Additionally, the approach would provide contextual information for the modification and implementation of recommendations. Adoption of recommendations implies accepting the recommendation without change, adaptation means changing the recommendation to suit local needs and contextualisation refers to adding context points to the recommendation to optimise implementation. 10 This study evaluates the applicability and feasibility of internationally developed evidence-based clinical recommendations and endorses them for the management of adults with CMSP in South Africa primary health care (SA PHC). The secondary objectives were to identify possible contextual factors that may influence the uptake of the recommendations and customise the recommendations for implementation in the intended context based on the contextual information.

Study design
A priori consensus methodology involving a group of local experts was used, 16 comprising two Delphi study rounds and a face-to-face consensus meeting. The Delphi process aims to structure group communication processes efficiently to develop a systematic consensus of opinion regarding specific circumstances. 16,17 Consensus methodologies are often used in decision-making regarding the appropriateness of clinical recommendations. 16,18 The consensus meeting allowed interactive discussion about key issues, including the implementation of recommendations. The study formed part of a broader project to develop tailored guidance for the PHC of CMSP in the SA context. The project phases comprised an analysis of context, a systematic review of CPGs, evidence synthesis, contextualisation (Delphi study) and an external review.

Study population and sampling
The study population comprised healthcare professionals involved in chronic pain management in the SA context. A multidisciplinary group of healthcare professionals was identified through purposive sampling and invited to participate. Panellists participating in consensus processes are typically individuals who have practical experience and interest in the topic being investigated. 16 The key sampling criteria considered the diversity of opinions and experiences to define the expert panel as follows: • Experience, expertise and interest in CMSP management and/or • Work experience in any of the different healthcare settings or sectors (public PHC, private practice, academic institutions, and professional organisations); and/or • Different clinical disciplines (medical doctors, clinical nurse practitioners, pharmacologists, physiotherapists, occupational therapists, psychologists, managers, researchers); and/or • Expertise in CPG writing and use; and • At least 3 years' experience in the field.
As it is important to include patients' views in guideline development, patient views were considered in a separate http://www.phcfm.org Open Access research phase, as part of implementation strategies for the recommendations. 19 Potential participants were identified by the authors through contributions in conferences focussing on pain management and public health, publications and public healthcare sector involvement. These potential participants were then requested to suggest other potential participants who may fit the inclusion criteria. 18 Twenty-six potential participants were identified and invited to participate via email and/or by personal discussion with the principal investigator. The purpose of the study, the process of developing a clinical recommendations, the consensus process, informed consent and the conflict of interest were explained. The aim was to recruit between 10 and 18 participants, as this number is recommended for consensus generation and group dynamics. 20

Development of evidence-based recommendations prior to consensus study
Prior to conducting the consensus study, we conducted a systematic review to identify high-quality CPGs for PHC of CMSP 14 and we extracted and synthesised 43 clinical recommendations from them. 15 Clinical practice guidelines were classified as high quality when they achieved a median score of 50% or more for the domain rigour of development on the Appraisal of Guidelines Research and Evaluation, Version II (AGREE II). 21 A summary of the systematic review findings is available in Table 1. The consensus approach described in this study utilised the list of 43 extracted recommendations to ascertain consensus about the applicability and feasibility of each recommendation for the SA PHC context.

Data collection
Participants completed an online Delphi survey via Stellenbosch University surveys (SUNsurvey), using an emailed survey link. The survey provided a short introduction about the study, a request for informed consent and instructions on how to complete the survey. The participants were presented with 43 recommendations, accompanied by the strength of the body of evidence (SoBE) for each recommendation. A five-point Likert scale (strongly agree [1], partly agree [2], undecided [3], partly disagree [4] and strongly disagree [5]) was used for voting. A non-applicable (n/a) key, which was considered a missing data point, was available if a statement was outside the scope of expertise of the participant. The survey requested participants to consider the applicability and feasibility of the clinical recommendation to endorse the recommendation for the SA context. Applicability or feasibility was defined as the ability to put a recommendation into practice when considering the barriers and facilitators to implementation. Applicability or feasibility considers the population in the intended setting, as well as the knowledge, skill, staff, time frames, equipment and resources that influence applicability. 22 The participants were able to add a comment per recommendation to elaborate on their vote. The survey was open for 2 weeks. Two reminders were sent before the closing date of the survey. Prior to the consensus meeting, all responses were downloaded from SUNsurvey, de-identified and summarised by the primary investigator (PI) and analysed by the research team.
Two weeks after the survey closed, a face-to-face consensus meeting was held. The purpose was to present the results of the first Delphi round and to discuss recommendations for which consensus was not reached after the Delphi round (i.e. recommendations for which the Likert responses were 3 or above; or no consensus). The meeting was audiotaped, and a summary of discussions and decisions was made by the PI. After the meeting, the research team revised the wording of the recommendations to ensure that they incorporated the panel comments and were grammatically correct.
The second-round online survey was administered 1 week after the consensus meeting. Participants could vote for the adapted or contextualised recommendations using the same Likert scale. The survey was open for 3 weeks. Two reminders were sent before the closing date of the survey.
The steps leading to the list of recommendations are summarised as follows: • Step 1: Conduct a systematic review to identify high-quality CPGs on the PHC of CMSP 14 • Step 2: Extract and synthesise a core list of recommendations for PHC of CMSP from existing high-quality CPGs 15 • Step 3: Panel evaluate and endorse or reject each recommendation for its applicability and feasibility in the SA PHC context (Delphi round 1)

Data management and analysis
The data were extracted from the online survey into an Excel spreadsheet. An explicit method for the combination of results for the two Delphi rounds was used. 16 The median was used as a measure of central tendency to represent the combined opinion of the participants. 16, 23 The interquartile range (IQR) was used to indicate the level of dispersion from the median. 17 Consensus could indicate the combined agreement or disagreement with a recommendation. Three criteria for consensus were applied. Firstly, an IQR of 1 or less was defined as consensus as this is a suitable consensus indicator for a five-unit scale. 17 Secondly, the study aimed to reach a consensus for 95% of items before terminating rounds (i.e. 39 out of the 43 recommendations). It was estimated that two to three rounds would be sufficient because round one started with predetermined evidence statements, instead of generating ideas (as for a classic Delphi study). 20 Each comment that participants added within the survey or consensus discussion remained linked to its recommendation and was analysed narratively. All recommendations on pharmacological management were externally reviewed by a pharmacologist, as we were unable to recruit a pharmacologist on the expert panel.

Quality assurance measures
For the Delphi rounds, voting was independent and confidential. The panel members signed a conflict-of-interest document to declare any actual or potential financial, professional affiliation and intellectual conflicts of interest that may have a direct influence on the content of the recommendations. The researchers were committed to neutrality in the process. In the consensus meeting, the PI facilitated the discussion, and the research team provided input to the rest of the group. This process is inherent to a consensus meeting (group decision-making). The researchers' engagement in the consensus meeting was grounded in the evidence statements and making sure that each member contributed equally. Only D.E. and Q.L. were involved in the data analysis and the rest of the research team contributed to the interpretation of the results. The researchers also completed a conflict of interest declaration. An independent research assistant (a social anthropologist) attended the consensus meeting and audited the data.

Ethical considerations
The study protocol was approved by the Stellenbosch University Health Research Ethics Committee (HREC), SA. Electronic (tick box) informed consent was obtained on the online Delphi questionnaire and written informed consent and conflict of interest declarations was obtained at the consensus meeting. The survey enabled confidential responses and anonymous extraction of data. Because of the nature of the in-person consensus meeting, participants were aware of each other's participation but were requested to uphold the confidentiality of the information shared. Participants were aware that they could exit the study at any point in time. The recommendations for the Conducting and REporting of DElphi Studies (CREDES) 18 were used for reporting.

Results
The researchers invited 26 potential participants and 17 of them participated in the Delphi round one, 14 participated in the consensus meeting and 13 in the second Delphi round.
The reasons for non-participation were time constraints and geographical distance (for the consensus meeting). Table 2 provides an overview of the participants' characteristics.
Physiotherapists were the largest professional group at the consensus meeting. Thirteen of the round one participants were from the Western Cape province.
Out of the original core set of 43 recommendations, 36 were endorsed (a median score of 1 or 2 and an IQR of 1 or less) during Delphi round one and seven were not endorsed. Table 3 presents the recommendations for which consensus was achieved in the first Delphi round, which indicated recommendations that could be adopted for the SA context.    We recommend referral to a pain management specialist when there is failure to achieve treatment goals; chronic pain is poorly controlled; there is significant distress and/or where interventional management or assessment is considered Good SoBE SIGN 1 1 Self-management Self-management We recommend self-management strategies and resources to be provided with other therapies in the treatment of patients with chronic pain to ensure active patient participation during early management as well as part of a long-term management  Table 4 shows the original and reformulated recommendations with no consensus or undecided ratings at the end of Delphi round one, and the results after discussion at the panel meeting and re-voting on these statements in the second Delphi round. Table 4 therefore indicates recommendations that needed to be adapted or contextualised to be acceptable or feasible for the SA context. However, two recommendations received no consensus, and one was undecided, after contextualisation.
The seven recommendations that were not endorsed in the first Delphi round and therefore had uncertain applicability and feasibility to the SA PHC sector comprised six recommendations about pharmacological management for the prescription of topical non-steroidal anti-inflammatory drugs (NSAIDs), topical rubefacients and antidepressants for chronic pain conditions. Low-level laser therapy (LLLT) was the only non-pharmacological intervention with uncertain applicability and feasibility for the SA context. Although the recommendations for topical NSAIDs and rubefacients were adapted, they were not endorsed in Delphi round two because of their limited availability and the risk of side effects. The recommendations for tricyclic antidepressants for the management of chronic low back pain (LBP) were amended by the panel.
After two Delphi rounds, the panel reached a consensus on 41 recommendations, of which a recommendation on pain neuroscience education was added and endorsed by the panel. All recommendations, endorsed in round two, as well as six in round one, were endorsed as 'partly agree' indicating that uncertainty about these remain, motivating the need for further investigation. Table 5 provides a summary of the contextual determinants listed by participants listed in the online Delphi rounds.

Discussion
This study provides information on the acceptability and feasibility of evidence-based CPG recommendations for the multimodal care of CMSP in SA PHC settings. Interdisciplinary care and the involvement of nurses, medical practitioners, physiotherapists, occupational therapists, mental health practitioners, social workers and patient participation are emphasised to achieve therapeutic goals ( Figure 1). We followed a structured and rigorous decisionmaking process via the participation of a panel of local experts who considered the evidence and context-specific circumstances for the applicability and feasibility of the recommendations. The study's findings may be useful for decision-making regarding CMSP, when considering PHC interventions in the context of the NHI in SA. The endorsed recommendations indicate the aspects that may be prioritised to address in policy and practice to address the burden of CMSP in this context; and they provide clear direction to clinicians regarding interventions for CMSP and their underpinning body of evidence. The participants identified several contextual factors ( Table 5) that may impact the implementability of the recommendations in practice. These contextual factors included organisation of health services at the primary care level, human resource capacity, practice patterns, access to care, intervention determinants, availability of equipment and patient-specific factors.
Although certain recommendations were endorsed as applicable and feasible for the SA context, they may not be readily implementable without a change in contextual conditions, particularly in remote or rural communities. The implementation of the endorsed recommendations will require strategic prioritisation and a multisectoral implementation plan.
The recommendations support a person-centred approach to care and collaborative decision-making with the patient and the multidisciplinary team. The main outcome of the collaboration is the formation of a therapeutic alliance between the healthcare provider and the patient. 35 With the given foundation, the healthcare provider can identify the patient's priority concerns about CMSP, while considering the individual patient's cultural beliefs, previous experience, values and beliefs regarding CMSP, to guide management and assessment decisions and to optimise the impact of the proposed interventions. 35 In SA, with its diverse culture and language, the focus on the patient's beliefs regarding pain is important, because the meaning and expression of pain is influenced by culture. 36 Despite the importance of patient-centred care, the panel raised some concerns regarding human resource constraints, limited access to healthcare providers and limited available consultation time in PHC. The given contextual impediments are confirmed by existing literature 8,37 and pose a threat to the approach to care as well as the implementation of several assessment and management options.
The multidimensional impact of CMSP requires holistic assessment. The panel agreed that the value of diagnostic imaging in chronic pain conditions should be carefully considered. This notion is supported by findings from a systematic review that negative diagnostic tests do not necessarily decrease the patient's fear or concern about pain. 38 However, investigations should be requested if requirements are met 39 (i.e. serious pathology suspected, unsatisfactory response to care and unexplained progression of clinical picture) and when treatment decisions may alter based on the results. There still appears to be some uncertainty (albeit consensus) regarding the use of a mechanism-based classification of pain as nociceptive, neuropathic or nociplastic, as advocated in the ICD-11. 3 The uptake of the classification approach in practice can be optimised using validated outcome measures, and it has several advantages for communication between healthcare providers. 3 The panel agreed that validated outcome measures are valuable for assessing the impact of pain, as well as the quality of life and functional ability. The applicability of outcome measures may be influenced by their availability in different languages spoken in the different regions of SA, as well as their appropriateness for lifestyles, values, socio-economic factors and cultural beliefs in the intended target group. There is thus an impetus to develop a set of outcome measures that is contextually relevant for the local PHC context to assess the emotional, psychological, functional occupational and social impact of CMSP. Because the experience of chronic pain is influenced by social and cultural factors. 36 Consistent with a multimodal approach to CMSP care, the panel recognised the importance of physical therapy, psychological therapy and socio-environmental support as a core first-line management for CMSP. Multidisciplinary management is important because of the close relationship between chronic pain, physical activity, psychological distress and the social determinants of health. 3,5,40 A multidisciplinary healthcare professional should therefore be involved in CMSP care to strengthen the person-centred, biopsychosocial approach. While these recommendations are supported by evidence, the panel emphasised that access to rehabilitation and mental health professionals is constrained in the SA PHC context, when compared with high-income health systems, because of their relative scarcity. 8,37,41 Mental health conditions and the reported lack of access to mental healthcare are an international concern. In addition, physical inactivity is a risk factor for the development of common mental disorders, emphasising the need for innovative PHC, primary health care   solutions to address the interaction between chronic pain, physical activity and mental well-being. 5 Therefore, there is an impetus to support SA PHC healthcare systems and organisations to enable interdisciplinary and intersectoral collaboration. 8 The scarcity of multidisciplinary pain management programmes and pain management specialists in the SA context emphasises the need for interdisciplinary collaboration in the PHC context to achieve therapeutic goals.
Patient empowerment for self-management within a supportive environment is a key outcome of CMSP care. Empowerment towards self-management should commence early in the patient management pathway. The various recommendations focussing on patient education support this notion. Chronic musculoskeletal pain, because of its persistent nature, requires ongoing and supported selfmanagement. 42 Patient education can lead to increased understanding of the condition, enhanced healthcare literacy, behaviour change and adherence. 43,44 However, education needs to be contextually relevant and take culture and health literacy into account aid patients' understanding of CMSP as emphasised by the panel by published literature. 5,36 Patient and family education requires a time investment, which was listed as an important need. Empowering and supportive environments can be achieved via community-based support programmes, to maintain gains made in the formal therapeutic environment, 43 through peer support and public involvement. 42 In the SA context, the Western Cape on Wellness (WoW) 45 programme is an example of such a community-based support initiative.
Several factors influenced the endorsement of recommendations related to the pharmacological management of CMSP. The prescription of analgesics should be based on the step-wise approach and within the guidance of the SA Essential Medicines List for PHC (EML PHC). 46 The endorsement for the use of opioids for moderate to severe CMSP came with several caveat recommendations, comprising careful patient selection, risk assessment and informed consent. The longterm use of opioids remains controversial, and the WHO released a statement cautioning on its use and advising further research in this area. 47 The cautions regarding opioid prescription are supported by research because of associated side effects and the quality of the body of evidence on the topic. A Cochrane review 48 found positive effects for pain and function in chronic LBP; however, the studies were of short duration, with limited data on long-term effects, and moderate to very low quality of evidence. Another Cochrane review 49 concluded that there was not enough evidence to support the long-term use of tramadol alone, or tramadol in combination with paracetamol, for pain or functional improvements associated with osteoarthritis. These authors also reported a high incidence of adverse effects attributed to tramadol (nausea, dizziness and tiredness), which resulted in high attrition rates in the studies. 49 For this study, it is acknowledged that most of the prescribers on the panel were based in public healthcare with opioid access limited to tramadol and morphine. There was a lack of representation of prescribers from the private healthcare sector, who have access to a wider variety of opioids for prescription. Therefore, the use of opioids in the management of CMSP should be carefully considered, based on the risk of addiction, risk of adverse effects and questionable long-term efficacy.
The recommendations regarding the anti-depressant drugs amitriptyline, fluoxetine and duloxetine were the focus of discussion during the consensus meeting in the study because of uncertainties that existed about them. The use of antidepressant medication in CMSP is an important consideration because depression is a common co-morbidity with CMSP 5 and depression is one of the leading causes of years lived with disability in sub-Saharan Africa. 32 Adaptations were made on the phrasing of these recommendations, based on information from the EML PHC 31 and the burden of disease. For example, the recommendation on tricyclic antidepressants for the management of chronic LBP was contextualised based on the value judgement of the panel and considering the burden of disease in SA. In addition, Su et al. 50 found that antidepressant therapy can improve outcomes of chronic pain even in the absence of clinical depression. It is acknowledged that, at the time of the study, the EML PHC 31 was used, while the EML PHC 46 is currently available. In summary, practical and policy implementation factors influenced participants' choices for recommendations on pharmacological management, namely medication inclusion in the EML and availability in the PHC sector.
Certain recommendations were not endorsed because of concerns about their availability in the intended context, cost, safety and efficacy. The use of topical NSAIDs was not endorsed based on the risks associated with using oral and topical NSAIDs concurrently. A systematic review of the efficacy of topical NSAIDs concluded that they can be effective in the relief of osteoarthritis pain. 33 It is therefore acknowledged that this recommendation might have benefitted from another Delphi round in which stable scoring might have been reached. 17 The panel were undecided as to whether LLLT is applicable in the intended context in view of its limited availability, high cost and training required. The availability and access to electrotherapy equipment may have played a role in the decision-making, as acknowledged by the panel. There is also limited evidence to support or refute the effectiveness of LLLT for the treatment of LBP (subacute or chronic). 51 The heterogeneity of the populations,

Socio-environmental concerns
We suggest that the clinician identify social and environmental stressors influencing the patient with chronic pain and refer the patient as appropriate (e.g. to a social worker). 40,52 Occupational rehabilitation We suggest that clinicians consider referring patients with chronic musculoskeletal pain for vocational assessment and rehabilitation to facilitate return to work and to reduce work-related disability. 53 Community support programme We suggest that the clinician consider referring the patient to a community support group, to enhance the focus on supported self-management. 43 interventions and comparison groups in the LLLT review provided insufficient data to draw firm conclusions. These findings indicate the need for higher-quality randomised controlled trials on the topic.
Based on the findings of the study, a patient pathway was synthesised to offer a visual overview of the recommendations and different decision-making points in a patient's journey through the PHC sector. The discussion from the consensus meeting indicated that for the holistic management of people with chronic pain, additional interventions need to be considered for the local context, which had not been listed in the original CPGs. These additional recommendations form part of the patient pathway and are summarised in Table 6.

Strengths and limitations of the study
The strength of this study was the systematic method that was followed to evaluate and endorse the recommendations. Through the electronic Delphi survey, the multidisciplinary panel confidentially provided their opinions about the applicability and feasibility of the recommendations for the intended context. We believe that the panellists used their scientific and contextual knowledge, and clinical experience integrated with their own belief system to make decisions. 18 The endorsement of the CPG recommendations and the identification of context factors highlighted uniquely SA circumstances and pragmatic considerations. The panel provided information on factors that may influence the translation of the recommendations in practice. A thorough analysis of the context factors and the development of a multilevel implementation plan is recommended to enhance the uptake of recommendations into practice. A future study will provide an in-depth analysis of contextual factors that may influence the implementation of these endorsed recommendations in the SA PHC context. Patients and potential end-users also provided input regarding the recommendations, 19 and the results of these inputs will be available in future publications.
The following limitations need to be considered when interpreting the findings of this study. Despite the delay in the publication of information, an updated search identified only one CPG that could potentially have been included in the synthesis process. The study did not include an analysis of the level of stability over Delphi rounds. 17 A stability analysis was not conducted because of the focus on the two consensus criteria (the IQR and the number of rounds) as well as pragmatic considerations (time and resources). We opted to conduct a consensus meeting to provide an opportunity for discussion of potentially challenging recommendations. The consensus meeting was held in-person, which constrained participation. An online meeting was not considered, as the meeting took place in the pre-coronavirus disease 2019 (COVID-19) era when online meetings were not the norm. The selection of panel members may have influenced consensus. 16 Despite a range of professions being invited to participate, the panel for the consensus meeting was skewed towards women and rehabilitation professionals. Only half of the invited participants participated in the second round, which may have led to participation bias. A major shortcoming was that we were unable to recruit a pharmacologist to the panel. However, a pharmacologist externally reviewed the recommendations to aid in quality assurance. There was limited participation of private sector practitioners. Most participants were from the urban areas of the Western Cape, and there was a limited representation of rural practitioners. The findings can therefore only be generalised to the Western Cape setting. Further research is required regarding the feasibility of the recommendations in rural contexts. We therefore recommend alternative recruitment strategies for similar studies in future to ensure geographical representativeness and representation of diverse professionals in the expert panel. 18 We did not distinguish between the concepts of applicability versus feasibility, which could have yielded useful information. Furthermore, we acknowledge that while consensus was the focus, the utility of the study findings is dependent on the characteristics of the people who participated in the consensus process. Consensus does not necessarily imply the correct judgement but simply that there is agreement among the panellists. 18 Patients were not directly part of this process, although their perspectives were considered. 19 Patients therefore provided input on the contextual factors that should be considered for the recommendations; however, they did not have the opportunity to vote. It is acknowledged that the study took place prior to the COVID-19 pandemic and that the healthcare system changes and updates to the EML 46 are not reflected in this study.

Conclusion
A set of evidence-based recommendations for the PHC of CMSP pain in SA was endorsed by a multidisciplinary panel of local experts. This study summarises these recommendations for the assessment and management of CMSP, comprising pharmacological and non-pharmacological management to enhance a multidisciplinary approach. Future research should explore the influence of other context-specific factors within health systems and communities, which were reported as adjunct recommendations in this study as they could influence the uptake of the recommendations into practice.