Effectiveness of patient education as a stand-alone intervention for patients with chronic widespread pain and fibromyalgia: a systematic review and meta-analysis of randomized trials

Objective Patient education is recommended as an integral component of the therapeutic plan for the management of chronic widespread pain (CWP) and fibromyalgia (FM). The key purpose of patient education is to increase the patient’s competence to manage his or her own health requirements, encouraging self-management and a return to desired everyday activities and lifestyle. The aim of this systematic review was to evaluate the evidence for the benefits and potential harms associated with the use of patient education as a stand-alone intervention for individuals with CWP and FM through randomized controlled trials (RCTs). Method On 24 November 2021 a systematic search of PubMed, MEDLINE, Embase, CENTRAL, PsycINFO, CINAHL, ClinicalTrials.gov, American College of Rheumatology, European League Against Rheumatism, and the World Health Organization International Clinical Trials Registry Platform identified 2069 studies. After full-text screening, five RCT studies were found to be eligible for the qualitative evidence synthesis. Results Patient education as a stand-alone intervention presented an improvement in patients’ global assessment (standardized mean difference 0.79, 95% confidence interval 0.13 to 1.46). When comparing patient education with usual care, no intervention, or waiting list, no differences were found for functioning, level of pain, emotional distress in regard to anxiety and depression, or pain cognition. Conclusion This review reveals the need for RCTs investigating patient education as a stand-alone intervention for patients with FM, measuring outcomes such as disease acceptance, health-related quality of life, enhancement of patients’ knowledge of pain, pain coping skills, and evaluation of prioritized learning outcomes.

Chronic widespread pain (CWP) is prevalent in the background population, with an estimated prevalence of about 10%, and represents a major clinical challenge owing to the complexity of the disorder (1,2).CWP can be used as a stand-alone diagnosis (3), but is also considered the defining feature of fibromyalgia (FM), with diagnostic criteria further requiring the presence of other symptoms, including fatigue, non-refreshing sleep, and cognitive impairment (4,5).Patients with FM are the best described subgroup of patients presenting with CWP, and are believed to represent the upper end of a pain severity spectrum associated with greater disease burden and higher levels of disability than CWP (6,7).FM has an estimated prevalence of about 2% in the background population; the complexity of this disorder also results in major clinical challenges (8).
Patients with CWP and FM are best managed through a multidisciplinary effort, in which patient education plays an essential role based on a biopsychosocial approach (9,10).In the biopsychosocial model, chronic pain is viewed as an interaction among biological, psychological, and social-contextual factors (11).The primary focus is on the person experiencing pain instead of the individual symptoms and pathophysiology, and also on maintaining and improving functional ability and social integration rather than symptom reduction (11).In this context, patient education applied as an intervention strategy is well accepted.The key purpose is to increase the patient's competences to manage his or her own health requirements (12).Thus, the core component of pain education is to encourage patients to be self-sufficient, develop active coping skills, and pursue as normal a life pattern as possible (10,11,13).Nonpharmacological intervention strategies, with patient education, encouraging self-management, and return to desired everyday activities and lifestyle, are therefore considered first line intervention in evidence-supported stepped-care therapeutic models for the management of CWP and FM (8,10,14,15).
The aim of this systematic review was to evaluate the evidence for the benefits and potential harms associated with the use of patient education as a stand-alone intervention for individuals with CWP and FM.Our objective was to compare the effectiveness of patient educational efforts, relative to management as usual (control comparator) in patients with CWP and FM, with functioning defined at the level of activity and participation as the primary outcome among the major outcomes collected.

Method
This study was reported in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) statement and checklist (Online Supplementary File: PRISMA checklist) (16).Our protocol (Online Supplementary File: The protocol) was registered in PROSPERO (identifier: CRD42021282840) and followed the PRISMA-P guidelines (17,18).The protocol was submitted to PROSPERO (2021-11-08) before conducting the search.

Eligibility criteria
The PICO framework was used for each search strategy, with PICO defined as population, intervention, comparators, and outcome (17,18).Study design.Randomized controlled trials (RCTs) examining the benefits and/or harms of patient education provided as a stand-alone intervention, communicated by health professionals, for patients with CWP and FM, were included.The trial reports had to be peer reviewed and published as full text, thereby excluding trial reports available only as unpublished or non-peer-reviewed manuscripts and conference abstracts.No restrictions were imposed regarding age, sex, trial duration, setting, or publication date.No limitations were made on applied language; thus, foreign-language papers were translated as far as possible.
Condition or domain being studied.In this study, the 1990 American College of Rheumatology (ACR) definition of CWP was applied (19).FM patients were defined according to the 1990 ACR classification criteria, i.e. presenting with CWP and a minimum of 11/18 tender points, or fulfilment of the ACR 2016 diagnostic criteria for FM (19,20).
Patient education.The intervention, patient education, had to have a biopsychosocial approach, as described in the introductory section: it had to be communicated by health professionals (i.e.not patients), with monodisciplinary or interdisciplinary delivery (e.g.medical doctors, nurses, physiotherapists, occupational therapists, psychologists), individually or group based, through lectures and/or group discussions.Patient education delivered using telemedicine was included.
Comparator.Usual care, no intervention, and waiting list were considered eligible as control comparators.However, when using usual care as a comparator, both groups had to receive an equal amount of standard care (e.g. by their primary physician).Comparisons not allowing for a clear distinction in effect between intervention and comparator (e.g.sham, or other psychological interventions) were not considered eligible.
Outcome.Trials were included independently of the outcome measures reported.

Information sources
The search was conducted on 24 November 2021 through a systematic search of PubMed, MEDLINE (via Ovid), Embase (via Ovid), CENTRAL (via Cochrane Library), PsycINFO (via EBSCO), CINAHL (via EBSCO), Clin-icalTrials.gov,ACR, European League Against Rheumatism (EULAR), and the World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP).Literature saturation was evaluated by further hand searching of references and relevant reviews.

Search strategy
The search strategy was developed by PD, MBP, and RC, with assistance from EEW and KA.The search strategies for all databases was conducted using the PICO framework, as recommended by the Cochrane Collaboration Guidelines/Handbook and the Updated Method Guidelines for Cochrane Musculoskeletal Group Systematic Reviews and Meta-analyses (21,22).The search (Online Supplementary File: Search strategy) was revised and assisted by a research librarian at Copenhagen University Library not otherwise associated with the project.

Selection process
Duplicates were removed using the Mendeley reference management software (version 1.19.4).Two reviewers (PD, KA) performed the screening of title/abstract and full-text assessment using the Covidence online tool.Disagreements were resolved by discussion or consultation with a third reviewer (EEW/RC).Covidence and Mendeley were used to manage citations.The PRISMA flow diagram was used to summarize trial selection process.All trials eligible for synthesis according to the PICO framework were included in the qualitative synthesis; if there were sufficient data, these eligible trials could also be encompassed in the quantitative synthesis (i.e. the meta-analysis).

Data collection process and data items
Data from eligible trials and for all outcomes of interest were extracted into standardized forms through Covidence by two independent reviewers (PD, KA).The extracted data were baseline characteristics of the studies, study methodology, demographic information, and detailed information on the intervention.
Our primary outcome for benefits was functioning, defined at the level of activity and participation [e.g.measured by the Fibromyalgia Impact Questionnaire (FIQ) function subscale].Secondary outcomes were health-related quality of life (HRQoL) [e.g.measured by the EuroQol 5 Dimensions (EQ-5D) questionnaire], level of pain (e.g.FIQ item 'level of pain'), emotional distress (anxiety and depression) [e.g.measured by the Brief Pain Inventory (BPI) or FIQ items], patient ratings of global assessment (e.g.measured by FIQ total score), and pain cognition related to coping skills, including catastrophizing, self-efficacy, pain-related fear, and acceptance [e.g.measured by the Pain Catastrophizing Scale (PCS)].For each outcome domain, any measures available from the included trials were accepted.
Outcomes for harms were adverse events (AEs) based on total number of withdrawals, withdrawals due to AEs, and the number of patients with serious adverse events (SAEs), including deaths.The collected outcomes were based on the Core Outcome Sets, endorsed by the Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials (IMMPACT) and Outcome Measures in Rheumatology (OMERACT) recommendations (23)(24)(25).
Data for outcomes were collected at baseline, postintervention, and at 6 months' follow-up.Trials reporting data only as final scores were converted to change scores, following the methods for conversion described in the Cochrane Handbook (22).Disagreements were resolved through discussion or with a third reviewer (EEW/RC).Authors were contacted to obtain unpublished data.

Study risk of bias assessment
Two authors (PD, KA) independently assessed randomization, blinding, adequacy of analyses (i.e.intentionto-treat approach), and the risk of selective outcome reporting, using the Cochrane Collaboration's risk of bias tool (RoB 1.0) (26).Disagreements were resolved through discussion.Each risk of bias domain was rated as low, high, or unclear risk of bias, according to the definitions summarized in the Cochrane Collaboration's risk of bias instrument (26).Finally, each RCT was assigned an overall risk of bias in terms of low risk (low for all key domains), high risk (high for at least one key domain), and unclear risk (unclear for at least one key domain).

Effect measures and synthesis methods
Continuous outcomes were expressed as standardized mean differences (SMDs) with 95% confidence intervals (95% CIs), calculated by dividing the mean difference by the standard deviation for the given outcome (22,27).Continuous outcome data were preferably extracted based on change scores along with the corresponding dispersion.Binary outcome data were converted to SMDs (28).Six-month follow-up data were used as the preferable timepoint for the evidence synthesis, when possible; otherwise, data were the post-intervention measure that was available.Relative risks (RRs) were the preferred measure for the harm measures, collected as binary data, and were applied to the control group risks, generating absolute risks (29).
Random effects meta-analysis was used by default to combine the trial results (30); restricted maximum likelihood (REML)-based meta-analyses were used to combine the trials (31).Evaluation of heterogeneity was based on the inconsistency index (I 2 ), measuring the percentage of variation across studies due to heterogeneity and not due to chance (32,33).Values of I 2 were interpreted accordingly: values < 30% as low heterogeneity, > 75% as high heterogeneity, and values in between as moderate heterogeneity (22,32).Statistical analyses were conducted using R software version 3.6.2(R Core Team, 2020).

Certainty assessment
The Grading of Recommendations Assessment, Development and Evaluation (GRADE) system was used to rate the overall certainty of the evidence (34,35).Evidence from RCTs starts at high quality, with subsequent rating down in cases of serious study limitations (risk of bias), the degree of inconsistency, indirectness, imprecision, and the apparent risk of publication bias (34,35).The GRADE ratings of very low-, low-, moderate-, or high-quality evidence reflect the extent to which the effect estimates are credible (34,35).The quality of the evidence across the included trials was independently assessed by PD and KA, and disagreements were resolved at a consensus meeting with RC.

Study selection
We identified 2069 studies by searching across the databases and found 56 potentially eligible studies (Figure 1).A 6% conflict between reviewers was revealed after screening the titles and abstracts (Online Supplementary File: Excluded studies during full text assessment along with reason for exclusion).After full-text screening, five studies were found to be eligible for inclusion; four in English (36)(37)(38)(39) and one in Spanish (40).The study by Bosch Romero et al (40) was included only in the qualitative synthesis as they reported insufficient data to be included in the quantitative synthesis.

Trial characteristics
The five eligible studies included 397 participants, all diagnosed with FM.The majority of participants were women (96%), with study proportions ranging from 94.2% to 100%.Study characteristics are presented in Table 1 lectures, because of the internet-based intervention (37).The content in the studies by Barrenengoa-Cuadra et al and Saracoglu et al was pain neuroscience education (PNE): explaining neurobiology and neurophysiology (38,39).Soares and Grossi included education and discussions about the body, FM, pain, sleep hygiene, stress, medication, managing crises, ergonomic education, and self-management (36).Williams et al taught participants about FM, and behavioural and cognitive skills to help with symptom management (37).Of the two authors (36,39) contacted for additional unpublished data, one responded (39).

Effect of patient education: benefits
An overview of the results of meta-analyses together with the judgement of certainty in the evidence is presented in Table 2.The meta-analysis of the primary outcome, functioning, included three trials (36)(37)(38) and showed no significant difference between patient education and the control group (Figure 2).The only significant difference was found regarding the outcome global assessment (Figure 3).The meta-analysis indicated a significant difference favouring patient education.The trial by Williams et al (37) used a binary outcome measurement with the Patient Global Impression of Change (PGIC) to measure patients' global assessment; otherwise, the outcomes were continuous.For the remaining outcomes of benefits, no significant difference between patient education and the control group was indicated.High heterogeneity was present in all of the meta-analyses (see Figures, Online Supplementary

Effect of patient education: harms
The results pertaining to AEs only include the total number of withdrawals, since there were no withdrawals due to AEs and no patients with SAEs or deaths reported in any of the included studies.The meta-analysis on the total number of withdrawals showed no significant differences between patient education and the control group (see Figure, Online Supplementary File: Forest plots, Figure S7).

Risk of bias within studies
Risk of bias assessments for each trial are shown in Figure 4.Only two trials described an adequate randomization process (37,38).There was a high risk of detection in all of the trials.Three trials had a high risk of reporting bias.Saracoglu et al neglected to report the Numeric Pain Rating Scale (NPRS) for pain intensity post-intervention and at follow-up (39).Soares and Grossi reported data on the Pain Questionnaire (PQ), the Symptom Checklist-90 -Revised (SCL-90-R) and the Interview Schedule for Social Interaction (ISSI), but neglected to report the results (36).Soares and Grossi also neglected to report follow-up data on the waiting list group (36).Bosch Romero et al neglected to report standard deviations, 95% CIs, or p-values for all outcome measures (40).None of the trials described concomitant conditions or treatments, giving them an unclear risk of other biases.‡ Indirectness includes differences between study population, intervention, and outcomes.§ Evaluating whether the 95% CI includes both benefit and harms.|| Publication bias could not be assessed owing to the small pool of published studies.CI, confidence interval; GRADE, Grading of Recommendations Assessment, Development and Evaluation; HRQoL, health-related quality of life; RCT, randomized controlled trial; SMD, standardized mean difference.(39) was assessed post-intervention, and for Williams et al (37) and Barrenengoa-Cuadra et al (38) at 6 months' follow-up.Only Williams et al (37) used the Patient Global Impression of Change (PGIC); the other studies used the Fibromyalgia Impact Questionnaire (FIQ).The treatment effect (TE) is reported as the standardized mean difference (SMD) with corresponding standard error (SE), calculated from binary data from the study by Williams et al (37), and continuous data from the studies by Soares and Grossi (36), Barrenengoa-Cuadra et al (38), and Saracoglu et al (39).CI, confidence interval.

Study
All included trials had a considerable risk of bias, leaving the overall risk of bias as high for all studies.

Certainty of evidence
The certainty of evidence for each outcome is presented in Table 2.Only AEs with the total number of withdrawals obtained a moderate certainty of evidence.Global assessment was evaluated as low, whereas the rest of the outcomes had a very low certainty of evidence.All were rated down because of high risk of bias (Figure 4), high heterogeneity, and extremely wide 95% CIs, and because they were all small studies.Prespecified subgroup analyses were not performed, owing to the small number of trials in the meta-analysis.

Discussion
Patient education as a stand-alone intervention given to patients with FM presented an improvement in global assessment with a low certainty of evidence.No differences were found between patient education and usual care, no intervention, or waiting list, regarding functioning, HRQoL, level of pain, emotional distress in regard to anxiety and depression, or pain cognition.For outcomes of harm, no differences were found in the total number of withdrawals.
Guidelines for the management of CWP and FM recommend stepped-care intervention models with patient education as an important first line intervention component (8,10,14,15).Still, evidence supporting these recommendations is scarce.We found that patient education was unable to influence the patients' level of symptoms and self-reported functional ability.Patient education as an intervention is, however, targeted at producing the changes in knowledge, attitudes, and skills necessary to maintain or improve health and encourage self-management (10,14,15,41).Thus, patient education becomes a means to reach certain goals and may not necessarily lead to optimal medical health, including a reduction in patients' symptom burden or perceived level of functioning.The effectiveness of patient education might be better judged on outcomes such as disease acceptance and evaluation of learning outcomes, including enhancement of patients' knowledge on pain and self-care strategies.The recommended overall goal in treating patients with FM is improving HRQoL (8).None of the included trials in this study presented data on the participants' HRQoL.
The results from this review are in agreement with previous systematic reviews by Elizagaray-García et al and García-Ríos et al (42,43).They concluded that patient education provided as a stand-alone intervention had no effect on pain, quality of life or functional ability.The review by Elizagaray-García et al briefly describes the search strategy, leaving the question of whether the search was broad enough (42).None of the reviews performed meta-analyses (42,43).The reviews included two and four RCTs on patient education, respectively (42,43).Both reviews included the RCT from Van Oosterwijck et al, which concluded the opposite of the review authors, i.e. that education about pain physiology (i.e.PNE) was useful in the treatment of FM (44).PNE consists of educational sessions describing the neurobiology and neurophysiology of chronic pain and pain processing, with particular focus on the role of the central nervous system and deemphasizing tissue-based explanations, i.e. the concept that pain means damage.It promotes patients' understanding of chronic pain and may help them to reverse unhelpful and often fearful ways of thinking about pain, which may discourage rehabilitation (39,41,45).A review by Saracoglu et al (46) found that PNE may improve functional status, pain-related symptoms, anxiety, and depression in patients with FM.Saracoglu et al evaluated the effect of adding PNE to a multimodal treatment, which included exercise (46).In contrast to our review, the comparator group also received the multimodal treatment but without the PNE (46).Suso-Martí et al ( 45) also investigated the effect of PNE for patients with FM.They allowed the comparator to be non-active interventions or minimal interventions besides no intervention (45).They found no effect on anxiety and pain catastrophizing, in accordance with our review, and low-quality evidence that PNE can decrease the impact of FM and pain intensity (45).
There is still limited research evaluating the most effective use and delivery of patient education, including PNE.The educational intervention in the trials included in this review had very different durations and dissimilar content (36)(37)(38)(39)(40). Also, the guidelines on managing patients with FM do not include recommendations on the delivery method, length, or duration (10).Our review included an internet-based RCT (37).Despite there being no contact with the participants, they significantly improved their pain intensity and physical functional status [measured by the 36-item Short Form Health Survey (SF-36) physical functioning scale] and an improvement in global assessment was found (37).An internet-based patient education Bosch Romero (2002) [40] Soares (2002) [36] Williams (2010) [37] Barrenengoa-Cuadra (2021) [38] Saracoglu (2021) [39] Selection bias Performance bias Detection bias Attrition bias Reporting bias Other bias Overall bias intervention could perhaps be an efficient delivery method, but only to a subgroup of patients.This is because patients with FM reporting lower baseline levels of pain found internet-based treatment to be beneficial, as opposed to those with high levels of pain (47).This was independent of the participants' age, duration of FM, and educational level (47).

Strengths and limitations
This review has several strengths, including a thorough search strategy, and double study selection and data extraction.The review was based on a detailed protocol that was prospectively registered in an online registry before conducting the search.However, several limitations need consideration.First, only a limited number of RCTs evaluating patient education as a stand-alone intervention has been published.Therefore, trial inclusion in this review was limited.Furthermore, all included trials scored a high risk of bias.This was partly due to the included trials having small patient populations and the difficulty of not being able to blind participants or personnel in an intervention concerning education.Even though the interventions all considered patient education, the duration of the interventions, delivery methods, and content were very diverse.A large variety of questionnaires was used for data collection in the individual trials.All of this probably added to the large degree of heterogeneity found in this review.The large spread in disease duration across the included trials, which could influence the results, is particularly noteworthy.The trial by Soares and Grossi involved participants with the shortest duration of FM and was the only trial with no significant effect of patient education on any of the outcomes.Another limitation may be the use of classification criteria versus diagnostic criteria across trials.Three of the included trials used ACR 1990 criteria (36,37,40), whereas two used ACR 2010/2016 criteria (38,39).Originally, we had also expected to find studies with patients diagnosed with CWP as a stand-alone diagnosis.However, all of the included studies only presented patients diagnosed with FM, thereby primarily including patients with FM representing the upper end of the disease severity spectrum and mostly women.Furthermore, we had planned to conduct several subgroup analyses, but this was not possible because of the small number of trials.

Implications
This review reveals a need for future well-designed RCTs investigating patient education as a stand-alone intervention not only for patients with FM, but also for patients with CWP, to cover the entire disease spectrum.
The patient population of FM is known to be heterogeneic, and previous studies have found different subgroups of FM to have different responses to standardized intervention programmes (48,49).However, studies using clinically relevant subgroups to direct interventions and predict outcome are still missing.Interventions may best be delivered when matched to the patients' needs.None of the included trials in this study reported on tailored patient educational programmes guided by learning priorities.

Conclusion
This review found that patient education as a standalone intervention can improve global assessment in patients with FM.No differences were found between the comparator and patient education with regard to functioning, HRQoL, level of pain, emotional distress (anxiety and depression), or pain-related cognition.Concerning harm, no difference was found in the total number of withdrawals.To improve the quality of therapeutic patient education evidence, we propose that future trials should include patients with CWP and that effectiveness assessment should include outcome measures such as disease acceptance, HRQoL, enhancement of patients' knowledge of pain, pain coping skills, and evaluation of prioritized learning outcomes.

Figure 1 .
Figure 1.Flow diagram showing the selection of the trials.

Figure 2 .Figure 3 .
Figure 2. Functioning.Functioning for Soares and Grossi (36) was assessed post-intervention with the Arthritis Self-Efficacy Scale (ASES) function subscale, for Williams et al (37) at 6 months' follow-up with the 36-item Short Form Health Survey (SF-36) physical functioning subscale, and for Barrenengoa-Cuadra et al (38) at 6 months' follow-up with the Fibromyalgia Impact Questionnaire (FIQ) function subscale.CI, confidence interval; SD, standard deviation; Std.mean difference, standardized mean difference.

Figure 4 .
Figure 4. Risk of bias in included randomized clinical trials evaluated with the Cochrane Collaboration's risk of bias instrument (26)., high risk of bias;, unclear risk of bias; , low risk of bias.

Table 1 .
Study characteristics of the eligible trials.

Table 2 .
GRADE evidence profile of patient education as a stand-alone intervention included in randomized trials.
*Evaluated using the Cochrane Collaboration's risk of bias instrument (26).† Values of I 2 > 75% were interpreted as high heterogeneity.