Multi-disciplinary interventions for chronic pain involving education: A systematic review

Background There have been growing recommendations to include education in multi-disciplinary interventions targeting chronic pain management. However, effects of this strategy on short- and long-term self-management of chronic pain, remain largely unexplored. Objectives 1. To provide an updated overview of studies that report on the impact of patient education in multi-disciplinary interventions, on self-management of chronic pain; 2. To explore associations between education and chronic pain self-management techniques; and 3. To identify the format and duration of suitable chronic pain interventions targeted at patient self-management. Methods Design: Narrative systematic literature review of randomised or controlled study designs. Data Sources: PubMed, CINAHL, EMBASE, PsycINFO. Participants: Adult patients with chronic pain of any aetiology participating in multi-disciplinary programs that included education. Main outcome measures: Assessments of level of pain, function, quality of life, self-efficacy, self-management, and any other relevant assessments. Study Appraisal and Synthesis Methods: PRISMA guidelines, Cochrane Risk of Bias tool, and TIDieR model. Results Database searching identified 485 potential papers. After removal of duplicates, and irrelevant articles by title and abstract, 120 full-text articles were reviewed and 27 studies were included in this systematic review. Studies were predominantly from the United States (n = 8; 29.6%). Over one hundred outcome measures were identified across all studies, with significant variation also observed in terms of how chronic pain duration was defined, and how education was delivered to participants. Overall, positive benefits of education were reported. Conclusions Education, as part of multi-disciplinary programs, is likely to improve self-management and self-efficacy in people with chronic pain of any aetiology. Heterogeneity in terms of: chronic pain duration; educational resources; healthcare professionals; and outcome measures, were identified as limitations. Further research, in the form of Randomised Controlled Trials addressing these limitations, is recommended.


Conclusions
Education, as part of multi-disciplinary programs, is likely to improve self-management and self-efficacy in people with chronic pain of any aetiology. Heterogeneity in terms of: chronic pain duration; educational resources; healthcare professionals; and outcome measures, were identified as limitations. Further research, in the form of Randomised Controlled Trials addressing these limitations, is recommended.

Background
Chronic pain is pain that lasts or recurs for more than three months [1][2][3]. Globally, prevalence of chronic pain has been reported as 37.3% of the population in developed countries and 41.1% in developing countries [4]. In Australia, chronic pain is the fourth most common chronic condition [1], affecting almost 20% of the population [5]. Chronic pain is complex in terms of aetiology and management approaches [6][7][8], and despite the existence of pharmacological therapy [9,10], remains highly resistant to treatment [5,10]. Chronic pain significantly and negatively impacts upon individual lives, leading to physical disability, mental health problems and long waitlists for specialist health services, as well as economic costs to health services, patients and the community [1,4,[11][12][13][14]. A sustainable solution is needed to reduce these negative impacts and promote effective self-management, with increased recognition of the role of public education around pain and its management, and co-ordinated multi-disciplinary (MD) care [1,9].
Public education promotes patient awareness of the somatic, psychological and social aspects of chronic pain [1,[6][7][8][9], and supports self-management with the aim to improve quality of life [1]. Education alone can have a limited impact on chronic pain [15][16][17][18][19][20][21][22][23][24], yet when combined with MD programs, there seems to be additional benefits for chronic pain management, such as increased patient confidence and self-efficacy [25][26][27]. MD care refers to the collaboration of healthcare professionals from a range of disciplines to deliver comprehensive patient care that meets the needs of the individual [28]. MD programs designed for chronic pain management generally include patient education and other services or care, such as medication reviews, self-management programs, allied health and community-based services, in various delivery modes, such as face-to-face, teleconferencing, and/or internet-based sessions [1,9]. Including education in MD programs has been associated with medication optimization, reduced pain catastrophizing, and reduced utilisation of secondary care services [9,25,26,29]. However, research has highlighted the need for Randomised Controlled Trials (RCTs) to confirm the immediate and sustained impact of these programs on health outcomes [25,26].
There has been limited reporting of empirical evidence of MD interventions; questions remain regarding the efficacy of individual versus group education and the optimal combination of educational topics [1,27,[30][31][32]. Although there are some randomised and/or controlled studies [33-59] and reviews in this area [27,[30][31][32], they have specific limitations. The most detailed systematic review to date was published by Scascighini et al. in 2008 [27]. MD programs, with or without education components, were reported to be more effective than no treatment or standard medical treatment, in participants with chronic, non-specific musculoskeletal pain [27]. The authors concluded that incorporation of once-weekly patient education sessions in MD programs was a superior strategy compared to other medical treatment [27]. A strength of the review was the exclusion of non-RCT studies, however, limitations included: an incomplete reporting of the search strategy; a narrow scope focussing solely on cognitive-behavioural and psychological graded interventions; and limited assessment of methodological quality and risk of bias assessments of included studies [27]. More recent meta-analyses were restricted to small numbers of included papers and the investigation of education as a solitary intervention rather than as part of MD programs [31,32]. Another narrative review around preventive interventions focused on post extremity trauma, that is, one pain type only [30].
Questions remain on how to best incorporate education into MD programs for chronic pain. Such knowledge remains critical in an era when the prevalence of chronic pain is predicted to rise with an ageing population [5,29]. The aim of this systematic literature review was to update and extend Scascighini et al. 's review [27]. This included: 1. Using a well-defined and comprehensive search strategy to identify studies that report the impact of patient education in a broad range of MD interventions; 2. Exploring for associations between education and chronic pain self-management techniques; and 3. Identifying format and duration of suitable chronic pain interventions targeted at self-management from methodological quality and risk of bias assessments. The findings of this review will inform researchers, healthcare professionals and/or people experiencing chronic pain about the current evidence of MD interventions, as well as highlight critical aspects to include in future programs alongside realistic expectations of effectiveness.

Methods
A systematic literature review was performed according to the methods outlined in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [60]. PRISMA guidelines constitute an evidence-based protocol for developing and reporting narrative systematic reviews and meta-analyses [60], and are recommended by the Cochrane community and the EQUATOR network when reporting database searches in systematic reviews [61,62]. Prior to commencing the systematic review, all authors agreed to relevant definitions and a robust search strategy and study selection process (see Supporting information); two attempts at contacting Scascighini et al. [27] for a detailed search strategy were unsuccessful. No protocol exists for this work.

Definitions
For the purpose of this review, the definition of chronic pain was: pain of any aetiology persisting for more than three months [3]. Interventions were deemed MD if they involved collaboration of healthcare professionals from a range of disciplines to deliver comprehensive care to the patient [28]. Education was defined as instructions to inform participants about self-management and/or medication-taking techniques for their chronic pain. The education could be delivered via a range of modalities, including face-to-face, teleconferencing, internet-based sessions and/or use of other multimedia content. Participants included persons with chronic pain who consented to take part in the research, such as those who volunteered, persons who were referred to pain clinics or health centres, individuals who were on sick leave from work, or persons living in retirement communities. Study designs included in this review were RCTs and/ or other randomised or controlled study designs (e.g. cluster randomised trials). treatment approach") and text words (e.g. "self care"). These databases represent the four most appropriate databases that were likely to produce a broad range of peer-reviewed literature across the disciplines of medicine, nursing, psychology and allied health. The search was conducted on five occasions from mid-2017 to August 2019, with the search strategy consistently adopted across all databases to ensure identification of eligible studies. The final list of studies was exported to a referencing management system (EndNote 1 ) in August 2019.

Eligibility criteria and study selection
The following inclusion and exclusion criteria were established (Table 1).
Eligibility for inclusion/exclusion was first assessed via independent duplicate manual screening of article titles and abstracts (S.J. and S.M.). Disagreements were resolved by discussion and consensus with a third author (M.K.). The same process was undertaken when reviewing all relevant full-text articles. A supplementary search method, involving a review of reference lists of included full-text articles (i.e. snowballing), was adopted.

Data extraction
Information was extracted into a Microsoft Excel 1 spreadsheet and included characteristics such as country, sample size, number of chronic pain types and/or sites, and members of the MD team. The outcome measures were noted as total number of outcome measures used, those that showed significant results and whether such measures were validated. Information extracted into a separate Microsoft Word 1 document included more detailed description of the individual interventions including type and delivery mode, as well as study setting, patient details and age range, findings and follow-up. The objectives, randomisation and blinding procedures, as well as group allocation details, were also noted.

Risk of bias and quality of reporting
Risk of bias and quality of reporting assessments were undertaken by the first author (S.J.), following concurrent independent testing (S.J. and M.K.) of the first six included articles against a checklist developed from the Cochrane Collaboration Risk of Bias tool [63][64][65] and the Template for Intervention Description and Replication (TIDieR) [66,67]. Where articles lacked information, the respective study protocols were consulted. These steps ensured the completeness of reporting and helped determine the reliability and replicability of included studies [63][64][65][66][67]. Any disagreements were resolved through discussion between all authors.

Data analysis
Given the heterogeneity of study methods and outcome measures across all included studies as well as the broad nature of the review, statistical comparisons between studies were not possible. Hence, a narrative description of the data was adopted as a suitable method to meet the review aims [68]. Other narrative studies published in evidence-based journals and databases such as PLoSONE [69] and Cochrane [61], are known to have used the same methodology.
The authors systematically analysed all reported outcomes across all included studies. Thematic analysis was performed from all outcomes and only those findings that aligned with the objectives of the work, have been reported in this paper.

Search results
The search strategy identified 485 records of which 69 duplicates were removed. An additional 297 records were excluded through title and abstract screening, with a total of 120 full-text articles retrieved for further evaluation. Twenty seven studies met the inclusion criteria (see Fig 1) [33-59]. A concise overview of data from these included studies is provided in Tables 2 and 3.

Characteristics of identified randomised and/or controlled studies
Findings from the 27 included studies that aligned with the objectives of this work are summarised below.  [40,43,47,56], with two studies each providing tapes [40,47] and CDs [43,56], and one study providing videos [56]. Three studies used emails as an additional educational tool [41,  Participant check-ins or monitoring periods between active intervention sessions varied widely, ranging from at least daily checks [49] to outcome checks every two months [41]. These checks all preceded actual follow-up periods. Three of the 27 studies did not involve follow-up periods [41,42,46]. Of the remaining 24 studies, the maximum duration of follow-up � Various terminology as used to specify pain type across all studies (e.g. "chronic widespread" and "chronic non-malignant" to describe widespread pain) ο One study with each of the following professionals: Acupuncturist, Massage Therapist, Nutritionist, Social Worker and Surgeon § Numbers do not add up to 27 as studies had more than one type of health professional Of the 27 studies included, 11 did not specify which healthcare professionals were involved. One mentioned an "activities director", one mentioned "statistician", one  Chronic and disabling musculoskeletal pain pain score � 4 and Roland Morris Disability score � 7 (n = 241).

Outcomes
Intervention outcomes and effectiveness were generally evaluated immediately after intervention conclusion, followed by re-evaluation at follow-up. The majority of studies (n = 23) showed statistically significant results (p<0.05 and/or large effect size) for one or more outcome measures [33-40, 42-47, 49-54, 56, 58, 59].

Risk of bias and quality of reporting
Risk of bias (Table 4 and Fig 2)  and a mix of computer-generated randomisation, sealed envelopes and blocked randomisation [56]. The three remaining studies involved biased randomisation in pain centres that could not be paired [43], did not specify how randomisation was achieved [47], or adopted a nonrandomised design [59].
Reporting bias and other bias were generally low across all 27 studies. Based on the information outlined in study protocols, three studies showed selective outcome reporting [33,35,55]. Other than a potential conflict of interest in one study [54] and design issues in one nonrandomised study [59], no additional sources of bias were identified.
Generally, the quality of reporting was satisfactory when assessed against the TIDieR checklist [66,67] (Fig 3). On average, eight items were reported across all 27 included studies. One study [56] reported all 12 checklist items, while another [35] reported all 11 checklist items as it did not involve any modifications during the course of study. All included studies described

Summary of findings
Overall, education as part of MD interventions, either as the active intervention or the comparator, seemed to show positive outcomes for participants. These positive outcomes generally appeared to be sustained or improved at follow-up. Topics around physical activity and cognitive-behavioural strategies were most common. The majority of included studies offered takehome materials in the form of written supplements and homework to help participants re-visit content at their own convenience. Interventions were conducted as one-on-one and/or group weekly didactic sessions, targeting any type of chronic pain. MD teams across all included studies consisted mostly of physiotherapists, psychologists and nurses. However, variable risks of bias and heterogeneity in outcome measures were observed across all 27 included studies.

Discussion
This systematic review identified that education, as part of MD interventions, appeared to contribute to an immediate amelioration in chronic pain management, with benefits seemingly sustained or improved long-term. This was despite the observed heterogeneity in: participant type and duration of chronic pain; study methods; outcome measures; and quality of reporting. Commonalities identified across the included studies provide important insights to all relevant stakeholders about potentially successful strategies for future adoption. Low risk of bias is shaded green; High risk of bias is shaded red; Unclear risk of bias is shaded yellow https://doi.org/10.1371/journal.pone.0223306.t004 Review of education in multi-disciplinary interventions for chronic pain

Common aspects across included studies and deductions
Most education sessions were run in groups, for a minimum of two hours each, over a four-to ten-week period. They incorporated direct, face-to-face interactions between educators and participants, and provided take-home materials and homework activities. These characteristics all seemed to optimise group outcomes and brought positive changes to participants' lifestyles and quality of life over time, a finding similar to that reported in a review of cancer pain management [75]. Duration of education sessions also aligned with previous recommendations of effective instruction time of 2.5 hours for low back pain [76].
In terms of skill sets of recruited healthcare professionals, physiotherapists, psychologists and nurses were the most common intervention facilitators, potentially indicative of a need to focus on education around physical activity, mental health (notably, cognitive behavioural strategies) and general lifestyle changes. Communication between participants and these three types of healthcare professionals is likely to reshape beliefs and behaviour of people with chronic pain, an outcome identified in previous research involving various healthcare professionals and health conditions such as low back pain and palliative care [28, [77][78][79][80].

Directions for future research
Moving forward, researchers should incorporate all fore-mentioned education characteristics in future RCTs of MD programs. Additionally, future RCTs should: expand on the MD team; use only validated outcome measures; extend follow-up periods to investigate the longitudinal effects on participants and health systems; and analyse the effectiveness of chronic pain management strategies in developing and non-developing countries.
1. Expanding the MD team. Almost half of the included studies (n = 11) did not report on the full composition of their MD teams, thereby creating a lack of transparency around the exact range and diversity of professionals involved. Chronic pain, by its very nature, is a complex medical condition that could require a broad array of healthcare professionals for optimal management. Research on the impact of bigger and more diverse MD teams than those used historically are recommended. The immediate and long-term impact of such teams on participants (e.g. the inclusion of GPs, pain specialists, pharmacists, psychologists, physiotherapists, nurses, nutritionists, occupational therapists and social workers), can only be hypothesised at this stage, as highlighted by previous preliminary studies [25,26]. Involvement of pharmacists in only two interventions was surprising and might indicate that these professionals are overlooked in chronic pain management, despite them being strategically placed in primary care to monitor, advise, support and refer consumers with chronic pain [9]. Future RCTs exploring the role of pharmacists in MD chronic pain management, especially in terms of medication management and counselling, are warranted to further investigate the positive contributions of these professionals [25,26].
A number of other healthcare professionals were overlooked in MD chronic pain management; occupational therapists, nutritionists and social workers were involved in only one study each. These healthcare professionals have an important role in providing assistance to patients with chronic pain. Occupational therapists can assist patients to reach optimal health and well-being through participation in everyday activities [81]. Nutritionists can educate patients on the impact of food in reducing chronic inflammation, a symptom associated with chronic pain, as well as assist with optimal weight management [82]. Social workers can establish, monitor and improve practice and ethical standards within MD teams, as well as advise patients on relevant and available services, such as subsidised healthcare plans [83]. Future RCTs investigating MD teams that also involve these healthcare professionals, are warranted.
2. Using only validated outcome measures. Instead of combined use of validated and non-validated outcome measures (as observed across most studies in this review), it is recommended that only evidence-based measures, such as the Pain Self-Efficacy Questionnaire (PSEQ), Assessment of Quality of Life-6D (AQoL-6D) and Patient Global Impression of Change (PGIC), be utilised in future RCTs. These represent reliable and consistent ways of assessing participants' ability to physically and emotionally cope with and manage pain over time [74,[84][85][86][87][88][89][90]. Given that VAS, RMDQ and SF-36 detected significant changes [33-35, 37, 42, 43, 49, 51-54] in this review, these tools are recommended for similar interventions in the future. Consistent use of only validated outcome measures will facilitate comparison of outcomes across studies. Only then, will other systematic frameworks, such as Grading of Recommendations Assessment, Development and Evaluation (GRADE) [91], be used effectively to make clinical practice recommendations.
3. Extending follow-up periods to assess longitudinal effects. This review has only reported positive, short-term (less than two years) lifestyle benefits of education, as part of MD programs, in individuals experiencing multiple types of chronic pain. Yet, longitudinal lifestyle, behavioural changes and economic impact (e.g. in the form of savings incurred to participants, health systems and the economy) over multiple years, remain other aspects to be investigated. The authors propose an extension of follow-up periods to analyse these effects.

Conducting RCTs in developing and non-developing countries.
Despite a high proportion of the population in developing countries reporting chronic pain [4,92], this review has mostly considered studies from developed nations (n = 25). Local traditions, cultural background and/or local policies tend to vary among developed, developing and non-developing countries, and can significantly influence diagnosis and treatment of chronic pain in these areas [10,92]. With an inevitably ageing population and a predictable rise in prevalence of chronic pain globally [5,29], investigating the provision of cost-effective MD programs that incorporate education in developing and non-developing countries is another avenue for future research. Only then, can chronic pain be truly and effectively managed on a global scale.

Strengths and limitations
The initial aim of this review was to update the work of Scascighini et al. [27]. Despite multiple attempts to contact the authors, a detailed search strategy of the 2008 review was not obtained. Consequently, an alternative up-to-date search strategy, aimed at better identifying relevant articles, was developed. While the total number of studies included in this work (n = 27) was not dissimilar to the 35 and 31 studies analysed by Scascighini et al. and Berube et al. respectively [27,30], this work set out with a broader focus of exploring the effects of any type of education, delivered as part of MD interventions, on any chronic pain type. Inclusion of all possible forms of education was a strength of this review and addressed the authors' original objective of reporting on the ideal combination of education methods to optimise self-management of chronic pain.
Another strength was the use of several databases to identify relevant studies, a strategy previously identified as broadening the scope of research and improving the quality of systematic reviews [93]. Setting language limits to English, French, Dutch and German, rather than English alone, added to this strength by capturing a wider range of articles and increasing the generalisability of conclusions. Comprehensive assessments of methodological quality and risk of bias, as well as analysis of longitudinal effectiveness in relation to follow-up periods [94], further consolidated conclusions of this review. These strategies were seldom reported in earlier work [27,[30][31][32].
Conversely, heterogeneity of included studies was a limitation as it made application of the Cochrane Collaboration Risk of Bias tool [63][64][65] and the TIDieR model [66,67] difficult. The comparative process proved especially challenging due to the diversity in individual studies in terms of: 1. inclusion criteria for participants, especially in terms of duration of their pain; 2. education strategies adopted; 3. the array of validated and non-validated or customised assessment tools utilised to monitor participant progress; and 4. the variety of healthcare professionals involved. Additional inherent limitations of the TIDieR tool also imply potential uncertainty in the determination of quality of intervention reporting and reproducibility of specified methods [95]. As stipulated by Cotterill et al. (2018), TIDieR limits its quality and reproducibility assessments only to the context or setting specified by the individual study analysed, thereby failing to provide an overview of quality and reproducibility of similar studies in different contexts or settings over time [95].
Another limitation was restriction of the authors' search strategy to only four databases and four languages, which implied that other relevant studies may have been missed. No grey literature was searched and it is likely that studies identifying negative effects of education may not have been reported.

Gaps in existing literature
Interpretation of results from this review calls for caution as variable risks of bias, uncertainties around quality, and lack of consistency (e.g. in terms of MD teams and outcome measures), were reported as shortcomings of included studies. There is an increased need for greater homogeneity in research methodology, such as use of intention-to-treat, adequate concealment and consistent use of only validated outcome measures. Detailed reporting of interventions to ease reproducibility, for example in terms of: type of professional recruited; type of materials used; and intervention adherence measures, is also required. Clarity around participant characteristics such as exact duration of chronic pain, will further increase reliability and generalisability of any findings promoting self-management and self-efficacy in people with this condition. Lastly, given the high proportion of studies from developed nations, conclusions from this review may not be generalised to people with chronic pain in developing and non-developing nations.

Conclusion
Participant education remains an understudied aspect of MD interventions focused on chronic pain management. Compared to previous work, this review has attempted to present reliable and comprehensive evidence around the effects of including education in such programs. MD programs that include four-to ten-week education around topics of physical activity, cognitive behavioural strategies and general lifestyle, generally point towards significantly positive results, particularly in relation to self-management practices and self-efficacy. Such programs should be adopted as part of a patient's pain management plan in the future. However, some research gaps remain, including a lack of consistency in reporting participant chronic pain duration; use of standardised and validated outcome measures in combination with non-validated measures; and lack of homogeneity in research methodology. A need for more research that expands on the MD team, uses only recommended outcome measures, extends follow-up periods and is implemented in various developing and non-developing countries, is warranted.