Received: 14 August 2018

| Revised: 20 December 2018

| Accepted: 30 December 2018

DOI: 10.1002/ejp.1358

REVIEW ARTICLE

The effectiveness of self‐management interventions in adults with

chronic orofacial pain: A systematic review, meta‐analysis and
meta‐regression

Vishal R. Aggarwal1 | Yu Fu2 | Chris J. Main3 | Jianhua Wu1

1
Faculty of Medicine & Health, School of
Dentistry, University of Leeds, Leeds, UK
Abstract
2
Faculty of Medicine & Health, School of Background: Psychosocial risk factors associated with chronic orofacial pain are
Healthcare, University of Leeds, Leeds, UK amenable to self‐management. However, current management involves invasive
3
Research Institute for Primary Care & therapies which lack an evidence base and has the potential to cause iatrogenic harm.
Health, Keele University, Keele, UK
Objectives: To determine: (a) whether self‐management is more effective than usual
Correspondence care in improving pain intensity and psychosocial well‐being and (b) optimal com-
Vishal R. Aggarwal, Faculty of Medicine & ponents of self‐management interventions.
Health, School of Dentistry, University of
Leeds, Leeds, UK.
Databases and data treatment: Cochrane Oral Health Group Trials Register,
Email: [email protected] Cochrane Central Register of Controlled Trials, MEDLINE, EMBASE, PsycINFO,
WHO International Clinical Trials Registry Platform and Clinical Trials.gov were
searched. Meta‐analysis was used to determine effectiveness, and GRADE was used
to rate quality, certainty and applicability of evidence.
Results: Fourteen trials were included. Meta‐analyses showed self‐management was
effective for long‐term pain intensity (standardized mean difference [SMD] −0.32,
95% confidence interval [CI] −0.47 to −0.17) and depression (SMD −0.32, 95% CI
−0.50 to −0.15). GRADE analysis showed a high score for certainty of evidence for
these outcomes and significant effects for additional outcomes of activity interfer-
ence (−0.29 95% CI −0.47 to −0.11) and muscle palpation pain (SMD −0.58 95% CI
−0.92 to −0.24). Meta‐regression showed nonsignificant effects for biofeedback on
long‐term pain (−0.16, 95% CI −0.48 to 0.17, p‐value = 0.360) and depression
(−0.13, 95% CI −0.50 to 0.23, p‐value = 0.475).
Conclusions: Self‐management interventions are effective for patients with chronic
orofacial pain. Packages of physical and psychosocial self‐regulation and education
appear beneficial. Early self‐management of chronic orofacial pain should be a prior-
ity for future testing.
Significance: This systematic review provides clear evidence for effectiveness of
combined biomedical and psychological interventions (incorporating self‐manage-
ment approaches) on long‐term outcomes in the management of chronic orofacial
(principally TMD) pain. Self‐management should be a priority for early intervention
in primary care in preference to invasive, irreversible and costly therapies. Further
research is needed firstly to clarify the relative effectiveness of specific components
of self‐management, both individually and in conjunction, and secondly on outcomes
in other types of chronic orofacial pains.

Eur J Pain. 2019;23:849–865. wileyonlinelibrary.com/journal/ejp © 2019 European Pain Federation ‐ EFIC® | 849
 |

15322149, 2019, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejp.1358 by UNIVERSIDAD DE VALPARAISO, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
850    AGGARWAL et al.

1 | BACKGROU N D dental practitioners (31%) (Bell, Smith, Rodgers, Flynn, &

Malone, 2008). General medical practitioners do not have the
Persistent pain in the face or mouth is a frequent cause for infrastructure or knowledge to manage chronic orofacial pain
consultation in both primary dental and medical care, and in and indeed find it difficult (Peters et al., 2015). Patients are
a substantial proportion of cases, it can become both chronic therefore referred from specialist to specialist and have mul-
and disabling (Aggarwal, McBeth, Zakrzewska, Lunt, & tiple tests, investigations and often invasive and irreversible
Macfarlane, 2008; Macfarlane, Blinkhorn, Davies, Kincey, treatments that do not improve symptoms (Beecroft et al.,
& Worthington, 2002). Subjects who report orofacial pain 2013; Durham et al., 2007; Elrasheed et al., 2004; Pfaffenrath
for three months or more report increased pain level and dis- et al., 1993). Costs of TMD alone in the United States are
ability and are also more likely to seek treatment and take in the region of $4 billion annually (Gatchel et al., 2006),
medication (Macfarlane et al., 2002). Chronic orofacial pain and a study examining the costs to the UK National Health
(OFP) is the characteristic feature of a number of clinical con- Service (Durham, Shen et al., 2016) showed that consultation
ditions, such as temporomandibular joint disorder (TMD), costs were a significant proportion (p < 0.001) of cumulative
burning mouth syndrome, atypical odontalgia and atypical healthcare utilization costs of patients with persistent orofa-
facial pain that are difficult to diagnose and treat (Durham, cial pain. This imposes a huge burden on already stretched
Exley, Wassell, & Steele, 2007; Elrasheed, Worthington, healthcare resources. The descriptive epidemiology of chronic
Ariyaratnam, & Duxbury, 2004; Pfaffenrath, Rath, Pöllmann, orofacial shows a strong association with psychosocial risk
& Keeser, 1993). TMD is globally the most common oro- factors (Aggarwal et al., 2008; Bair et al., 2016; Slade et al.,
facial pain condition, and in the United States, a preva- 2007, 2016) and a co‐occurrence with other long‐term con-
lence of 6% in women and 3.5% in men has been reported ditions like chronic widespread pain (CWP), irritable bowel
(Lipton, Ship, & Larach‐Robinson, 1993); in the United syndrome (IBS) and chronic fatigue (CF) (Aggarwal et al.,
Kingdom, the prevalence of chronic orofacial pain is similar 2006; Bair et al., 2016; Slade et al., 2016).
at 7% (Aggarwal, McBeth, Zakrzewska, Lunt, & Macfarlane, In line with a global drive to curb the epidemic of
2006). The American Academy of Orofacial pain suggests ­noncommunicable diseases and long‐term conditions, UK
that in any given year 10% of women and 6% of men (approx- government policy places an emphasis on using self‐man-
imately 20 million adults) have TMD pain (Gatchel, Stowell, agement to improve management of long‐term conditions
Wildenstein, Riggs, & Ellis, 2006). Reports from European through patient participation and ownership of their own
studies also have similar prevalence figures (6.7%) for TMD health care (Department of Health, 2001, 2005). Self‐man-
(Johansson, Unell, Carlsson, Söderfeldt, & Halling, 2003). agement approaches (where the person takes an active role
Patients with chronic orofacial pain are likely to be fre- in managing their condition rather than a passive one that
quent consulters to primary, secondary and tertiary care is more dependent on others) are increasingly accepted for
and undergo multiple investigations to determine an organic chronic pain (Nicholas & Blyth, 2016). This term refers to
cause for their symptoms—although underlying organic pa- all actions taken by individuals to manage the symptoms,
thology is rarely found (Durham et al., 2007; Elrasheed et treatment, physical and psychosocial consequences and
al., 2004; Pfaffenrath et al., 1993). Management of chronic lifestyle changes inherent in living with a chronic condition
orofacial pain by dentists tends to focus on correction of local (Barlow, Wright, Sheasby, Turner, & Hainsworth, 2002).
mechanical factors, such as teeth grinding and malocclusion. Self‐management interventions aim to increase the capac-
However, evidence in the form of Cochrane systematic re- ity, confidence and efficacy of the individual and are in-
views has shown little or no beneficial effects of invasive creasingly viewed as core strategies of the management of
physical therapies such as irreversible occlusal adjustments chronic conditions (Kennedy et al., 2013). Education and
(Koh & Robinson, 2003) and oral splints (Al‐Ani, Gray, skill development are two common components of those in-
Davies, Sloan, & Glenny, 2005; List & Axelsson, 2010). terventions that are tailored to influence individual's cogni-
Indeed, an audit of 101 consecutive referrals of persistent tive, behavioural and emotional responses to maintain and
orofacial pain to a secondary care Oral Surgery department strengthen a satisfactory quality of life (Barlow, 2001). The
(Beecroft, Durham, & Thomson, 2013) showed that patients boundary between “active” and “passive” treatment how-
had been treated in nine different hospitals, referred to 15 ever is not absolute, and it could be argued that anything
distinct specialties with a mean of seven consultations per done by the patient in an endeavour to better manage their
specialty. Overall 341 treatment attempts had been made and symptoms, function or associated distress could be viewed
only 24% yielded a successful outcome. The study concluded as self‐management. However, the term self‐management
that there was a need for evidence‐based management and approach normally has a specific cognitive or behavioural
specialist regional centres (Beecroft et al., 2013). focus and is normally contrasted with passive treatment pri-
Patients with orofacial symptoms also frequently consult marily delivered by a healthcare practitioner. Currently, it is
their general medical practitioner (69%) rather than general normally taken to apply to pain coping strategies employed
 |

15322149, 2019, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejp.1358 by UNIVERSIDAD DE VALPARAISO, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
AGGARWAL et al.    851

by the patient to help manage their pain and its impact. This
aligns with TMD interventions which aim to target these fac-
2 | M ETHODS
tors using techniques such as psychoeducation, relaxation, This systematic review and meta‐analysis were undertaken
jaw posture control, cognitive behaviour therapy (CBT) and following the Preferred Reporting Items for Systematic re-
biofeedback as per previous studies (Goldthorpe, Lovell, views and Meta‐Analyses (PRISMA) statement (Liberati et
Peters, McGowan, & Aggarwal, 2016; Litt, Shafer, Ibanez, & al., 2009; Moher, Liberati, Tetzlaff, Altman, & PRISMA
Kreutzer, 2009; Turner, Holtzman, & Mancl, 2007; Turner, Group, 2009). This study is registered with PROSPERO
Whitney, Dworkin, Massoth, & Wilson, 1995). These stud- (CRD42017060158; Aggarwal, Fu, Main, & Wu, 2018).
ies have not only outlined components for biopsychosocial
interventions for chronic orofacial pain and TMD but also
explored the mechanisms by which self‐care interventions 2.1 | Criteria for considering studies for
involving both psychosocial self‐care and jaw posture con- this review
trol can bring about change in patients with chronic orofacial
pain. Guided self‐care interventions can target vicious cycles 2.1.1 | Types of studies
associated with both fear‐avoidance behaviour (central pain Randomized controlled trials which included self‐manage-
processing mechanisms) and “anxiety–pain–tension” cycles ment of chronic orofacial pain compared with any other form
involving muscle over activity linked to emotional stress (de- of treatment such as surgery, usual care, pharmacological
pression, anger, fears and anxieties about the pain), which treatment and/or waiting list controls.
in turn may increase pain by precipitating activity in psy-
chophysiological systems. By changing patient beliefs and
developing coping strategies, self‐management interventions 2.1.2 | Types of participants
have the potential to induce a return to normal functioning. Adults over 18 years of age with chronic orofacial pain de-
(Goldthorpe, Lovell et al., 2016; Litt et al., 2009; Turner et fined as those diagnosed with the following conditions:
al., 2007, 1995). Such interventions are noninvasive and have temporomandibular disorders (TMD), atypical facial pain,
the potential, if effective, to be applied across health care and atypical odontalgia and burning mouth syndrome. Other
delivered by general medical practitioners to whom patients terms used to describe these conditions were also included
with orofacial symptoms frequently consult. in the search strategy, for example, myofacial pain, myo-
Key components of such interventions have included psy- fascial pain related to the facial region, craniomandibular/
choeducation, relaxation, CBT and biofeedback (Goldthorpe, oromandibular dysfunction, mandibular stress syndrome, fa-
Lovell et al., 2016; Litt et al., 2009; Turner et al., 2007, cial arthromyalgia, masticatory muscle disorder, masticatory
1995). However, biofeedback, in particular EMG biofeedback myalgia, TMJ syndrome, stomatodynia, persistent idiopathic
(Gatchel et al., 2006), requires not only expensive equipment facial pain and persistent dento‐alveolar pain.
but also time spent on training and particularly time spent by
patients on practice. This may not be amenable to self‐man-
agement particularly for interventions that need to be deliv-
2.1.3 | Types of outcome measures
ered remotely by telephone or Web‐based interactions. Primary outcomes
The aim of the current review was therefore to assess the 1. Pain intensity (short and/or long term) measured using
effectiveness of self‐management interventions compared a visual analogue scale or a validated categorical scale,
with usual care in the management of adults with chronic for example, Brief Pain Inventory and Multidimensional
orofacial pain. Pain Inventory.
Specific objectives: 2. Depression/Anxiety (long and short terms using validated
scales, e.g., Hospital Anxiety and Depression Scale).
1. To determine whether, in adults with chronic orofacial 3. Interference with life—pain impact on activities of daily
pain including temporomandibular disorders (TMD), living measured using, for example, Brief Pain Inventory
self‐management interventions more effective than usual and Multidimensional Pain Inventory.
care in improving long‐term outcomes related to pain
intensity and psychosocial well‐being.
2. To determine whether the biofeedback component of in-
2.1.4 | Types of interventions
terventions shows an additional treatment effect compared Self‐management interventions were defined as those that in-
to no biofeedback. cluded patient participation in the intervention. Table 1 illus-
3. To determine the effectiveness of self‐management for trates the components of the interventions. Trials were eligible
subtypes of chronic orofacial pain in particular TMD for inclusion into self‐management as they included patient
which is the most common subtype. participation through a patient manual and/or between session
 852
|

TABLE 1 Components of self‐management interventions for included studies

  

Self‐management Physical self‐regulation Psychological

Jaw posture
Patient Between relaxation and Breathing Cognitive Behaviour
Study details COFP subtype manual session work habit reversal Biofeedback techniques therapy therapy Education
Bergdahl et al. (1995) BMS √ √
Carlson et al. (2001) TMD √ √ Diaphragmatic Patients were instructed to wear
breathing the splint at night and were
provided with general informa-
tion regarding aetiology and
self‐care strategies for managing
myofascial pain
Crockett et al. (1986) TMD √ √ √ √ √
Dworkin et al. (1994) TMD √ √ √ √ √
Ferrando et al. (2012) TMD √ √ √ Psychoeducation
Gardea et al. (2001) TMD √ √ √ √ √ √ Education of stress and relation-
ship to anxiety, depression and
pain
Gatchel et al. (2006) TMD √ √ √ √ √ √ Education (mind–body relation-
ship to stress and body's reaction
to stress)
Goldthorpe et al. (2017) All subtypes √ √ √ √
Litt et al. (2010) TMD √ √ √ √ √
Shedden‐Mora et al. (2013) TMD √ √ √ √ √ √ Patients were educated about
symptoms and causes of their
TMD
Townsend et al. (2001) TMD √ √ √ Deep breathing
Turk et al. (1993) TMD √ √ √ √ √ Didactic education on link
between stress, muscle tension
and pain
Turk et al. (1996) √ √ √ √ √ Didactic education regarding the
association between stress,
increased muscle tension and
pain
Turner et al. (2006) TMD √ √ √ √
Total 10 12 11 6 11 11
AGGARWAL et al.

15322149, 2019, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejp.1358 by UNIVERSIDAD DE VALPARAISO, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
 |

15322149, 2019, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejp.1358 by UNIVERSIDAD DE VALPARAISO, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
AGGARWAL et al.    853

work as part of the intervention protocol. Other components the characteristics of trial participants, interventions, control
were education, psychological such as cognitive behaviour ther- groups and outcomes. Characteristics of included studies are
apy or its components (cognitive therapy, behavioural therapy) presented in Supporting Information Appendix S2. VA ex-
and physical self‐regulation, for example, posture control, habit tracted all the studies while JW and YF shared equally ex-
reversal, relaxation and/or biofeedback. Table 1 summarizes traction for the purpose of duplication. Any differences were
the intervention components of studies and how these map onto resolved by discussion. Differences involving risk of bias
self‐management. were resolved by using the most frequent option selected, for
example, if two of the three reviewers were in agreement then
we chose that option. There were no instances where there
2.1.5 | Search methods for
was disagreement between all 3 reviewers. Prior to extrac-
identification of studies
tion, the data extraction form was piloted using three studies
For the identification of studies included or considered for and all authors extracting the data participated in the piloting
this review, detailed search strategies were developed for so that they were clear about the extraction process. The data
each database searched. These were based on the search strat- extraction form was modified for ease of use following the
egy developed for MEDLINE (OVID) but revised appropri- pilot extractions.
ately for each database.
The following electronic databases were searched (to 29
2.3 | Assessment of risk of bias in
September 2017): The Cochrane Oral Health Group Trials
included studies
Register, the Cochrane Central Register of Controlled Trials
(CENTRAL), MEDLINE via OVID, EMBASE via OVID, The assessment of risk of bias in the included trials was un-
PsycINFO via OVID, WHO International Clinical Trials dertaken independently and in duplicate as part of the data ex-
Registry Platform, and Clinical Trials.gov. There were no traction process by three of the review authors (VA, JW and
restrictions regarding language or date of publication. The YF) as described above and in accordance with the Cochrane
search strategy used a combination of controlled vocabulary Handbook for Systematic Reviews of Interventions 5.0.2
and free text terms for identifying randomized trials (RCTs) (Higgins & Green, 2011). Included trials were assessed on
in MEDLINE. Details of the search strategy are provided in the following criteria:
Supporting Information Appendix S1.
The reference lists of all eligible trials were checked • adequate sequence generation
for additional studies. Where these had not already been • concealed allocation of treatment
searched, the journals were hand searched by the review au- • blinding of participants/caregivers (where feasible) and
thors if electronic copies were not available. outcome assessors
• incomplete outcome data
• selective outcome reporting
2.2 | Data collection and analysis
• any other bias relevant to the study
2.2.1 | Selection of studies
A description of the quality items was tabulated for each
The title and abstracts of relevant articles and reports from the included trial, along with a judgement of low, high or uncer-
search strategy outlined in Supporting Information Appendix tain risk of bias. Criteria for risk of bias judgements regard-
S1 were screened independently by two review authors (VA ing allocation concealment were as described in the Cochrane
and JW). Full reports were obtained where trials met the inclu- Handbook for Systematic Reviews of Interventions (Higgins
sion criteria or where a clear decision could not be made from & Green, 2011):
the title or abstract. Disagreements were resolved by discus-
sion, and full reports of all studies potentially meeting the in- • Low risk of bias—adequate concealment of the allocation
clusion criteria were obtained. Full reports were used to assess (e.g., sequentially numbered, sealed, opaque envelopes or
trials where inclusion was unclear and reasons for rejection centralized or pharmacy‐controlled randomization).
were clear upon examining full reports. Main reasons for rejec- • Uncertain risk of bias—uncertainty about whether the al-
tion were as follows: Studies were not randomized controlled location was adequately concealed (e.g., where the method
trials and had the wrong disease definition and/or patient group. of concealment is not described or not described in suffi-
cient detail to allow a definite judgement).
• High risk of bias—inadequate allocation concealment
2.2.2 | Data extraction and management
(e.g., open random number lists or quasi‐randomiza-
Data were extracted, independently and in duplicate, using tion such as alternate days, date of birth or case record
a previously prepared data extraction form which included number).
 |

15322149, 2019, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejp.1358 by UNIVERSIDAD DE VALPARAISO, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
854    AGGARWAL et al.

A summary assessment of the risk of bias for the pri-

mary outcome (across domains) within and across studies
2.8 | Quality of evidence
was undertaken. Within a study, a summary assessment of We used the Grading of Recommendations Assessment,
low risk of bias was given when there was a low risk of bias Development and Evaluation (GRADE) approach to assess
for all key domains, unclear risk of bias when there was an the quality and certainty of the body of evidence per outcome,
unclear risk of bias for one or more key domains, and high in accordance with the Cochrane Handbook for Systematic
risk of bias when there was a high risk of bias for one or Reviews of Interventions 5.0.2 (Higgins & Green, 2011).
more key domains. Across studies, a summary assessment For the most important outcomes, we used the programme
was rated as low risk of bias when most information is from GRADE pro GDT 2015 to generate a certainty of evidence
studies at low risk of bias, unclear risk of bias when most table (Table 2). Starting from an assumed level of high qual-
information was from studies at low or unclear risk of bias, ity, this reduced the quality of the evidence by one or more
and high risk of bias when the proportion of information levels if there were one or more limitations in the risk of bias,
was from studies at high risk of bias sufficient to affect the consistency, and/or precision of the pooled estimate. The
interpretation of the results. level of evidence as then rated as either high, moderate, low
or very low depending on the number of limitations.

2.4 | Measures of treatment effect

For dichotomous outcomes, treatment effects were ex-
2.9 | Assessment of intervention components
pressed as risk ratios with 95% confidence intervals while
for continuous outcomes mean differences with 95% con-
2.9.1 | Meta‐regression
fidence intervals were used. All analyses were performed Simple mixed‐effects meta‐regression was used to inves-
using R version 3.4.1 (https://cran.r-project.org/; R Core tigate whether biofeedback provided additional treatment
Team, 2013). effect. We performed meta‐regression on outcome meas-
ures of long‐term pain and depression between patients
with biofeedback and those without biofeedback.
2.5 | Assessment of heterogeneity
Clinical heterogeneity was accounted for by inclusion criteria
for uniform disease definition, assessing components of the 3 | RESULTS
interventions and outcome measures included in the trials.
Statistical heterogeneity was assessed by means of Cochrane
Q, where a large Q value indicates the presence of hetero-
3.1 | Description of studies
geneity, and the I2 statistic where I2 gives the percentage of A detailed description of the studies is in the characteristics
variability in the effect estimate that is due to heterogeneity of included and excluded studies presented in Supporting
rather than to chance. Suggested thresholds for the interpreta- Information Appendix S2.
tion of I2 are as follows: Less than 40% indicate there is no
problem with heterogeneity, 30%–60% indicates a moderate
problem, 60%–90% a substantial problem and 75% and over
3.2 | Results of the search
considerable heterogeneity (Higgins & Green, 2011). The initial search strategy yielded 1,104 references which
were assessed blind and independently by VA and JW, and
based on the abstracts and titles, these were reduced to 48
2.6 | Assessment of reporting biases
relevant manuscripts (Figure 1). Main reasons for exclusion
Reporting biases were assessed through funnel plots for out- were that a large proportion of studies were not trials and oth-
comes that were reported by more than five studies. Egger's ers were not on chronic orofacial pain.
test was used to test the statistical significance of reporting All the 48 manuscripts identified above were extracted
biases for each outcome. by the lead author VA. Extraction was duplicated by sharing
blind and independently between the other co‐authors (JW,
YF). Sixteen manuscripts were relevant for analysis and are
2.7 | Data synthesis
presented in the characteristics of included studies table in
Meta‐analyses were only carried out if trials were of simi- Supporting Information Appendix S1. A number of trials that
lar comparisons reporting the same outcome measures. were duplicates of the same study were merged. Reasons for
Estimates of effect were combined using a random‐effects exclusion at this stage were interventions not compatible with
model. Mean differences or standardized mean differences self‐management, had the wrong disease definition and/or pa-
were used for the same outcomes with different scales. tient group, and they were not randomized controlled trials. Of
 |

15322149, 2019, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejp.1358 by UNIVERSIDAD DE VALPARAISO, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
AGGARWAL et al.    855

TABLE 2 GRADE analysis showing certainty of evidence for self‐management compared to usual care for chronic orofacial pain

Anticipated absolute effects

No. of participants Certainty of the Risk with Risk difference with

Outcomes (studies) follow‐up evidence (GRADE) usual care self‐managementa
Pain short term (≤3 months) assessed 779 (14 RCTs) ⨁⨁⨁⨁ – SMD 0.06 SD lower (0.22 lower to 0.09
with: VA, YF, JW HIGH higher)
Pain long term (>3 months) assessed 757 (12 RCTs) ⨁⨁⨁⨁ – SMD 0.32 SD lower (0.47 lower to 0.17
with: VA, YF, JW HIGH lower)
Muscle palpation pain long term 143 (3 RCTs) ⨁⨁⨁⨁ – SMD 0.58 SD lower (0.92 lower to 0.24
(>3 months) assessed with: VA, HIGH lower)
YF, JW
Activity interference/disability 527 (10 RCTs) ⨁⨁⨁⨁ – SMD 0.29 SD lower (0.47 lower to 0.11
(>3 months) assessed with: VA, HIGH lower)
YF, JW
Depression long term (>3 months) 524 (8 RCTs) ⨁⨁⨁⨁ – SMD 0.32 SD lower (0.5 lower to 0.15
assessed with: VA, YF, JW HIGH lower)
Note. CI: Confidence interval; SMD: standardized mean difference.
a
The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and
its 95% CI).

FIGURE 1 PRISMA flow diagram

the 16 studies which met all eligibility criteria and hence in- study (Komiyama, Kawara, Arai, Asano, & Kobayashi, 1999)
cluded in this review, Dworkin's 2 studies (Dworkin, Huggins displayed results graphically and we did not have means
et al., 2002; Dworkin, Turner et al., 2002) and Komiyama's and standard deviations to pool these studies. Authors were
 |

15322149, 2019, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejp.1358 by UNIVERSIDAD DE VALPARAISO, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
856    AGGARWAL et al.

contacted to obtain data but only provided means and no stan- (3 months or less) and long‐term (more than 3 months) pain
dard deviations or did not respond. This left 14 studies for in- intensity and long‐term measures for muscle palpation pain,
clusion in the final meta‐analysis (Figure 1). activity interference and depression.

3.3 | Included studies 3.4 | Risk of bias in included studies

All of the included trials had comparable control groups Risk‐of‐bias plots are displayed in Figure 2a,b, the former show-
comprising usual treatment which involved conservative ing the overall risk of bias and the latter individual plots for each
treatment composed of education, counselling and an in- study. Figure 2c shows funnel plots for publication bias.
traoral flat‐plane appliance. The Bergdahl study (Bergdahl,
Anneroth, & Ferris, 1995) included a control group of atten-
3.5 | Blinding (performance bias and
tion placebo and the Townsend study (Townsend, Nicholson,
detection bias)
Buenaver, Bush, & Gramling, 2001) included a waiting list
control with no intervention and were therefore not pooled It is notable that due to the nature of the intervention, blind-
in the meta‐analysis as they had different comparators. They ing was difficult where the intervention and controls were con-
were however used for the GRADE analysis (Table 2). cerned. However, it was possible for outcome assessment, and
The interventions for self‐management were as de- for the purposes of this review, we evaluated whether included
fined previously. Outcome measures included short‐term studies had blinded outcome measurement. This was reported

A B

F I G U R E 2 (a) Overall risk of bias. (b) Risk of bias for individual studies. Green circles with “+” symbol indicate low risk of bias; yellow
circles with “?” symbol indicate unclear risk of bias; red circles with “–” symbol indicate high risk of bias. (c) Funnel plots for outcomes reported
by more than five studies. Dots outside the funnel indicate outliers (Egger's test p‐values: short‐term pain = 0.35; long‐term pain = 0.52; activity
interference = 0.34; long‐term depression = 0.69)
 |

15322149, 2019, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejp.1358 by UNIVERSIDAD DE VALPARAISO, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
AGGARWAL et al.    857

by seven of the included studies (Carlson, Bertrand, Ehrlich, 1986; Dworkin et al., 1994; Ferrando et al., 2012; Gardea et
Maxwell, & Burton, 2001; Dworkin et al., 1994; Ferrando et al., 2001; Goldthorpe et al., 2017; Litt et al., 2010; Shedden‐
al., 2012; Gardea, Gatchel, & Mishra, 2001; Goldthorpe et Mora et al., 2013; Turk, Zaki, & Rudy, 1993).
al., 2017; Shedden‐Mora, Weber, Neff, & Rief, 2013; Turner, Due to substantial heterogeneity (I2 = 62%), the results
Mancl, & Aaron, 2006), and three did not report at all (Bergdahl of these studies could not be pooled (Figure 3). Hence, no
et al., 1995; Gatchel et al., 2006; Litt, Shafer, & Kreutzer, overall conclusions could be drawn for this domain. Of the
2010). The remaining studies were unclear (Figure 2b). The studies that did not have quantitative data for this outcome,
overall risk of bias was deemed low in this area (Figure 2a). the Komiyama paper (Komiyama et al., 1999) showed no
differences in pain intensity between the self‐management
3.6 | Incomplete outcome data (attrition intervention and control groups. In contrast, the Dworkin
bias) comprehensive care programme study (Dworkin, Huggins et
al., 2002) showed significant improvement in short‐term pain
Only three trials did not report on incomplete outcome data; intensities between self‐management and usual care.
nine fully reported this (Bergdahl et al., 1995; Carlson et al.,
2001; Dworkin et al., 1994; Ferrando et al., 2012; Gardea
et al., 2001; Gatchel et al., 2006; Goldthorpe et al., 2017; 3.10.2 | Self‐management interventions
Shedden‐Mora et al., 2013; Turner et al., 2006), and the one versus usual care—pain (long term)
was unclear (Litt et al., 2010) and risk of bias (Figure 2b) was Nine studies provided data on this outcome (Carlson et al.,
therefore low for this domain (Figure 2a). 2001; Dworkin et al., 1994; Gardea et al., 2001; Gatchel et
al., 2006; Goldthorpe et al., 2017; Litt et al., 2010; Shedden‐
3.7 | Allocation (selection bias) Mora et al., 2013; Turk et al., 1993; Turner et al., 2006).
Due to low heterogeneity (I2 = 7%), the results of the
This was not reported by only three of the included studies studies could be pooled for the purpose of statistical analysis
(Ferrando et al., 2012; Litt et al., 2010); fully reported by (Figure 4). This showed a statistically significant difference in
four studies (Gardea et al., 2001; Goldthorpe et al., 2017; favour of self‐management interventions (SMD −0.32, 95%
Shedden‐Mora et al., 2013; Turner et al., 2006); and the re- CI −0.47 to −0.17), and this represented a 16% improvement
maining studies were unclear (Figure 2b). Overall, the risk of in long‐term pain for self‐care versus usual care for patients
bias in this area was therefore low (Figure 2a). with chronic orofacial pain (Figure 4).
Considering subgroups of interventions, statistically sig-
3.8 | Selective reporting (reporting bias) nificant differences were observed for self‐care CBT (SMD
−0.26, 95% CI −0.45 to −0.07) and combined biofeedback
None of the included trials had selective reporting and there- and CBT (SMD −0.46 95% CI −0.72 to −0.20) (Figure 4).
fore were assessed as being at low risk of bias for selective Of the studies that did not have quantitative data for this out-
reporting (Figure 2a). come, the Dworkin self‐care intervention (Dworkin, Huggins
et al., 2002) showed significant (p < 0.05) improvement in
3.9 | Publication bias long‐term pain intensity while the comprehensive care pro-
gramme study (Dworkin, Turner et al., 2002) did not.
There were only two outliers for short‐term pain intensity and
one for long‐term pain intensity and activity interference for
funnel plots (Figure 2c) which may indicate the existence of
3.10.3 | Self‐management interventions
publication bias. However, formal tests showed that this was
versus usual care—muscle palpation pain (long
not statistically significant (Egger's test, p‐value for short‐
term)
term pain = 0.35, long‐term pain = 0.52, activity interfer- Overall only three studies provided data on this outcome
ence = 0.34 and long‐term depression = 0.69). (Carlson et al., 2001; Turk, Rudy, Kubinski, Zaki, & Greco,
1996; Turk et al., 1993).
Only three studies provided data on this outcome, and
3.10 | Effectiveness of self‐management
because there was substantial heterogeneity (I = 63%), the
interventions
pooled results were unreliable although they showed a sig-
nificant improvement in muscle palpation pain (SMD −0.58
3.10.1 | Self‐management interventions
95% CI −0.92 to −0.24) (Table 3). There were insufficient
versus usual care pain (short term)
data to draw any conclusions regarding any of the individ-
Nine studies provided comparable data for this outcome ual interventions with regard to muscle palpation pain (long
(Carlson et al., 2001; Crockett, Foreman, Alden, & Blasberg, term). Of the studies that did not have quantitative data for
 |

15322149, 2019, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejp.1358 by UNIVERSIDAD DE VALPARAISO, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
858    AGGARWAL et al.

FIGURE 3 Comparison—any self‐management intervention versus usual care outcome—pain short term (3 months or less)

this outcome, the Dworkin self‐care intervention (Dworkin, Huggins et al., 2002) showed significant (p < 0.05) im-
Huggins et al., 2002) showed significant (p < 0.05) improve- provement in this outcome while the comprehensive care
ment in this outcome. programme study (Dworkin, Turner et al., 2002) did not.

3.10.4 | Self‐management interventions 3.10.5 | Self‐management interventions

versus usual care—activity interference (long versus usual care—depression (long term)
term)
A total of seven studies provided data for the statistical analy-
A total of eight studies provided data for this outcome sis for this outcome (Carlson et al., 2001; Gatchel et al., 2006;
(Carlson et al., 2001; Dworkin et al., 1994; Ferrando et al., Goldthorpe et al., 2017; Litt et al., 2010; Shedden‐Mora et
2012; Gardea et al., 2001; Goldthorpe et al., 2017; Litt et al., al., 2013; Turk et al., 1996, 1993).
2010; Shedden‐Mora et al., 2013; Turk et al., 1996). Overall seven studies provided data on this outcome and
Eight studies provided data for this outcome, and there there were statistically significant differences in favour of
was a significant effect of the pooled results (SMD −0.29 psychosocial interventions (SMD −0.32, 95% CI −0.50 to
95% CI −0.47, −0.11) (Table 3). However, because there was −0.15) (Figure 5) and this represented a 25% improvement
substantial heterogeneity (I2 = 79%) the pooled results are in long‐term pain for psychosocial interventions versus
unreliable. Individually, there was statistically significant usual care. There was no heterogeneity (I2 = 0%; Figure 5).
difference for self‐care CBT (SMD −0.37 95% CI −0.57, Individually, both self‐care CBT and CBT/biofeedback
−0.16). Of the studies that did not have quantitative data for show statistically significant benefit over usual care with re-
this outcome, the Dworkin self‐care intervention (Dworkin, gard to depression (SMD −0.27, 95% CI −0.49 to −0.05)
 |

15322149, 2019, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejp.1358 by UNIVERSIDAD DE VALPARAISO, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
AGGARWAL et al.    859

FIGURE 4 Comparison—self‐management intervention versus usual care outcome—pain long term (>3 months)

TABLE 3 Effectiveness of self‐management compared to usual care on muscle palpation pain and activity interference

Outcomes Intervention Control No. of studies Pooled effect Heterogeneity (%)

Muscle palpation pain (>3 months) Combined self‐care Usual care 1 −0.39 (−0.91, 0.13) –
biofeedback and CBT
Self‐care CBT Usual care 2 −0.72 (−1.16, −0.27) 78
All intervention Usual care 3 −0.58 (−0.92, −0.24) 63
Activity interference/disability Combined self‐care Usual care 2 0.06 (−0.40, 0.52) 0
Long term (>3 months) biofeedback and CBT
Self‐care CBT Usual care 7 −0.37 (−0.57, −0.16) 85
All intervention Usual care 9 −0.29 (−0.47, −0.11) 79

and (SMD −0.41, 95% CI −0.68 to −0.13) respectively interventions on activity interference (SMD −0.29, 95% CI
(Figure 5). −0.47 to −0.11) and long‐term muscle palpation pain (SMD
−0.58, 95% CI −0.92 to −0.24). The effect for short‐term
pain remained nonsignificant (SMD −0.06, 95% CI −0.21
3.11 | Certainty of the evidence
to 0.09).
The certainty of the evidence was high for the main outcome
measures as assessed using GRADE criteria (Table 2). For
the key outcome measures of long‐term pain intensity and
3.12 | Subgroup analysis
depression, there were 757 participants (12 RCTs) and 524 A subgroup analysis for trials that only included TMD stud-
participants (8 RCTs), respectively. ies showed similar significant effects on long‐term pain and
For other outcome measures that were not pooled, depression SMD −0.34 (−0.50, −0.19) and −0.33 (−0.51,
the quality of evidence was also high and significant ef- −0.15), and results could be pooled due to low heterogeneity
fects were observed for the effects of self‐management (Table 4).
 |

15322149, 2019, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejp.1358 by UNIVERSIDAD DE VALPARAISO, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
860    AGGARWAL et al.

FIGURE 5 Comparison—any self‐management intervention versus usual care outcome—depression long term (>3 months)

the main outcome measures (pain and depression) were high

3.13 | Components of self‐management
using GRADE scores. For other outcome measures, the qual-
Meta‐regression was conducted to test whether biofeedback ity of evidence was also high in GRADE despite the hetero-
component showed an additional treatment effect compared geneity observed for these outcomes in the meta‐analysis.
with no biofeedback. The outcomes of long‐term pain and Self‐management interventions therefore also showed sig-
depression were used to assess this effect. Of the 11 studies nificant improvement on activity interference and long‐term
reporting long‐term pain, five studies also used biofeedback muscle palpation pain.
in the intervention. The coefficient estimate from meta‐re- The descriptive analysis of studies and interventions used
gression for using biofeedback was (−0.16, 95% CI −0.48 to showed that all but two of the included studies were on TMD
0.17, p‐value = 0.360). Of the eight studies reporting long‐ and that self‐management interventions for chronic orofacial
term depression, three studies also used biofeedback in the pain (mainly TMD) include education, physical (jaw pos-
intervention. The coefficient estimate from meta‐regression ture relaxation) and psychosocial (cognitive, behavioural)
for using biofeedback was (−0.13, 95% CI −0.50 to 0.23, self‐regulation. Meta‐regression showed that biofeedback did
p‐value = 0.475). not provide additional contribution to effect size. Given that
some types of biofeedback, such as masseter EMG biofeed-
back, require additional expensive equipment, training and
4 | D IS C U SS ION
particularly time for patients to practise, further evaluation is
required on the value of biofeedback in self‐management of
4.1 | Summary of main results
chronic orofacial pain.
This systematic review has shown for the first time that there
is strong evidence to support the use of self‐management in-
terventions to improve long‐term outcomes for patients with
4.2 | Implications for management of
chronic orofacial pain and TMD. There were significant ef-
chronic orofacial pain
fects for improvement in long‐term pain and depression, the Overall, the components identified by the review map onto a
studies were at low risk of bias, and there were sufficient biopsychosocial intervention model involving both physical
numbers of studies that could be pooled to give an overall and psychological approaches to the management of chronic
treatment effect. The quality and certainty of evidence for orofacial pain (mainly TMD). This is not dissimilar to
 |

15322149, 2019, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejp.1358 by UNIVERSIDAD DE VALPARAISO, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
AGGARWAL et al.    861

TABLE 4 Subgroup analysis of effectiveness of self‐management for TMD alone

Outcomes Intervention Control No. of studies Pooled effect Heterogeneity (%)

Long‐term pain Combined self‐care Usual care 4 −0.46 (−0.72, −0.20) 41.5
biofeedback and CBT
Self‐care CBT Usual care 5 −0.28 (−0.47, −0.09) 0
All interventions Usual care 9 −0.34 (−0.50, −0.19) 10
Long‐term depression Combined self‐care Usual care 3 −0.41 (−0.68, −0.13) 26.7
biofeedback and CBT
Self‐care CBT Usual care 4 −0.28 (−0.51, −0.05) 16.3
All interventions Usual care 7 −0.33 (−0.51, −0.15) 12

approaches identified for management of chronic back pain components. Indeed, self‐reports of jaw parafunction, psycho-
(with which TMD co‐occurs) and which have been shown to social factors and reporting of other somatic symptoms have
be cost‐effective (Hill et al., 2011, 2013; Main, Sowden, Hill, been shown to be the strongest predictors of TMD the large
Watson, & Hay, 2012). prospective OPERRA study (Slade et al., 2007, 2016). These
Physical self‐regulation and education as active compo- risk factors lend themselves to the biopsychosocial approach
nents for TMD self‐management are supported by a Delphi identified by the findings of the current systematic review.
study. It showed that main components of a standard self‐ It is important to note that the trials included in the current
care programme of TMD were agreed to comprise educa- review were mainly on TMD. The physical self‐regulation
tion; self‐exercise; self‐massage; thermal therapy; dietary (jaw posture relaxation) component is therefore relevant to
advice and nutrition; and parafunctional behaviour (Durham, TMD alone rather than all facial pain subtypes as TMD is
Al‐Baghdadi et al., 2016). However, it did not include psy- commonly associated with parafunctional habits (Durham,
chological components which were shown to be integral in Al‐Baghdadi et al., 2016). Future research needs to explore
the management of TMD in our current systematic review. the effects of self‐management on all facial pain subtypes
Previous studies using a predominantly psychosocial ap- as per the study by Goldthorpe et al., (2017) and determine
proach (Goldthorpe et al., 2017) identified the need for phys- whether physical self‐regulation components are effective for
ical self‐regulation as an additional component. It was not other subtypes of chronic orofacial pain.
included in their patient manual, but recognized as an im-
portant component for management of patients in their trial.
Indeed, current recommendations for TMD management
4.3 | Implications for future research
(The European Pain Federation ) state that physiotherapy and The studies eligible for inclusion in this review were con-
pain management psychology can be useful. This is in agree- ducted in secondary care where patients had developed long‐
ment with the descriptive components of self‐management standing chronic orofacial pain. Given the effectiveness of
interventions identified in our review that show packages of self‐management in this group of patients, future studies
both physical and psychosocial components appear benefi- need to be conducted in primary care to explore whether
cial. Future research needs to explore how these approaches early intervention can improve outcome by preventing chro-
interact separately and/or combined in a single intervention. nicity. This certainly appears to be the case for early inter-
Indeed, this can have implications for pain management pro- vention in tertiary care (Gatchel et al., 2006). Future trials
grammes including those for orofacial pain which tend to also need to standardize outcome measures so that they can
address physical and psychosocial management separately, be comparable across trials. In the current review, we were
for example, by referral to a physiotherapist and/or clinical able to compare effectiveness for pain intensity and physi-
psychologist. It may be that such interventions delivered as cal and emotional functioning using outcomes available in
a package by skilled clinicians using a biopsychosocial ap- the included trials. Of these, only outcomes for pain inten-
proach may be more appropriate. Indeed, it has been found sity and emotional functioning (depression was the only out-
to be effective for physiotherapists to deliver a self‐manage- come across trials that was measured) could be pooled in the
ment package (comprising education, physical and psycho- meta‐analysis. Physical functioning represented by activity
social components) for biopsychosocial management of back interference could not be pooled due to high heterogeneity.
pain (Hill et al., 2011, 2013; Main et al., 2012). Both future Outcome measures for these domains (pain intensity, physi-
trials and current pain management programmes for chronic cal and emotional functioning) need to be standardized for
orofacial pain and TMD should prioritize a biopsychosocial future trials so that results can be compared across trials
approach that includes education, physical and psychosocial and pooled for meta‐analyses. Core outcome measures for
 |

15322149, 2019, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejp.1358 by UNIVERSIDAD DE VALPARAISO, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
862    AGGARWAL et al.

chronic pain in clinical trials have been clearly defined by assessment. The GRADE scoring showed that the certainty of
initiatives such as IMMPACT, and these would appear to be evidence was high for all the outcome measures. Therefore, the
an appropriate benchmark (Turk et al., 2008) for future trials quality of the evidence was high for trials included in the re-
on chronic orofacial pain and TMD. Indeed, there are several view. The component analysis showed that all trials included
dimensions of emotional functioning like fear of pain, cata- self‐management and physical and psychosocial self‐regulation.
strophizing and anxiety that are relevant to pain management Only six studies included biofeedback for which we were able
but due to the lack of homogeneity in their measurement we to conduct a meta‐regression. This showed that biofeedback
were unable to assess their effects. alone does not produce an effect in the meta‐regression model
Future work that explores the mechanisms by which these with very low residual heterogeneity (I2 = 7%).
interventions bring about change is also needed to inform The results of the current review update the findings of our
outcome measures. For example, Turner et al. (2007) exam- previous Cochrane systematic review (Aggarwal et al., 2011),
ined potential mediators, moderators and predictors of patient which showed that psychosocial interventions were effective
improvement with CBT. It was a novel study that examined in improving long‐term outcomes for patients with chronic
whether pre‐ to post‐treatment process variable changes medi- orofacial pain. However, that review failed to acknowledge
ated CBT effects on subsequent outcomes (Turner et al., 2007). the importance of the components within the interventions
The results showed that changes in perceived pain control and and grouped all interventions into a psychosocial group. In
self‐efficacy were important in explaining the treatment ef- addition, the evidence was weak as few studies were included
fects of CBT on the outcomes and should be considered in and an overall quality assessment of the quality of evidence
designing future behavioural interventions for TMD. A fur- was not conducted. Other systematic reviews in this area (Liu
ther study by Litt et al. (2010) also showed that somatization, et al., 2012; Randhawa et al., 2016) have suffered from meth-
self‐efficacy and readiness for treatment were significant odological shortcomings due to the limited amount of stud-
moderators. Work by our group assessed processes of engage- ies, lack of meta‐analysis and including interventions with
ment with a self‐management intervention, and this showed a number of disparate components all of which have led to
that key mechanisms of change centred around: identification inconclusive findings.
with the intervention; feeling believed and understood; obtain-
ing a plausible explanation for symptoms; degree of perceived
effort required to engage; acceptance of having a long‐term 4.5 | Potential biases
condition; and receiving demonstrative, positive feedback Given that the majority of interventions were delivered by
(Goldthorpe, Peters, Lovell, McGowan, & Aggarwal, 2016). a therapist, bias arising from therapeutic alliance related
These studies indicate that self‐efficacy, pain control, and un- to the quality of doctor–patient relationship may be pre-
derstanding and accepting the chronicity of the conditions are sent which can drive nonspecific effects (placebo effect in
important biopsychosocial predictors of patient improvement clinical practice and Hawthorne effect in clinical studies).
and should be incorporated into future interventions. This is Further, included studies were conducted in tertiary care
similar to other chronic pain conditions like chronic back pain settings which specialized in the management of chronic
whereby mediators like obtaining a plausible explanation for orofacial pain. This may affect the generalizability of the
symptoms and knowledge of the condition have led to the results as patients in these settings are likely to represent
development of public health approaches (Roland, Waddell, the more severe and intractable cases of chronic orofacial
Klaber Moffett, Burton, & Main, 2002; Waddell, Moffett, & pain and hence share common characteristics. Future tri-
Burton, 2004; Williams et al., 2009). Such approaches need als need to explore early management of chronic orofacial
to be considered for chronic orofacial pain and TMD, and in- pain in primary care using these interventions. While we
deed, specific self‐management advice can be included in both concluded that overall risk of bias was low and indeed tri-
primary care dental and medical practices. Over‐the‐counter als by Turner et al., (2006) and Shedden‐Mora et al., (2013)
pain relief such as nonsteroidal anti‐inflammatory drugs are were completely free of the domains of bias assessed, data
particularly useful for TMD pain and can be incorporated into permitting, we would have used sensitivity analyses to ex-
self‐management plans. This will avoid the need for costly in- amine the effect of concealed allocation, intention‐to‐treat
vasive and irreversible procedures like surgery, occlusal reha- analysis and blind outcome assessment on the overall esti-
bilitation and splints. mates of effect.

4.4 | Quality of the evidence in the

review and comparison to previous reviews 5 | CONCLUSIONS
The risk of bias pertaining to each item discussed in the re- The findings of this review provide strong evidence for
sults section was low for the majority of domains used in the the use of noninvasive self‐management interventions for
 |

15322149, 2019, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejp.1358 by UNIVERSIDAD DE VALPARAISO, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
AGGARWAL et al.    863

patients with chronic orofacial pain (mainly TMD). The myofascial pain: A systematic review. Journal of Dental Education,
components of these interventions included physical self‐ 69, 1242–1250.
regulation (jaw posture regulation), psychosocial (cognitive Bair, E., Gaynor, S., Slade, G. D., Ohrbach, R., Fillingim, R. B.,
Greenspan, J. D., … Maixner, W. (2016). Identification of clusters
and behavioural) self‐regulation and education. Future work
of individuals relevant to temporomandibular disorders and other
needs to prioritize the use of these interventions in early
chronic pain conditions: The OPPERA study. Pain, 157, 1266.
management of chronic orofacial pain including TMD. https://doi.org/10.1097/j.pain.0000000000000518
Barlow, J. (2001). How to use education as an intervention in osteoar-
thritis. Best Practice & Research Clinical Rheumatology, 15, 545–
ACKNOWLEDGEMENTS 558. https://doi.org/10.1053/berh.2001.0172
Barlow, J., Wright, C., Sheasby, J., Turner, A., & Hainsworth, J. (2002).
The authors would like to thank Professor Danielle van der Self‐management approaches for people with chronic conditions: A
Windt for her critique on the final draft which helped im- review. Patient Education and Counseling, 48, 177–187. https://doi.
prove the manuscript. org/10.1016/S0738-3991(02)00032-0
Beecroft, E., Durham, J., & Thomson, P. (2013). Retrospective exam-
ination of the healthcare 'journey' of chronic orofacial pain patients
CONFLICTS OF INTEREST referred to oral and maxillofacial surgery. British Dental Journal,
214, E12.
None declared. Bell, G., Smith, G., Rodgers, J., Flynn, R., & Malone, C. (2008). Patient
choice of primary care practitioner for orofacial symptoms. British
Dental Journal, 204, 669. https://doi.org/10.1038/sj.bdj.2008.523
AUTHOR CONTRIBUTIONS
Bergdahl, J., Anneroth, G., & Ferris, H. (1995). Cognitive therapy in the
Vishal R. Aggarwal contributed to study design, studies selec- treatment of patients with resistant burning mouth syndrome: A con-
tion, data extraction, risk of bias assessment, data analysis and trolled study. Journal of Oral Pathology & Medicine, 24, 213–215.
https://doi.org/10.1111/j.1600-0714.1995.tb01169.x
interpretation, and drafting the manuscript. Yu Fu involved in
Carlson, C. R., Bertrand, P. M., Ehrlich, A. D., Maxwell, A. W., &
systematic search in databases, studies selection, data extrac- Burton, R. G. (2001). Physical self‐regulation training for the man-
tion, risk of bias assessment, data interpretation and revising agement of temporomandibular disorders. Journal of Orofacial
the manuscript. Chris J. Main contributed to interpretation Pain, 15, 47–55.
of data analysis, critique of relevant literature and theory un- Crockett, D., Foreman, M., Alden, L., & Blasberg, B. (1986). A compar-
derpinning the introduction and discussion, and revising the ison of treatment modes in the management of myofascial pain dys-
manuscript. Jianhua WU involved in studies selection, data function syndrome. Biofeedback and Self‐Regulation, 11, 279–291.
https://doi.org/10.1007/BF01000164
extraction, risk of bias assessment, data analysis and interpre-
Department of Health (2001). The expert patient: A new approach to
tation, and revising the manuscript. All authors have read the
chronic disease management for the 21st century. London, UK:
article, discussed the results and commented on the manuscript. Department of Health.
Department of Health (2005). Self care–a real choice; self care sup-
port–a practical option. London, UK: Department of Health
R E F E R E NC E S
London.
Aggarwal, V., Fu, Y., Main, C., & Wu, J. (2018). Assessment of the Durham, J., Exley, C., Wassell, R., & Steele, J. (2007). 'Management is
relative effectiveness of self‐management interventions in the man- a black art'–professional ideologies with respect to temporomandib-
agement of adults with chronic orofacial pain: a systematic review, ular disorders. British Dental Journal, 202, E29.
meta‐analysis and meta‐regression. PROSPERO: International pro- Durham, J., Shen, J., Breckons, M., Steele, J., Araujo‐Soares, V., Exley,
spective register of systematic reviews. C., & Vale, L. (2016). Healthcare cost and impact of persistent oro-
Aggarwal, V. R., Lovell, K., Peters, S., Javidi, H., Joughin, A., & facial pain: The DEEP Study cohort. Journal of Dental Research,
Goldthorpe, J. (2011). Psychosocial interventions for the manage- 95, 1147–1154.
ment of chronic orofacial pain. Cochrane Database of Systematic Durham, J., Al‐Baghdadi, M., Baad‐Hansen, L., Breckons, M., Goulet,
Reviews, 11, CD008456. J.‐P., Lobbezoo, F., … Peck, C. (2016). Self‐management pro-
Aggarwal, V. R., McBeth, J., Zakrzewska, J. M., Lunt, M., & Macfarlane, grammes in temporomandibular disorders: Results from an interna-
G. J. (2006). The epidemiology of chronic syndromes that are fre- tional Delphi process. Journal of Oral Rehabilitation, 43, 929–936.
quently unexplained: Do they have common associated factors? Dworkin, S. F., Huggins, K. H., Wilson, L., Mancl, L., Turner, J.,
International Journal of Epidemiology, 35, 468–476. Massoth, D., … Truelove, E. (2002). A randomized clinical trial
Aggarwal, V., McBeth, J., Zakrzewska, J., Lunt, M., & Macfarlane, using research diagnostic criteria for temporomandibular disorders‐
G. (2008). Are reports of mechanical dysfunction in chronic oro‐ axis II to target clinic cases for a tailored self‐care TMD treatment
facial pain related to somatisation? A population based study. program. Journal of Orofacial Pain, 16, 48–63.
European Journal of Pain, 12, 501. https://doi.org/10.1016/j. Dworkin, S. F., Turner, J. A., Mancl, L., Wilson, L., Massoth, D.,
ejpain.2007.08.002 Huggins, K. H., … Truelove, E. (2002). A randomized clinical trial
Al‐Ani, Z., Gray, R. J., Davies, S. J., Sloan, P., & Glenny, A.‐M. (2005). of a tailored comprehensive care treatment program for temporo-
Stabilization splint therapy for the treatment of temporomandibular mandibular disorders. Journal of Orofacial Pain, 16, 259–276.
 |

15322149, 2019, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejp.1358 by UNIVERSIDAD DE VALPARAISO, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
864    AGGARWAL et al.

Dworkin, S. F., Turner, J. A., Wilson, L., Massoth, D., Whitney, C., Komiyama, O., Kawara, M., Arai, M., Asano, T., & Kobayashi, K. (1999).
Huggins, K. H., … Truelove, E. (1994). Brief group cognitive‐be- Posture correction as part of behavioural therapy in treatment of my-
havioral intervention for temporomandibular disorders. Pain, 59, ofascial pain with limited opening. Journal of Oral Rehabilitation,
175–187. https://doi.org/10.1016/0304-3959(94)90070-1 26, 428–435. https://doi.org/10.1046/j.1365-2842.1999.00412.x
Elrasheed, A., Worthington, H., Ariyaratnam, S., & Duxbury, A. (2004). Liberati, A., Altman, D. G., Tetzlaff, J., Mulrow, C., Gøtzsche, P. C.,
Opinions of UK specialists about terminology, diagnosis, and treat- Ioannidis, J. P., … Moher, D. (2009). The PRISMA statement for
ment of atypical facial pain: A survey. British Journal of Oral reporting systematic reviews and meta‐analyses of studies that eval-
and Maxillofacial Surgery, 42, 566–571. https://doi.org/10.1016/ uate health care interventions: Explanation and elaboration. PLoS
S0266-4356(04)00154-8 Medicine, 6, e1000100.
Ferrando, M., Galdón, M. J., Durá, E., Andreu, Y., Jiménez, Y., & Lipton, J., Ship, J., & Larach‐Robinson, D. (1993). Estimated prevalence
Poveda, R. (2012). Enhancing the efficacy of treatment for temporo- and distribution of reported orofacial pain in the United States. The
mandibular patients with muscular diagnosis through cognitive‐be- Journal of the American Dental Association, 124, 115–121. https://
havioral intervention, including hypnosis: A randomized study. Oral doi.org/10.14219/jada.archive.1993.0200
Surgery, Oral Medicine, Oral Pathology and Oral Radiology, 113, List, T., Axelsson, S. (2010). Management of TMD: Evidence from sys-
81–89. https://doi.org/10.1016/j.tripleo.2011.08.020 tematic reviews and meta‐analyses. Journal of Oral Rehabilitation
Gardea, M. A., Gatchel, R. J., & Mishra, K. D. (2001). Long‐term ef- 37, 430–451. https://doi.org/10.1111/j.1365-2842.2010.02089.x
ficacy of biobehavioral treatment of temporomandibular disorders. Litt, M. D., Shafer, D. M., Ibanez, C. R., Kreutzer, D. L., & Tawfik‐
Journal of Behavioral Medicine, 24, 341–359. Yonkers, Z. (2009). Momentary pain and coping in temporoman-
Gatchel, R. J., Stowell, A. W., Wildenstein, L., Riggs, R., & Ellis, E. dibular disorder pain: Exploring mechanisms of cognitive behav-
(2006). Efficacy of an early intervention for patients with acute tem- ioral treatment for chronic pain. Pain, 145, 160–168. https://doi.
poromandibular disorder–related pain: A one‐year outcome study. org/10.1016/j.pain.2009.06.003
The Journal of the American Dental Association, 137, 339–347. Litt, M. D., Shafer, D. M., & Kreutzer, D. L. (2010). Brief cognitive‐
https://doi.org/10.14219/jada.archive.2006.0183 behavioral treatment for TMD pain: Long‐term outcomes and mod-
Goldthorpe, J., Lovell, K., Peters, S., McGowan, L., & Aggarwal, V. erators of treatment. Pain, 151, 110–116. https://doi.org/10.1016/j.
(2016). Management of medically unexplained chronic orofacial pain.2010.06.030
pain: Justification for behavioural interventions. Journal of Sleep Liu, H., Liang, Q., Xiao, P., Jiao, H., Gao, Y., & Ahmetjiang, A. (2012). The
Medicine and Disorders, 3, 1038. effectiveness of cognitive‐behavioural therapy for temporomandibu-
Goldthorpe, J., Lovell, K., Peters, S., McGowan, L., Nemeth, I., Roberts, lar disorders: A systematic review. Journal of Oral Rehabilitation,
C., & Aggarwal, V. R. (2017). A pilot randomized controlled trial 39, 55–62. https://doi.org/10.1111/j.1365-2842.2011.02239.x
of a guided self‐help intervention to manage chronic orofacial pain. Macfarlane, T. V., Blinkhorn, A. S., Davies, R. M., Kincey, J.,
Journal of Oral & Facial Pain & Headache, 31, 61–71. https://doi. & Worthington, H. V. (2002). Oro‐facial pain in the com-
org/10.11607/ofph.1665 munity: Prevalence and associated impact. Community
Goldthorpe, J., Peters, S., Lovell, K., McGowan, L., & Aggarwal, V. Dentistry and Oral Epidemiology, 30, 52–60. https://doi.
(2016). 'I just wanted someone to tell me it wasn't all in my mind org/10.1034/j.1600-0528.2002.300108.x
and do something for me': Qualitative exploration of acceptability of Main, C., Sowden, G., Hill, J., Watson, P., & Hay, E. (2012). Integrating
a CBT based intervention to manage chronic orofacial pain. British physical and psychological approaches to treatment in low back
Dental Journal, 220, 459. pain: The development and content of the STarT Back trial's ‘high‐
Higgins, J. P., & Green, S. (2011). Cochrane handbook for systematic risk’ intervention (StarT Back; ISRCTN 37113406). Physiotherapy,
reviews of interventions. Version 5.1.0. The Cochrane Collaboration. 98, 110–116. https://doi.org/10.1016/j.physio.2011.03.003
Hill, J. C., Whitehurst, D. G., Lewis, M., Bryan, S., Dunn, K. M., Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., & PRISMA Group
Foster, N. E., … Somerville, S. (2011). Comparison of stratified (2009). Preferred reporting items for systematic reviews and meta‐
primary care management for low back pain with current best prac- analyses: The PRISMA statement. PLoS Medicine, 6, e1000097.
tice (STarT Back): A randomised controlled trial. The Lancet, 378, Nicholas, M. K., & Blyth, F. M. (2016). Are self‐management strategies
1560–1571. https://doi.org/10.1016/S0140-6736(11)60937-9 effective in chronic pain treatment? PainManagement, 6, 75–88.
Hill, J., Whitehurst, D., Lewis, M., Bryan, S., Dunn, K., Foster, N., Peters, S., Goldthorpe, J., McElroy, C., King, E., Javidi, H., Tickle,
… Somerville, S. (2013). A randomised controlled trial and eco- M., & Aggarwal, V. R. (2015). Managing chronic orofacial pain:
nomic evaluation of stratified primary care management for low A qualitative study of patients', doctors', and dentists' experiences.
back pain compared with current best practice: The STarT Back British Journal of Health Psychology, 20, 777–791. https://doi.
trial [ISRCTN37113406]. The Bone & Joint Journal, 95, 28–28. org/10.1111/bjhp.12141
Johansson, A., Unell, L., Carlsson, G. E., Söderfeldt, B., & Halling, A. Pfaffenrath, V., Rath, M., Pöllmann, W., & Keeser, W. (1993). Atypical
(2003). Gender difference in symptoms related to temporomandib- facial pain–application of the IHS criteria in a clinical sample.
ular disorders in a population of 50‐year‐old subjects. Journal of Cephalalgia, 13, 84–88. https://doi.org/10.1177/03331024930
Orofacial Pain, 17, 29–35. 13S1218
Kennedy, A., Bower, P., Reeves, D., Blakeman, T., Bowen, R., Chew‐ R Core Team (2013). R: A language and environment for statistical com-
Graham, C., … Gardner, C. (2013). Implementation of self man- puting. Vienna, Austria: R Foundation for Statistical Computing.
agement support for long term conditions in routine primary care ISBN 3-900051-07-0.
settings: Cluster randomised controlled trial. BMJ, 346, f2882. Randhawa, K., Bohay, R., Côté, P., van der Velde, G., Sutton, D., Wong,
Koh, H., & Robinson, P. (2003). Occlusal adjustment for treating and J. J., … Mior, S. (2016). The effectiveness of noninvasive inter-
preventing temporomandibular joint disorders. Cochrane Database ventions for temporomandibular disorders. The Clinical Journal of
of Systematic Reviews, 1, CD003812. Pain, 32, 260–278. https://doi.org/10.1097/AJP.0000000000000247
 |

15322149, 2019, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejp.1358 by UNIVERSIDAD DE VALPARAISO, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
AGGARWAL et al.    865

Roland, M., Waddell, G., Klaber Moffett, J., Burton, K., & Main, C. people with pain. Pain, 137, 276–285. https://doi.org/10.1016/j.
(2002). The back book: The best way to deal with back pain: Get pain.2007.09.002
back active. Norwich, UK: The Stationery Office. Turner, J. A., Holtzman, S., & Mancl, L. (2007). Mediators, moderators,
Shedden‐Mora, M. C., Weber, D., Neff, A., & Rief, W. (2013). and predictors of therapeutic change in cognitive–behavioral ther-
Biofeedback‐based cognitive‐behavioral treatment compared with apy for chronic pain. Pain, 127, 276–286. https://doi.org/10.1016/j.
occlusal splint for temporomandibular disorder: A randomized con- pain.2006.09.005
trolled trial. The Clinical Journal of Pain, 29, 1057–1065. https:// Turner, J. A., Mancl, L., & Aaron, L. A. (2006). Short‐and long‐term ef-
doi.org/10.1097/AJP.0b013e3182850559 ficacy of brief cognitive‐behavioral therapy for patients with chronic
Slade, G., Diatchenko, L., Bhalang, K., Sigurdsson, A., Fillingim, R., temporomandibular disorder pain: A randomized, controlled trial.
Belfer, I., … Maixner, W. (2007). Influence of psychological factors Pain, 121, 181–194. https://doi.org/10.1016/j.pain.2005.11.017
on risk of temporomandibular disorders. Journal of Dental Research, Turner, J. A., Whitney, C., Dworkin, S. F., Massoth, D., & Wilson, L.
86, 1120–1125. https://doi.org/10.1177/154405910708601119 (1995). Do changes in patient beliefs and coping strategies predict
Slade, G., Ohrbach, R., Greenspan, J., Fillingim, R., Bair, E., Sanders, A., temporomandibular disorder treatment outcomes? The Clinical
… Smith, S. (2016). Painful temporomandibular disorder: Decade Journal of Pain, 11, 177–188.
of discovery from OPPERA studies. Journal of Dental Research, Waddell, G., Moffett, J. K., & Burton, K. (2004). The neck book.
95, 1084–1092. https://doi.org/10.1177/0022034516653743 London, UK: The Stationery Office. ISBN-0-11-703321-9.
The European Pain Federation. Temporomandibular disorders. Retrieved Williams, N. H., Amoakwa, E., Burton, A. K., Hendry, M., Lewis, R.,
from http://www.orofacialpain.org.uk/downloads/Orofacial%20 Jones, J., … Wilkinson, C. (2009). The hip and knee book: Helping
pain%20patient%20leaflets/EYAP2013%20Patient%20Leaflet%20 you cope with osteoarthritis. London, UK: The Stationery Office.
TMD.pdf ISBN-9780117067295.
Townsend, D., Nicholson, R. A., Buenaver, L., Bush, F., & Gramling,
S. (2001). Use of habit reversal treatment for temporomandibular
pain in a minimal therapist contact format. Journal of Behavioural
Therapy and Experimental Psychiatry, 32, 221–239. SUPPORTING INFORMATION
Turk, D. C., Rudy, T. E., Kubinski, J. A., Zaki, H. S., & Greco, C.
Additional supporting information may be found online in
M. (1996). Dysfunctional patients with temporomandibular dis-
orders: Evaluating the efficacy of a tailored treatment protocol.
the Supporting Information section at the end of the article.  
Journal of Consulting and Clinical Psychology, 64, 139. https://doi.
org/10.1037/0022-006X.64.1.139
Turk, D. C., Zaki, H. S., & Rudy, T. E. (1993). Effects of intraoral appli- How to cite this article: Aggarwal VR, Fu Y, Main CJ,
ance and biofeedback/stress management alone and in combination Wu J. The effectiveness of self‐management
in treating pain and depression in patients with temporomandibular interventions in adults with chronic orofacial pain: A
disorders. Journal of Prosthetic Dentistry, 70, 158–164. https://doi. systematic review, meta‐analysis and meta‐regression.
org/10.1016/0022-3913(93)90012-D Eur J Pain. 2019;23:849–865. https://doi.org/10.1002/
Turk, D. C., Dworkin, R. H., Revicki, D., Harding, G., Burke, L. B., ejp.1358
Cella, D., … Hertz, S. (2008). Identifying important outcome do-
mains for chronic pain clinical trials: An IMMPACT survey of