Psychological treatment of dental anxiety among adults: a systematic review

Authors

  • Ulla Wide Boman,

    Corresponding author
    1. Public Dental Service, Region Västra Götaland, Clinic of Oral Medicine, Göteborg, Sweden
    • Department of Behavioral and Community Dentistry, Institute of Odontology, The Sahlgrenska Academy, University of Gothenburg, Göteborg, Sweden
    Search for more papers by this author
  • Viktor Carlsson,

    1. Department of Behavioral and Community Dentistry, Institute of Odontology, The Sahlgrenska Academy, University of Gothenburg, Göteborg, Sweden
    2. Public Dental Service, Region Västra Götaland, Clinic of Oral Medicine, Göteborg, Sweden
    Search for more papers by this author
  • Maria Westin,

    1. Public Dental Service, Region Västra Götaland, Clinic of Oral Medicine, SU Östra Hospital, Göteborg, Sweden
    Search for more papers by this author
  • Magnus Hakeberg

    1. Department of Behavioral and Community Dentistry, Institute of Odontology, The Sahlgrenska Academy, University of Gothenburg, Göteborg, Sweden
    2. Public Dental Service, Region Västra Götaland, Clinic of Oral Medicine, Göteborg, Sweden
    Search for more papers by this author

Ulla Wide Boman, Department of Behavioral and Community Dentistry, Institute of Odontology, The Sahlgrenska Academy, University of Gothenburg, PO Box 450, SE-405 30 Göteborg, Sweden

E-mail: ulla.wide-boman@odontologi.gu.se

Abstract

The aim was to investigate the efficacy of behavioural interventions as treatment of dental anxiety/phobia in adults, by conducting a systematic review of randomized controlled trials (RCTs). The inclusion criteria were defined according to the Patients, Interventions, Controls, Outcome (PICO) methodology. The study samples had documented dental anxiety, measured using validated scales [the Dental Anxiety Scale (DAS) or the Dental Fear Survey (DFS)], or fulfilled the psychiatric criteria for dental phobia. Behavioural interventions included were based on cognitive behavioural therapy (CBT)/behavioural therapy (BT), and control conditions were defined as information, sedation, general anaesthesia, and placebo/no treatment. The outcome variables were level of dental anxiety, acceptance of conventional dental treatment, dental treatability ratings, quality of life and oral health-related quality of life, and complications. This systematic review identified 10 RCT publications. Cognitive behavioural therapy/behavioural therapy resulted in a significant reduction in dental anxiety, as measured using the DAS (mean difference = −2.7), but the results were based on low quality of evidence. There was also some support that CBT/BT improves the patients' acceptance of dental treatment more than general anaesthesia does (low quality of evidence). Thus, there is evidence that behavioural interventions can help adults with dental anxiety/phobia; however, it is clear that more well-designed studies on the subject are needed.

The prevalence of dental anxiety is around 20% in different population-based studies [1-7]. Severe dental anxiety, including dental phobia and avoidance of dental care, has been reported to be approximately 5% in the general population [8, 9]. Dental anxiety is similar to specific phobias because of the pronounced avoidance tendencies and because it interferes in several ways with the afflicted person's life [10, 11]. Furthermore, the condition often presents serious problems to the providers of dental care [12]. Studies in population and clinical samples show that severe dental anxiety is related to poor dental health [13-19]. Psychological and social problems may also be present in individuals with long-standing dental anxiety. Negative consequences in social relationships have been reported, and such negative consequences may include embarrassment about poor oral health and reduced self-confidence, as well as increased frequency of sick-leave/absence from work and less involvement in social contexts [20-23]. Berggren [24] has presented a biopsychosocial vicious-circle model to describe how severe dental anxiety, together with its psychosocial concomitants, shapes and maintains severe dental anxiety over time.

The terms dental anxiety and dental phobia are often used interchangeably to denote individuals with severe dental anxiety and a behaviour pattern of avoidance of dental care. Severe dental anxiety is commonly defined by the use of cut-off points on validated self-report scales, while a phobia is defined according to criteria in psychiatric manuals [25].

Adult patients with severe dental anxiety may be referred to Special Care Dentistry Clinics/Clinics of Oral Medicine. These clinics usually provide adapted dental care, including different types of sedation (general anaesthesia, and nitrous oxide-, intravenous, and peroral sedation), and, to a varying degree, behavioural interventions for treating dental anxiety. These methods can be combined to customize individual treatment strategies. Treatment of dental anxiety has the aim to decrease dental anxiety and facilitate conventional dental care in the short- and long-term perspective. Behavioural interventions applied in the dental setting are often based on principles from learning, social learning, and cognitive theory [26]. Behavioural therapy (BT)/cognitive behavioural therapy (CBT) is the most accepted form of psychological treatment for anxiety related to particular situations and objects [27, 28]. Both behavioural interventions (such as exposure, systematic desensitization, and relaxation) and cognitive interventions (such as cognitive restructuring) fall under the general term of CBT, and, in clinical practice, they are often combined [27, 28].

In general, there is a need, within the health-care system (both in medical care and in dental care) to evaluate different types of treatment for diseases and conditions. The obvious reason is that the best treatments should be used routinely with regard to effectiveness, cost, and patient-related outcomes. Moreover, it is equally important to reduce, or even terminate, the use of ineffective treatment methods. The current standard when evaluating health-care techniques is systematic reviews applying strict protocols according to the Cochrane standard, for example, to scrutinize scientific papers that report on specific treatments and methods [29-31]. Studies conducted using a randomized controlled trial (RCT) design are considered to provide the most reliable evidence when investigating the efficacy of different interventions [32].

In a previous systematic review of behavioural interventions for dental anxiety in adults, published in 2004, the authors concluded that, despite extensive heterogeneity, the changes in dental anxiety represented medium to large effect sizes, and the effects were generally lasting [26]. This review was based on a literature search in two databases, PubMed and Psychlit, of papers published from 1966 to 2001 [26]. The review included different study designs, and no systematic evaluation of study quality was made. To the best of our knowledge, there are no recent systematic reviews/meta-analyses of the efficacy and/or effectiveness of behavioural interventions as a treatment mode for dental anxiety/phobia in adults.

The aim of the present study was to investigate the efficacy of behavioural interventions as treatment of dental anxiety/phobia in adults. This was carried out by conducting a systematic review applying a strict methodology and using the Health Technology Assessment (HTA)/Cochrane/the Grading of Recommendations Assessment, Development and Evaluation (GRADE) model [30, 31, 33], based on RCTs and with strict criteria to validate patient selection as well as the internal and external validity of the trials.

In this paper, we focussed on BT and CBT as broad descriptors of the type of behavioural interventions of interest to evaluate treatment of dental anxiety. To narrow the research question it was decided to evaluate whether behavioural interventions based on CBT/BT are effective for the treatment of dental phobia or dental anxiety, compared with information on dental care, sedation, experience of dental treatment under general anaesthesia, or placebo/no treatment.

Material and methods

This project comprises a systematic review and a meta-analysis.

Inclusion criteria

The inclusion criteria were defined according to the Patients, Interventions, Controls, Outcome (PICO) methodology [31]. The inclusion criteria were as follows: documented severe dental anxiety of the study population, as measured using validated psychometric scales or fulfilling the criteria for dental phobia according to psychiatric manuals [Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) [34] or the International Classification of Diseases, Tenth Revision (ICD-10) [35]]; intervention being based on CBT or BT, including exposure therapy, systematic desensitization, or relaxation therapy; the presence of a control condition, specified as information, sedation, general anaesthesia, or placebo/no treatment; the following specified outcome variables: level of dental anxiety (measured using validated scales), acceptance of conventional dental treatment (without sedation), dental treatability ratings, quality of life and oral health-related quality of life, and complications; and RCT and systematic review as the study design.

The validated self-report scales measuring dental anxiety that were used in the studies were the Dental Anxiety Scale (DAS) [36] and the Dental Fear Survey (DFS) [37]. The DAS consists of four items describing imaginary dental situations including ‘appointment tomorrow’ and different treatment situations. In this scale, responses are scored from 1 (no anxiety) to 5 (extreme anxiety), giving total scores of 4–20. Dental Anxiety Scale scores of 8–9 have been reported in the general population, and DAS scores of 13 or above have been reported among dental anxiety patients [38-40]. The DFS consists of 20 items covering anticipatory anxiety, physiological reactions, and situational anxiety. In this scale, responses are scored from 1 (no anxiety) to 5 (high intensity of anxiety), giving a total score of 20–100. The subscales anticipatory anxiety, physiological reactions, and situational anxiety have been confirmed by factor analysis [41]. Average DFS scores range from 35 to 45 in the general population and are above 60 in patients with extreme dental anxiety [40].

Literature search and selection

An electronic literature search was performed of the databases PubMed, The Cochrane Library, Embase, CINAHL, and PsycINFO, and of a number of HTA databases, from January 1970 to August 2011. English, Danish, Norwegian, or Swedish publications were accepted. The main MeSH terms used were ‘Dental Anxiety’, ‘Behaviour Therapy’, ‘Cognitive Therapy’, ‘Desensitisation, Psychological’, and ‘Relaxation Therapy’. The MeSH terms and the full search strategies in are given in Table S1. Reference lists of relevant articles were scrutinized for additional references. After removal of duplicates, a total of 990 articles was identified, of which 848 abstracts were excluded. After evaluating the remaining 142 articles, another 86 articles were excluded. Fifty-six articles were sent to the work group for assessment. Ten of these articles were included in the present analysis. (See Supporting Information for excluded references no. 54–98.) Two librarians at the regional HTA centre in Region Västra Götaland, Sweden, conducted the literature searches and excluded the abstracts in consultation with the authors. A flow chart of the literature search and the selection procedure is presented in Fig. S1.

Rating of quality of individual studies

The articles included in the report were critically appraised according to their scientific quality, using a checklist for RCTs from the Swedish Council on Technology Assessment in Health Care [33], which is in accordance with the principles used at other HTA centres globally and follows the CONSORT checklist [32]. The criteria covered external and internal validity and study precision [33], formulated in 35 questions covering study population (eligibility, and exclusion before randomization); allocation to intervention (randomization method and implementation); result of randomization (if groups were comparable on relevant variables); blinding (of patients, providers of treatment, and evaluators); dropouts (number, and reasons for dropout); compliance with treatment; measures of primary and secondary effects and complications (prespecified, adequately reported); results and precision (adequately reported, power analysis); and conflicts of interest. Each question was answered on a four-grade scale (acceptable, unclear, non-acceptable, and non-applicable), and an overall quality rating was then assigned to each study, as high, moderate or low quality, following the Swedish Council on Technology Assessment in Health Care checklist [33].

Rating of evidence across studies

The quality of the evidence obtained for each outcome measure specified for this literature review was rated according to GRADE [30, 31], taking into account the summarized quality of all studies providing results for the outcome. The quality of the evidence for each outcome was assessed at four levels [31]: high grade (image) was defined as a high level of confidence in the evidence reflecting the true effect, meaning that further research is very unlikely to change our confidence regarding the effect estimate. Moderate quality (image) was defined as moderate confidence in the evidence reflecting the true effect, and that further research could change our confidence in the effect estimate and may change the estimate. Low grade (image) was defined as low confidence in the evidence reflecting the true effect, with further research being likely to change the confidence in the effect estimate and also likely to change the estimate. Insufficient grade (image) means that evidence is either unavailable or does not permit estimation of an effect. The collective judgment of the strength of the evidence was based on the following criteria [31, 33]: study limitations; consistency; directness, precision; publication bias; magnitude of effect; and other important factors, such as dose–response gradient.

Each included article and the evidence for each specified outcome was assessed individually by at least three reviewers, and overall assessments were then agreed upon by all authors. Disagreements were solved by consensus.

Statistical analysis

Weighted mean differences were used for continuous outcomes. Statistical heterogeneity between trials was determined using the chi-square test and the I2 statistic (> 0.1, I2 < 25) [42]. A funnel plot was used to examine publication bias. Statistical analyses were performed using Review Manager (RevMan) [Computer program]. Version 5.1. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2011.)

Results

The systematic literature review identified 10 RCT publications, comprising seven different trials [43-52], that fulfilled the inclusion criteria. Five of the publications originated from two different RCTs. Berggren & Linde [43] and Berggren [44] compared BT with dental care under general anaesthesia. The BT was given by a psychologist, with each patient receiving an average of six sessions. Willumsen et al. [50, 51] and Willumsen & Vassend [52] compared three different treatments in 10 sessions performed at weekly intervals: CBT, applied relaxation, and nitrous oxide sedation. A CBT-trained dentist gave the treatment. de Jongh et al. [45] evaluated a one-session cognitive treatment method, information, and a waiting-list condition. The treatment was given by a dentist/psychologist. Gatchel [46] tested a 30-min videotaped dental anxiety-reduction programme with behavioural techniques vs. a placebo condition. Getka & Glass [47] compared BT and CBT with positive dental experience and with a waiting-list condition. The treatments comprised six sessions given by a graduate psychology student. Haukebø et al. [48] tested one- and five-session exposure treatments vs. a waiting-list condition. A dentist with special CBT training provided the treatment. Moses & Hollandsworth [49] tested stress inoculation, coping skills, education alone, and a waiting-list condition. The treatment was provided by a psychologist in one session lasting 2.5–3.5 h.

All 10 publications evaluated treatment effects on dental anxiety, using the DAS [43-52] (Table 1). Four of the publications also evaluated the treatment effect on dental anxiety using the DFS [48, 50-52] (Table 2). Four studies were of moderate quality [43, 44, 50, 51] and six of low quality [45-49, 52].

Table 1. Description of studies (randomized controlled trials) reporting on the outcome variable Dental Anxiety Scale
AuthorsYearCountryNumber of patientsWithdrawals and dropoutsStudy groupsaCommentsaQuality
InterventionControl
  1. a

    The Dental Anxiety Scale includes four items, each rated 1–5, with a higher rating denoting more dental anxiety.*The results are presented as mean sum scores or as mean sum scores ± SD. The mean sum score value ranged from 4 to 20, and the cut-off for dental anxiety was >12. Quality may vary according to outcome.

  2. AR, applied relaxation; BT, behavioural therapy; CBT, cognitive behavioural therapy; CS, coping skills; CT, cognitive therapy; DA, dental anxiety; E, education; GA, general anaesthesia; NO, nitrous oxide sedation; PDE, positive dental experience; SI, stress inoculation.

Berggren & Linde [43]1984Sweden

99

50/49

BT 8.3 ± 3.2

GA 10.8 ± 2.5

< 0.001

Before treatment:

BT 16.6 ± 3.1

GA 16.7 ± 2.4

Moderate
Berggren [44]1986Sweden

99

84

44/40

15

6/9

BT 9.1 ± 4.2

GA 12.2 ± 4.2

< 0.05

2-yr follow-up of Berggren & Linde [43]Moderate
Moses & Hollandsworth [49]1985USA

24

6/6/6/6

1 replaced

SI 15.6 ± 1.7

CS 16.3 ± 2.3

E 15.6 ± 1.7

Waiting list 17.5 ± 1.8

< 0.09

Before treatment:

SI 16.7 ± 1.1, CS 17.7 ± 1.7, E 17.3 ± 1.7, Waiting list 17.5 ± 1.3

Low
Gatchel [46]1986USA20?

Video programme: BT 13.5

6-month follow-up 13.3

Placebo 16.0

6-month follow-up 15.8

The 20 high-DA patients included in the analysis

SD not available

Low
Getka & Glass [47]1992USA41?

BT 9.6

CBT 9.6

Waiting list 13.5

PDE 14.9

P < 0.0001

Before treatment:

BT 15.4, CBT 14.9, Waiting list 14.9, PDE 15.8

SD not available

Low
de Jongh et al. [45]1995the Nether-lands

29

15/14

29

9

8

Cognitive intervention: CT 14.7 ± 2.8

1-yr follow-up 11.6 ± 3.2)

Information 17.8 ± 1.9

< 0.05

1-yr follow-up 11.4 ± 3.2

Before treatment:

CT 17.5 ± 1.5

Information 17.8 ± 2.2

1-month follow-up

No post-test for waiting-list condition

Low
Willumsen et al. [50]2001Norway

65

21/20/21

AR = 2

NO = 1

CT 9.3 ± 2.9 (= 21)

AR 8.1 ± 1.9 (= 20)

NO 10.0 ± 3.1 (= 21)

P > 0.05

Before treatment:

CT 17.0 ± 3.0, AR 17.8 ± 2.4, NO 17.0 ± 3.1

Moderate
Willumsen et al. [51]2001Norway62

CT = 3

AR = 1

NO = 2

CT 9.7 ± 3.5 (= 18)

AR 7.8 ± 3.2 (= 19)

NO 9.9 ± 3.7 (= 19)

P > 0.05

1-yr follow-up of Willumsen et al. [50]Moderate
Willumsen & Vassend [52]2003Norway62

CT = 9

AR = 5

NO = 7

CT 10.9 ± 4.3 (n = 12)

AR 9.9 ± 4.4 (n = 15)

NO 10.6 ± 3.9 (n = 14)

P > 0.05

5-yr follow-up of Willumsen et al. [50]Low
Haukebø et al. [48]2008Norway

40

10/10/20

40

1

9

1 (n = 10) or 5 (n = 9) sessions of exposure therapy 11.5 ± 3.0 (n = 19)

1-yr follow-up (n = 31)

1 session 10.4 ± 3.4

5 sessions 10.1 ± 3.2

Waiting list (n = 20)

16.6 ± 2.8

< 0.01

Before treatment:

Exposure therapy 17.2 ± 2.2

1 session 16.6 ± 2.0

5 sessions 16.6 ± 2.8

Waiting list 17.0 ± 2.8

Waiting-list group randomized to 1 or 5 sessions after 5 wk

Low
Table 2. Description of studies (randomized controlled trials) reporting on the outcome variable Dental Fear Survey
AuthorsYearCountryNumber of patientsWith-drawals and dropoutsStudy groupsaCommentsaQuality
InterventionControl
  1. a

    The Dental Fear Survey (DFS) includes 20 items, each rated 1–5, with a higher rating denoting more dental anxiety. *The results are presented as mean sum scores ± SD (range, 20–100), and the cut-off level for dental anxiety was a DFS score of >60. The DFS was also divided into three subscales (Behaviour, Arousal, and Situation, each scoring on a scale from 1 to 5), and the results of each subscale are presented as mean item scores SD. Quality may vary according to outcome.

  2. AR, applied relaxation; CT, cognitive therapy; NO, nitrous oxide sedation; NS, no significant difference.

  3. 1AR is statistically significant different from NO and CT.

  4. 2No significant difference between groups.

Willumsen et al. [50]2001Norway

65

21/20/21

3

CT (n = 21)

 Arousal 2.6 ± 1.0

 Situation 2.1 ± 0.7

AR (n = 20)

 Arousal 2.5 ± 1.1

 Situation 2.4 ± 1.0

NO (n = 21)

 Arousal 2.9 ± 0.9

 Situation 2.7 ± 0.9

NS (between groups)

Before treatment:

CT

 Arousal 3.6 ± 0.9

 Situation 4.0 ± 0.6

AR

 Arousal 3.7 ± 0.7

 Situation 4.0 ± 0.6

NO

 Arousal 3.7 ± 0.8

 Situation 4.2 ± 0.5

Two DFS subscales (mean item score)

Moderate
Willumsen et al. [51]2001Norway626

CT (n = 18)

 DFS 2.5 ± 0.81

 Behaviour 3.2 ± 1.61

 Arousal 2.5 ± 1.02

 Situation 2.5 ± 0.81

AR (n = 19)

 DFS 2.0 ± 0.71

 Behaviour 2.2 ± 1.41

 Arousal 2.2 ± 0.92

 Situation 1.8 ± 0.81

NO (n = 19)

 DFS 2.7 ± 1.01

 Behaviour 3.3 ± 1.21

 Arousal 2.6 ± 1.02

 Situation 2.6 ± 1.01

< 0.05

1-yr follow-up, Willumsen et al. [50]

DFS mean item score plus three subscales

Moderate
Willumsen &. Vassend [52]2003Norway62

19 (non-responders)

2 dropouts

CT (n = 12)

 DFS 2.8 ± 0.7

AR (n = 15)

 DFS 2.3 ± 0.9

NO (n = 14)

 DFS 2.7 ± 0.8

NS (between groups)

5 yr follow-up, Willumsen et al. [50]

DFS mean item score

Low
Haukebø et al. [48]2008Norway

40

10/10/20

11 (n = 10) or 5 (n = 9) sessions of exposure therapy 58.4 ± 14.1 (n = 19)

Waiting list (n = 20) 75.7 ± 8.8

< 0.01 (between groups)

Before treatment:

Exposure therapy 78.6 ± 7.7

Waiting list 75.6 ± 8.9

DFS mean sum score

Items 1 and 2 omitted in post-treatment assessments

Low

Effect on dental anxiety post-treatment

Meta-analyses of the post-treatment effect on dental anxiety measured using the DAS were performed for the five studies providing enough data [43, 45, 48-50], and showed a statistically significant decreased level of dental anxiety with a mean DAS score of 2.7 (< 0.0001) (Table 3). When the studies were subdivided according to type of control, decreased DAS scores were seen, with a mean of 2.0 for CBT/BT compared with anaesthesia/sedation (two RCTs, n = 161) (= 0.0006), and a mean of 3.3 DAS scores for CBT/BT compared with no treatment (three RCTs, n = 86) (P = 0.001) (Table 3). The two studies not included in the meta-analysis also showed lower dental anxiety levels for intervention compared with controls [46, 47].

Table 3. Behavioural therapy vs. controls: Dental Anxiety Scale (DAS) outcome after treatmentThumbnail image of

Two studies also reported a post-treatment effect on dental anxiety as measured using the DFS scale [48, 50]. One of the studies demonstrated a statistically significant reduction in dental anxiety [48]. No meta-analysis could be performed for outcomes measured using the DFS scale, as a result of inconsistencies of the reported data (different subscales and/or different versions of the scale were used).

Effect on dental anxiety at follow up

Six studies, from five trials, reported follow-up results on dental anxiety, measured using the DAS [44-47, 51, 52] (Table 1). Two studies were included in a meta-analysis (Table 4). These studies compared CBT/BT with anaesthesia/sedation [44, 51] (n = 79), and showed a decreased level of dental anxiety with a mean DAS score of 2.2 (P = 0.001) (Table 4).

Table 4. Behavioural therapy vs. anaesthesia/sedation: Dental Anxiety Scale (DAS) long term (1–2 yr) outcomeThumbnail image of

The paper of Willumsen & Vassend [52] was not included in the meta-analysis as the trial was already represented by the publication reporting follow-up data after 1 yr. This publication [52] found no difference in dental anxiety between intervention and control groups at the 5-yr follow-up. The remaining studies did not provide enough data to be included in the meta-analysis. However, de Jong et al. [45] reported no difference in dental anxiety between interventions and controls after 1 yr, whilst Getka & Glass [47] and Gatchel [46] reported decreased dental anxiety in the intervention groups compared with the control groups after 1 yr and 6 months, respectively.

Follow-up data using the DFS as a measure of dental anxiety found no differences between groups at follow up after 1 and 5 yr [51, 52].

Acceptance of conventional dental treatment measured using dental ‘treatability’ rating

Two studies reported the effect of CBT/BT on the acceptance of dental care, measured under controlled conditions and including a rating of patient behaviour [43, 48] (Table 5). One study of moderate quality reported a statistically significant positive effect of BT on the acceptance of conventional dental treatment, compared with general anaesthesia, with 80% successful dental sessions in the BT group and 53% in the general anaesthesia group (P = 0.009) [43]. The other study also reported a high proportion of success in the intervention group, but did not provide any control data for comparison [48]. There was insufficient data to perform a meta-analysis.

Table 5. Description of studies (randomized controlled trials) reporting on the outcome variable acceptance of dental treatment: dental treatability rating
AuthorsYearCountryNumber of patientsWith-drawals and dropoutsResultsCommentsQuality
InterventionControl  
  1. Quality may vary according to outcome.

  2. PICO, Patients, Interventions, Controls, Outcome.

Berggren & Linde [43]1984Sweden

99

50/49

 

Behavioural therapy

Successful 80% (n = 40)

General anaesthesia

Successful 53% (n = 26) P = 0.009 (chi-square test calculated from data)

Two sessions of dental treatment, rating of success or failure made by dentist using dentist rating scaleModerate
Haukebø et al. [48]2008Norway

40

10/10/20

11 (n = 9) or 5 (n = 10) sessions of exposure therapy. 92.3% success (completed all 14 steps of a behavioural test post-treatment)No control according to PICOBehavioural test, 14 steps, from entering room to filling cavity; success rated by dentistLow

Effects on quality of life/oral health-related quality of life and complications

These outcomes were not measured in the publications included.

Rating of quality of evidence according to GRADE

Table 6 presents a summary of the findings according to outcome, including grading of quality. There is some support for CBT/BT giving a clinically significant reduction in dental anxiety, as measured using the DAS; however, this conclusion is based on evidence of low quality (GRADE image). There is insufficient support for an effect of CBT/BT on dental anxiety, as measured using the DFS and based on evidence of very low quality (GRADE image). There is some support that behavioural therapy improves patients’ acceptance of conventional dental treatment more than does general anaesthesia, based on evidence of low quality (GRADE image).

Table 6. Summary of findings according to outcome, including grading of quality of evidence
Outcome variable/number of studiesStudy limitationsConsistencyDirectednessPrecisionPublication biasRelative effect (95% CI)Absolute effectQuality of evidence GRADE
  1. All included studies had a randomized controlled trial (RCT) design.

  2. BT, behavioural therapy; CBT, cognitive behavioural therapy; DAS, Dental Anxiety Scale; DFS, Dental Fear Survey; GRADE, the Grading of Recommendations Assessment, Development and Evaluation; NA, not applicable.

DAS

CBT vs. all controls (n = 5)

Serious limitations

Unclear randomization No blinding

Some inconsistency

Statistical heterogeneity

Some uncertainty

Advertisement recruitment in few studies

No imprecisionUnlikely−3.9 to −1.5Mean difference 2.7 less anxietyLow

DAS

CBT vs. anaesthesia/sedation (n = 2)

Serious limitations

Unclear randomization No blinding

No inconsistencyNo uncertaintyNo imprecisionUnlikely−3.2 to −0.9Mean difference 2.0 less anxietyLow

DAS long-term

CBT vs. anaesthesia/sedation (n = 2)

Serious limitations

Unclear randomization No blinding

Some inconsistency

Statistical heterogeneity

Some uncertainty

Advertisement recruitment in few studies

No imprecisionUnlikely−3.6 to 0.9Mean difference 2.3 less anxietyLow

DFS

CBT vs. anaesthesia/sedation, or waiting list (n = 4)

Serious limitations

Unclear randomization No blinding

Some inconsistency

Some uncertainty

All patients receive ‘good care’ and compassion

Serious imprecisionUnlikelyNANAVery low

Acceptance of dental treatment

BT vs. general anaesthesia (= 1)

Serious limitations

Unclear randomization No blinding

No inconsistencyNo uncertaintyNo imprecisionLikelyBT leads, approximately 1.5 times more often, to acceptance of conventional treatmentBT: 80% acceptance vs. general anaesthesia: 50% acceptanceLow

Discussion

The main finding of this systematic review and meta-analysis was that CBT/BT interventions may be effective in the treatment of dental anxiety/phobia. Treatments result in lower dental anxiety and in increased acceptance of dental treatment. However, the quality of evidence was low or very low, according to a strict evaluation scheme following the GRADE system in evidence-based medicine, which is used by most international Health Technology Assessment centres. The obvious reasons for these results are found in the scientific study designs of each publication included in the systematic review. Even though only RCTs were included, critical procedure violations, according to the accepted evaluation principles for such study designs, were found. Frequent problems with the studies were no blinding procedure, small sample sizes, and no information about the randomization procedure. Violation of such important features of an RCT results in low, or possibly moderate, study quality.

As far as we know, no previous systematic review, applying strict criteria according to the GRADE system, has been performed regarding the effect of behavioural treatments in dental anxiety/phobia. As many as 10 publications were found after the standardized exclusion procedure. Several of the included articles originated from the same study, typically reporting pre- to post-treatment effects and possibly follow-up data in a separate publication. All the included studies reported acceptable homogeneous patient groups with respect to high levels of dental anxiety, which is essential to be able to compare the outcome and efficacy of the different interventions. Moreover, and this is an important point when comparing the studies, the interventions differed somewhat with regard to specific behavioural treatment protocols, even if the core of the treatments was judged by the authors to belong to the behavioural/cognitive behavioural family of procedures, as far as could be deemed from the publication text. In addition, the control groups were exposed to different treatments/conditions, implicating the heterogeneity and thus decreasing the interpretability of the results. However, such weaknesses in systematic reviews are usually present when studies performed over a long time period, and from different health-care cultures and countries, are compared.

The occurrence of complications was not measured in the included publications, for any of the treatments or methodologies studied. As with all exposure-based behavioural treatments there is a risk of an increased anxiety level, as the treatment is anxiety-provoking initially. However, conventional treatment of dental-anxiety patients under general anaesthesia and/or sedation is associated with certain medical risks – the mortality rate is estimated at 1:100,000 of cases of general anaesthesia [53]. By including other types of study designs, such as observational and case–control studies, reports of complications may have been elucidated and reported.

The treatments evaluated in this study target individuals with severe dental anxiety/phobia, and it seems reasonable that this group should be given priority. But it may also be important to alleviate dental anxiety among those who are anxious, but still receive dental care on a regular basis without severe anxiety reactions. The gain would be to minimize a shift towards more negative dental-care behaviour, including avoidance of dental care.

There is a lack of well-designed studies, both RCTs and observational studies, as the present systematic review has identified significant knowledge gaps. There is a need for evaluations with clinically relevant outcome measures (acceptance of dental treatment, quality of life, oral-health-related quality of life, and dental status) as well as complications from treatments and long-term follow-up. Knowledge about the referral processes and care proceedings for patients with dental anxiety would also be valuable.

To conclude, this systematic review shows that behavioural interventions for the treatment of dental anxiety/phobia among adults decrease dental anxiety and increase the acceptance of dental care. This result indicates that behavioural interventions should be provided to adult patients with dental anxiety/phobia. However, the quality of the evidence was low or very low; thus, more well-designed studies, including a broader range of outcome measures, should be performed.

Acknowledgements

The authors would like to thank Maud Eriksson, Ann Liljegren, Petteri Sjögren, and Annika Strandell for their contributions to this work.

Conflicts of interest

The authors declare no conflicts of interest.

Ancillary