Introduction
Music performance anxiety (MPA) is one of the most frequently reported disorders among musicians. The prevalence rate is estimated between 15% and 25% (Spahn et al., Reference Spahn, Richter and Altenmüller2011). Due to the International Classification of Diseases (ICD-10) (Dilling and Freyberger, Reference Dilling and Freyberger2017), it is coded as a specific phobia; in the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) (American Psychiatric Association, 2013), it is classified as a subtype of social anxiety disorder (performance only subtype). A consensus on its definition has not been reached yet (Kenny, Reference Kenny2011).
Musicians suffering from MPA have problems in performance situations, for example, in front of an audience or during orchestra rehearsals. They display physiological [most reported: tachycardia, sweating, tremor, dry mouth, shortness of breath (Hiner et al., Reference Hiner, Brandt, Katz, French and Beczkiewicz1987; Wesner et al., Reference Wesner, Noyes and Davis1990)], emotional (like panic and stress) and cognitive symptoms (e.g. self-doubt or expectation of failure), often leading to avoidance (not performing) and safety behaviour (e.g. alcohol, distraction). For most musicians, MPA is present directly before and during performances, while about 21% suffer from anticipatory anxiety days before the feared situation takes place (van Kemenade et al., Reference van Kemenade, van Son and van Heesch1995). There are varying degrees of MPA severity. Some musicians being most affected even decide to end their career. Although stress-related mental disorders (like depression and anxiety disorders) are frequently observed as psychiatric comorbidities (Kenny, Reference Kenny2011), only about 15% of musicians affected from MPA seek help (Wesner et al., Reference Wesner, Noyes and Davis1990). Compared to the general working population, musicians showed more symptoms of anxiety and depression in a Norwegian study (Vaag et al., Reference Vaag, Bjoerngaard and Bjerkeset2016).
There are different theories regarding the aetiology of anxiety disorders. Following the so-called ‘biopsychosocial model’ of anxiety disorders, there are biological, psychological and social factors contributing to the development of MPA (Bandelow et al., Reference Bandelow, Michaelis and Wedekind2017). According to Kenny (Reference Kenny2011), a special risk factor increasing the vulnerability for MPA might be a highly demanding environment that in the same time provides little support. Besides, the exposure to early and frequent (self-) assessments in a competitive setting is seen as a specific psychological vulnerability for MPA (Kenny, Reference Kenny2011).
Altogether, there are three reviews dealing with the treatment options of MPA. Nagel (Reference Nagel2010) selected studies researching cognitive behavioural therapy (CBT) and psychodynamic therapy to treat MPA and found evidence for the efficacy of CBT in MPA. However, other treatment options have not been considered. Another systematic review on treatments for MPA (professional musicians and students) describes significant positive effects on MPA and performance quality for different CBT techniques, such as behavioural training, cognitive restructuring, self-instruction in combination with progressive muscle relaxation (PMR) and self-instruction in combination with attention training (Kenny, Reference Kenny2005). In this review, only English publications were included. The third review (Brugués, Reference Brugués2011) particularly found β-blockers and CBT effective but declared further need for research as a conclusion. Main reasons for that were small sample sizes, no randomization and methodological problems. Given the fact that Nagel (Reference Nagel2010) only focused on selected studies, Kenny (Reference Kenny2005) restricted the review to English publications and Brugués (Reference Brugués2011) pronounced a lack of methodological satisfactory studies, there is a need for updating the current state of research regarding MPA.
Therefore, the aim of the present systematic review was to summarize previously published literature on prevalence, risk factors and treatment effects of MPA among professional musicians respecting all languages. Furthermore, the quality of evidence was critically evaluated, to address the problem pronounced by Brugués (Reference Brugués2011).
Methods
The methods of the systematic literature research followed the PRISMA statement (Liberati et al., Reference Liberati, Altman, Tetzlaff, Mulrow, Gotzsche, Ioannidis, Clarke, Devereaux, Kleijnen and Moher2009; Moher et al., Reference Moher, Liberati, Tetzlaff and Altman2009) and the recommendations of the Cochrane Collaboration (Green and Higgins, Reference Green, Higgins, Green and Higgins2011). Search methods and inclusion criteria were recorded in a protocol in advance.
Study types
Case reports, case–control studies, cohort studies, cross-sectional studies and intervention studies published in peer-reviewed journals were included in the review. Studies of all languages and countries of origin were considered and native speakers were recruited for all foreign-language articles. Last literature search was conducted on 3 February 2018 and no time limit was set.
Primary outcome parameters
Prevalence, incidence, risk factors and treatment strategies of MPA were of interest.
Search methods
Studies using the terms fear of performing, podium anxiety, stage fright and performance anxiety were included.
The search was carried out in two parts: an electronic and a manual search. Electronic search was conducted via search algorithms in the databases MEDLINE, EMBASE, CINAHL, PsycArticles, PsycInfo and ERIC. Manual search included two journals: ‘Medical Problems of Performing Artists’ and the German journal ‘Musikphysiologie und Musikermedizin’. Complete search algorithms can be found in the online Supplementary Appendices (Appendix 1: search algorithms).
Population and selection of studies
Firstly, studies were selected at title, secondly, at abstract and lastly, at full text level. Therefore, pre-defined inclusion criteria were determined:
– Population:
○ Musicians from at least 16 years of age
○ Mixed populations with children/adolescents/adults were only included when subgroups were analysed separately. Only data of musicians from at least 16 years of age were included in the review
○ Mixed populations with musicians/actors/dancers were only included when subgroups were analysed separately. Only data of musicians were included in the review
○ Students at music schools, universities or conservatories
○ Professional musicians as well as music teachers
○ Musicians with MPA
– Outcome:
Prevalence, incidence, risk factors and therapy
Full-text examination was carried out by a five-person team, consisting of medical staff of the Berlin Centre for Musicians' Medicine, the Institute of Social Medicine, Epidemiology and Health Economics and the Department of Psychiatry and Psychotherapy of the Charité – Universitätsmedizin Berlin. A consensus conference with the entire five-member team took place when the inclusion of a study was ambiguous.
Data extraction
The following information was extracted from the studies and entered into tables sorted by study type: (1) authors, (2) publication date, (3) populations studied, (4) sample sizes of intervention group and, if applicable, control group, (5) type of intervention, (6) randomization status, (7) outcomes and (8) results.
Regarding results, prevalence in percentages, effect sizes, correlations, mean values with standard deviations or errors, significance values, odds ratios or confidence intervals were of interest. If none of those parameters was provided, results were adopted as indicated in the particular study. With the exception of percentages, no calculations were made based on the provided values.
Quality rating
Quality assessment tools, ensuring a standardized evaluation of studies were developed for each of the different study types (cross-sectional study, cohort study, case–control study and controlled intervention study), with the exception of case reports.
For quality assessment, the following instruments of the National Heart, Lung, and Blood Institute (National Heart Lung and Blood Institute, Last Updated April 2014) were used: ‘Quality Assessment of Controlled Intervention Studies’, ‘Quality Assessment of Observational Cohort and Cross-Sectional Studies’ and ‘Quality Assessment of Before-After (Pre-Post) Studies With No Control Group’. To those assessment tools, further elements from the quality assessment instruments of the Critical Appraisal Skills Programme (Critical Appraisal Skills Programme (CASP), 2013) and the ‘Methodology Checklists’ of the Scottish Intercollegiate Guidelines Network (Scottish Intercollegiate Guidelines Network) were added.
A scoring system was created to systematically rate each study. Qualitative criteria were postulated dichotomously in the form of ‘yes’ or ‘no’ questions (e.g. ‘Was the research question or objective in this paper clearly stated?’). Questions were formulated in such a way for each inclusion criterion that a ‘yes’ always meant that criteria were met. To obtain a final rating for each study, the number of criteria rated with ‘no’ was subtracted from the number of criteria rated with ‘yes’. If a question was not applicable to a study, zero points were awarded (see online Supplementary Appendices 2–5). The possible overall scores differed for each evaluation instrument and thus for each study type. Controlled intervention studies could reach a maximum of 18 points, cross-sectional studies could reach a maximum of 15 points, cohort studies without control group a maximum of 15 points and case–control studies a maximum of 14 points.
Results
The search resulted in 43 articles, comprising 21 intervention studies, 11 cross-sectional studies, one cohort study and 10 case reports (see Fig. 1 for an overview of search results).
Fig. 1. Overview of search and selected studies.
Of the intervention studies, nine comprised an active control group (James et al., Reference James, Griffith, Pearson and Newbury1977; Pearson and Simpson, Reference Pearson and Simpson1978; Sweeney and Horan, Reference Sweeney and Horan1982; James and Savage, Reference James and Savage1984; Gates et al., Reference Gates, Saegert, Wilson, Johnson, Shepherd and Hearne1985; Gates and Montalbo, Reference Gates and Montalbo1987; Montello et al., Reference Montello, Coons and Kantor1990; Stanton, Reference Stanton1994; Wells et al., Reference Wells, Outhred, Heathers, Quintana and Kemp2012), nine a waiting list group or no treatment control group (Sweeney and Horan, Reference Sweeney and Horan1982; Nagel et al., Reference Nagel, Himle and Papsdorf1989; Montello et al., Reference Montello, Coons and Kantor1990; Valentine et al., Reference Valentine, Fitzgerald, Gorton, Hudson and Symonds1995; Chang et al., Reference Chang, Midlarsky and Lin2003; Khalsa and Cope, Reference Khalsa and Cope2006; Khalsa et al., Reference Khalsa, Shorter, Cope, Wyshak and Sklar2009; Bissonnette et al., Reference Bissonnette, Dube, Provencher and Moreno Sala2015; Spahn et al., Reference Spahn, Walther and Nusseck2016), three were without control group (Kim, Reference Kim2005; Stern et al., Reference Stern, Khalsa and Hofmann2012; Juncos et al., Reference Juncos, Heinrichs, Towle, Duffy, Grand, Morgan, Smith and Kalkus2017) and six compared different interventions (Sweeney and Horan, Reference Sweeney and Horan1982; Brodsky and Sloboda, Reference Brodsky and Sloboda1997; Hinz, Reference Hinz2005; Kim, Reference Kim2008; Khalsa et al., Reference Khalsa, Shorter, Cope, Wyshak and Sklar2009; Wells et al., Reference Wells, Outhred, Heathers, Quintana and Kemp2012). Some studies are listed several times because they included different kinds of control groups and/or interventions.
Quality ratings for cross-sectional studies ranged from −11 to 6 points (maximum 15 points), the cohort study reached 0 out of 16 points and intervention studies ranged from −4 to 11 points (maximum 18 points). For detailed results see Tables 1–3.
Table 1. Cohort study and cross-sectional studies
N, number of participants; M, mean; s.d., standard deviation; outcome and results: prevalence, risk factor or treatment of MPA; rating: quality rating of study; MPA, music performance anxiety. K-MPAI, Kenny Music Performance Anxiety Inventory (Kenny, Reference Kenny, Williamon, Pretty and Buck2009); TAI-G, ‘Prüfungsängstlichkeitsinventar’, modified (Brandner, Reference Brandner2001); STAI-S, State Anxiety Inventory (Spielberger et al., Reference Spielberger, Gorssuch, Lushene, Vagg and Jacobs1982).
Table 2. Intervention studies
N, number of participants; M, mean; s.d., standard deviation; s.e., standard error; outcome and results: prevalence, risk factor or treatment of MPA; rating: quality rating of study; MPA, music performance anxiety. AATS, adaptation of Achievement Anxiety Test Scale (Alpert and Haber, Reference Alpert and Haber1960); ACQ, Anxiety Control Questionnaire (Rapee et al.., Reference Rapee, Craske, Brown and Barlow1996); AD, Anxiety Differential (Husek and Alexander, Reference Husek and Alexander1963); AMPS, Appraisal of Music Performer's Stress (Brodsky and Sloboda, Reference Brodsky and Sloboda1997); FZA-F, Assessment of solo musical performance (Mills, Reference Mills1987); FZAQ-F, Fragebogen Zur AuftrittsQualität – Fremdeinschätzung: (Spahn, et al., Reference Spahn, Nusseck and Walther2013); KAB, ‘Kurzfragebogen zur aktuellen Belastung’ (Müller and Basler, Reference Müller and Basler1993); K-MPAI, Kenny Music Performance Anxiety Inventory (Kenny, Reference Kenny, Williamon, Pretty and Buck2009); MPAQ, The Music Performance Anxiety Questionnaire (Lehrer et al., Reference Lehrer, Goldman and Strommen1990); MPASS, Music performance anxiety self-statement scale (Craske et al., Reference Craske, Craig, Kendrick, Hersen and Bellack1988): positive outlook and task-focused attention; MPQ, Music Performance Quality Rating Form (Educational Testing Service, 1998); MPSS, Music Performance Stress Survey (Brodsky and Sloboda, Reference Brodsky and Sloboda1997); NMAC, Nowlis mood adjective checklist (Nowlis, Reference Nowlis, Tomkins and Izard1966); PAI, Performance Anxiety Inventory (Nagel et al., Reference Nagel, Himle and Papsdorf1981); PAQ, Performance Anxiety Questionnaire (Cox and Kenardy, Reference Cox and Kenardy1993); PARQ, Performance Anxiety Response Questionnaire (Appel, Reference Appel1976); POA, ‘Podiumsangst’ (Schröder and Liebelt, Reference Schröder and Liebelt1999); PRCP, Personal Report of Confidence as a Performer (Appel, Reference Appel1976); STAI-S, State Anxiety Inventory; STAI-T, Trait Anxiety Inventory (Spielberger et al., Reference Spielberger, Gorssuch, Lushene, Vagg and Jacobs1982).
Table 3. Case studies
N, number of participants; M, mean; s.d., standard deviation; outcome and results: prevalence, risk factor or treatment of MPA; rating: quality rating of study; MPA, music performance anxiety. ACQ, Anxiety Control Questionnaire (Rapee et al., Reference Rapee, Craske, Brown and Barlow1996); ASI, Anxiety Sensitivity Index (Reiss et al., Reference Reiss, Peterson, Gursky and McNally1986); BAI, Beck Anxiety Inventory (Beck and Steer, Reference Beck and Steer1993); K-MPAI, Kenny Music Performance Anxiety Inventory (Kenny, Reference Kenny, Williamon, Pretty and Buck2009); MPQ, Music Performance Quality Rating Form (Craske et al., Reference Craske, Craig, Kendrick, Hersen and Bellack1988); PAI, Performance Anxiety Inventory (Nagel et al., Reference Nagel, Himle and Papsdorf1981).
Prevalence and incidence
Prevalence of MPA was between 16.5% and 60% (Fishbein et al., Reference Fishbein, Middlestadt, Ottani, Straus and Ellis1988; Middlestadt, Reference Middlestadt1990; Wesner et al., Reference Wesner, Noyes and Davis1990; van Kemenade et al., Reference van Kemenade, van Son and van Heesch1995; Krawehl and Altenmüller, Reference Krawehl and Altenmüller2000; Medeiros Barbar et al., Reference Medeiros Barbar, de Souza Crippa and de Lima Osório2014; Sousa et al., Reference Sousa, Machado, Greten and Coimbra2016). About one-third of the examined musicians indicated MPA to be a severe problem [24% (Fishbein et al., Reference Fishbein, Middlestadt, Ottani, Straus and Ellis1988; Middlestadt, Reference Middlestadt1990), 38% (Krawehl and Altenmüller, Reference Krawehl and Altenmüller2000), 39% (Medeiros Barbar et al., Reference Medeiros Barbar, de Souza Crippa and de Lima Osório2014), 21.5% (Sousa et al., Reference Sousa, Machado, Greten and Coimbra2016), 16.5% (Wesner et al., Reference Wesner, Noyes and Davis1990) and 58.7% (van Kemenade et al., Reference van Kemenade, van Son and van Heesch1995)], whilst about 60% of them at least once experienced some kind of MPA in their career (Krawehl and Altenmüller, Reference Krawehl and Altenmüller2000). MPA was mostly assessed by self-reports of patients and not by professionals according to ICD or DSM criteria. For incidence, no studies were found.
Risk factors
The majority of studies reported different gender distribution. Women were found to be more frequently affected by MPA than men or displayed higher scores on questionnaires addressing MPA (Fishbein et al., Reference Fishbein, Middlestadt, Ottani, Straus and Ellis1988; Middlestadt, Reference Middlestadt1990; Wesner et al., Reference Wesner, Noyes and Davis1990; Hildebrandt et al., Reference Hildebrandt, Nubling and Candia2012; Kenny et al., Reference Kenny, Driscoll and Ackermann2014). However, some studies could not find differences between men and women (van Kemenade et al., Reference van Kemenade, van Son and van Heesch1995; Kenny et al., Reference Kenny, Davis and Oates2004; Khalsa et al., Reference Khalsa, Shorter, Cope, Wyshak and Sklar2009). Regarding age, younger musicians seem to be more affected from MPA than older musicians (Steptoe and Fidler, Reference Steptoe and Fidler1987; Fishbein et al., Reference Fishbein, Middlestadt, Ottani, Straus and Ellis1988; Middlestadt, Reference Middlestadt1990; Kenny et al., Reference Kenny, Driscoll and Ackermann2014). With an age older than about 45–50, there is a tendency to less MPA. Two studies found no relationship between age and MPA (Wesner et al., Reference Wesner, Noyes and Davis1990; van Kemenade et al., Reference van Kemenade, van Son and van Heesch1995). In three studies, there was a positive relationship between neuroticism and MPA (Steptoe and Fidler, Reference Steptoe and Fidler1987; Valentine et al., Reference Valentine, Fitzgerald, Gorton, Hudson and Symonds1995; Hodapp et al., Reference Hodapp, Langendörfer and Bongard2009) and one study found a negative relationship between self-efficacy and MPA (Hodapp et al., Reference Hodapp, Langendörfer and Bongard2009). Another study found that symphonic orchestra musicians suffer most from MPA (van Kemenade et al., Reference van Kemenade, van Son and van Heesch1995). Regarding the type of instrument, one study found brass players to be most affected by MPA (Fishbein et al., Reference Fishbein, Middlestadt, Ottani, Straus and Ellis1988; Middlestadt, Reference Middlestadt1990), but another study could not find any differences between instruments (Kenny et al., Reference Kenny, Driscoll and Ackermann2014).
Treatment
Different treatments for MPA have been investigated. Most often research was conducted for CBT, relaxation, exercise and β-blockers. Not all studies reported how and if diagnoses of MPA were determined, and in four studies, all participants were unaffected of mental illness (and therefore also free from MPA in sense of ICD or DSM). Results of those four studies are reported separately. To measure MPA, different questionnaires were used [e.g. Kenny Music Performance Anxiety Inventory (K-MPAI) (Kenny, Reference Kenny, Williamon, Pretty and Buck2009), Performance Anxiety Inventory (PAI) (Nagel et al., Reference Nagel, Himle and Papsdorf1981) or State Anxiety Inventory (STAI-S) (Spielberger et al., Reference Spielberger, Gorssuch, Lushene, Vagg and Jacobs1982)]. Furthermore, self-ratings of anxiety or performance quality (rated by a jury or self-rating of musicians) were used as measures of MPA. There were different types of control groups ranging from waiting list to non-active or active control groups. Some studies are considered several times as they compared different interventions with each other.
Psychological counselling
Psychological counselling in general was rated to be helpful in about 60–62% by patients in three cross-sectional studies (Fishbein et al., Reference Fishbein, Middlestadt, Ottani, Straus and Ellis1988; Middlestadt, Reference Middlestadt1990; Kenny et al., Reference Kenny, Driscoll and Ackermann2014).
CBT
Ten studies investigated CBT. Different CBT techniques were examined and all showed positive effects on MPA. There were two intervention studies without control group: Brodsky and Sloboda (Reference Brodsky and Sloboda1997) found a significant reduction of MPA after a CBT intervention alone or plus relaxation with or without music. Juncos et al. (Reference Juncos, Heinrichs, Towle, Duffy, Grand, Morgan, Smith and Kalkus2017) researched 12 sessions of acceptance and commitment therapy (ACT). The authors found a significant reduction of MPA. Ratings of performance quality did not change before and after therapy.
In three studies, CBT was compared to a non-active or a waiting list control group. One of those studies investigated six sessions of virtual reality exposure training (Bissonnette et al., Reference Bissonnette, Dube, Provencher and Moreno Sala2015). MPA reduced significantly after therapy and the quality of performances after therapy improved significantly compared to before and compared to waiting list control group performances. Another study comprised six group sessions of cognitive therapy, PMR and weekly individual temperature biofeedback (Nagel et al., Reference Nagel, Himle and Papsdorf1989). MPA reduced significantly after the intervention and the CBT group had less MPA than the control group. Sweeney and Horan (Reference Sweeney and Horan1982) examined six group sessions of cognitive restructuring with or without cue-controlled relaxation and found that anxiety symptoms seen on a videotape of a performance significantly decreased after therapy compared to before. For cognitive restructuring with cue-controlled relaxation, MPA decreased significantly. An active control group (musical analysis training) and a waiting list control did not change in their MPA level at the same time (Sweeney and Horan, Reference Sweeney and Horan1982).
There were five case reports investigating systematic desensitization, verbal self-directed positive statements training, cognitive restructuring and ACT (Norton et al., Reference Norton, MacLean and Wachna1978; Rider, Reference Rider1987; Salmon, Reference Salmon, Freeman and Dattilio1992; Lazarus and Abramovitz, Reference Lazarus and Abramovitz2004; Juncos and Markman, Reference Juncos and Markman2016). All patients reported that MPA improved or was cured after therapy.
Psychoanalytic and psychodynamic therapy
There were two case reports investigating psychoanalytic and psychodynamic therapy in MPA: Safirstein (Reference Safirstein1962) reported that after 200 h of psychoanalysis, the patient had no more anxiety and Kenny et al. (Reference Kenny, Arthey and Abbass2016) investigated 10 sessions of intensive short-term dynamic psychotherapy and found a positive effect on MPA.
Music therapy
Music therapy improvisation and desensitization were investigated in four trials (Montello et al., Reference Montello, Coons and Kantor1990; Kim, Reference Kim2005, Reference Kim2008). All showed a significant reduction of MPA after the intervention. In two studies, music therapy resulted in significantly lower MPA compared to a waiting list condition and an active control group (discussion of musical topics) [both Montello et al. (Reference Montello, Coons and Kantor1990)]. One study compared music therapy to PMR and imagery. Both interventions equally reduced MPA (Kim, Reference Kim2008).
β-Blockers
As a pharmacological treatment approach of MPA, the effect of β-blockers was quite intensively studied. In one cross-sectional study, β-blockers were rated to be helpful by 93% of musicians (Kenny et al., Reference Kenny, Driscoll and Ackermann2014). One randomized controlled trial tested 20 mg nadolol v. placebo (Gates and Montalbo, Reference Gates and Montalbo1987). There was no significant difference in anxiety before performances and in quality ratings of performances between both groups. Another randomized controlled trial tested 0, 20, 40 and 80 mg of nadolol compared to placebo. The active groups did not differ regarding the quality of performance and self-rated anxiety levels from placebo but heart rate was significantly lower in all nadolol conditions. Comparing only nadolol groups to each other, 20 mg of nadolol revealed the best effects on MPA (Gates et al., Reference Gates, Saegert, Wilson, Johnson, Shepherd and Hearne1985). An intervention study compared 25 mg of propranolol (administered once 1–1.5 h before the performance) and a few weeks of PMR (Hinz, Reference Hinz2005). Propranolol reduced heart rate more significantly than PMR. There was no significant difference in MPA between the two groups, but for the PMR group, there was a significantly lower disease-related burden after the performance.
Exercise
Within a cross-sectional study, aerobic exercise was found to be effective against MPA in 70% of musicians and yoga was rated as a helpful intervention in 58% (Fishbein et al., Reference Fishbein, Middlestadt, Ottani, Straus and Ellis1988; Middlestadt, Reference Middlestadt1990). Two intervention studies found significant reductions of MPA in questionnaires after yoga interventions (8 weeks/14 classes) (Khalsa and Cope, Reference Khalsa and Cope2006; Stern, Reference Stern2012). One intervention study compared a yoga lifestyle intervention, yoga only and a no treatment control group. Both yoga groups lead to significant improvements in MPA. But at no time point, yoga groups differed significantly to the no treatment control group regarding MPA (Khalsa et al., Reference Khalsa, Shorter, Cope, Wyshak and Sklar2009).
Hypnotherapeutical interventions
In cross-sectional studies, hypnosis was rated to be helpful against MPA by 60–76% of musicians (Fishbein et al., Reference Fishbein, Middlestadt, Ottani, Straus and Ellis1988; Middlestadt, Reference Middlestadt1990; Kenny et al., Reference Kenny, Driscoll and Ackermann2014). Stanton (Reference Stanton1994) investigated hypnotherapy compared to an active control group (talking about courses). He found a significant reduction in MPA (measured by a questionnaire) after the intervention and at follow-up. At follow-up, hypnotherapy resulted in a significantly lower MPA than the control group intervention.
Relaxation techniques
Relaxation techniques were rated to be helpful by 12% of musicians and deep breathing was rated to be helpful by 41–78% of the musicians in cross-sectional studies (Hiner et al., Reference Hiner, Brandt, Katz, French and Beczkiewicz1987; Kenny et al., Reference Kenny, Driscoll and Ackermann2014).
Two studies examined PMR: MPA severity decreased significantly after music-assisted PMR plus imagery and PMR was as effective as music therapy (Kim, Reference Kim2008), and as already reported (see β-blockers), PMR lead to a lower burden directly after performance compared to β-blockers. Chang et al. (Reference Chang, Midlarsky and Lin2003) found a significant reduction of MPA after a meditation intervention but no significant difference to a waiting list control group in MPA after the intervention.
Six group sessions of cue-controlled relaxation resulted in a significant reduction of MPA, whilst an active (musical analysis training) and waiting list control group did not change within the same period of time (Sweeney and Horan, Reference Sweeney and Horan1982).
Mixed interventions
Two case reports investigated several interventions: Stanton (Reference Stanton1993) found decreased MPA in three patients after two sessions of hypnotherapy, success imagery and rational emotive therapy. Abilgaard (Reference Abilgaard2007) investigated the Alexander technique, stress management with mental training and reactivation of hobbies in a patient with MPA and alcohol abuse (after alcohol detoxification). After the treatment, the patient was not suffering from MPA anymore.
Other interventions
In cross-sectional studies, the following interventions were rated to be helpful against MPA: Alexander technique by 47% (Fishbein et al., Reference Fishbein, Middlestadt, Ottani, Straus and Ellis1988; Middlestadt, Reference Middlestadt1990), focusing on performance by 69% (Hiner et al., Reference Hiner, Brandt, Katz, French and Beczkiewicz1987), mock performance practice by 91% (Kenny et al., Reference Kenny, Driscoll and Ackermann2014) and positive self-talk by 65% of musicians (Kenny et al., Reference Kenny, Driscoll and Ackermann2014).
Alexander Technique significantly improved MPA compared to a no treatment control group (Valentine et al., Reference Valentine, Fitzgerald, Gorton, Hudson and Symonds1995).
Psychodramatic treatment was described as helpful by the author of a case report (Moreno, Reference Moreno1946).
A seminar with video feedback, body awareness and cognitive strategies resulted in a significant improvement of MPA and rated performance compared to a no treatment control group (Spahn et al., Reference Spahn, Walther and Nusseck2016).
Studies with musicians not suffering from MPA
The four studies with participants being mostly free from MPA examined the influence of benzodiazepine, β-blockers and biofeedback on anxiety during performances.
The benzodiazepine diazepam (2 mg administered 1 h before a performance) had no significant influence on anxiety (self- and observer ratings during performances) compared to placebo (James and Savage, Reference James and Savage1984).
In contrast, two studies examining β-blockers showed some effect on anxiety: nadolol (40 mg administered 4 h before a performance) resulted in a better bow control in string players and a significantly lower pulse rate during the performance than placebo. Other observer ratings and self-ratings of performances did not differ between nadolol and placebo (James and Savage, Reference James and Savage1984). Oxprenolol (40 mg administered 90 min before the performance) significantly reduced self-reported anxiety ratings, pulse rate and blood pressure during performance compared to placebo (James et al., Reference James, Griffith, Pearson and Newbury1977; Pearson and Simpson, Reference Pearson and Simpson1978).
For low-frequency heart rate variability (HRV) biofeedback plus slow breathing, no significant differences could be found in anxiety compared to reading or just slow breathing. Taking both slow breathing groups together, high-frequency HRV and low-frequency/high-frequency ratio improved significantly during anxious anticipation compared to reading. Only for highly anxious participants, slow breathing groups reduced anxiety significantly compared to the control group (Wells et al., Reference Wells, Outhred, Heathers, Quintana and Kemp2012).
Discussion
The results of the present review impressively demonstrated that the research on MPA currently suffered from certain methodological weaknesses and is characterized by a high degree of heterogeneity.
First, study designs, term usage and surveyed result parameters differed widely. Mainly the terms MPA, performance anxiety or stage fright were used, without clarifying what exactly was meant by these terms. Usage ranged from some excitement while being on stage up to clinically relevant MPA diagnosis. A definition of MPA according to the criteria of the established diagnostic classification systems (ICD-10, DSM-IV or DSM-5), however, did not take place in any study. Therefore, it often remained unclear what was exactly measured and a direct comparison of studies was not possible.
All studies showed methodological deficiencies, as it is reflected in quality ratings. Especially the selection of participants was problematic. Some studies did not report diagnostic inclusion criteria and if musicians suffered from MPA or if healthy musicians were examined. It seems to be essential to first screen musicians with standardized instruments [e.g. IDCL-Checklists (Hiller et al., Reference Hiller, Zaudig and Mombour1994), SKID-I and II (Fydrich et al., Reference Fydrich, Renneberg, Schmitz and Wittchen1997; Wittchen et al., Reference Wittchen, Zaudig and Fydrich1997) or Composite International Diagnostic Interview (Robins et al., Reference Robins, Wing, Wittchen, Helzer, Babor, Burke, Farmer, Jablenski, Pickens, Regier, Satorius and Towle1988)] and report if the sample suffered from MPA, as defined in the ICD-10, DSM-IV or DSM-5. A criteria-based sample is needed to investigate prevalence, risk factors or treatment options. Moreover, valid instruments are needed to assess disorder-specific symptom severity. In the past, unspecific measurements of anxiety like the STAI-T or STAI-S (Spielberger et al., Reference Spielberger, Gorssuch, Lushene, Vagg and Jacobs1982) or self-developed questionnaires were used to assess (changes in) MPA, making it impossible to interpret the results or compare the outcomes to other studies. There are disorder-specific instruments available measuring MPA [like K-MPAI (Kenny, Reference Kenny, Williamon, Pretty and Buck2009), Performance Anxiety Questionnaire (PAQ) (Cox and Kenardy, Reference Cox and Kenardy1993) or PAI (Nagel et al., Reference Nagel, Himle and Papsdorf1981)] with the K-MPAI being validated (Chang et al., Reference Chang, Kenny and Burga2018). For German studies, there is a validated German version of the PAQ available: the ‘Bühnenangstfragebogen’ by Fehm et al. (Reference Fehm, Hille and Becker2002).
Regarding intervention studies, the quality of methods was limited due to several reasons: often participants were not randomized to different groups or could choose between different interventions or waiting list. Interventions were only rarely blinded making it possible for participants to expect some interventions, especially in comparison to waiting list controls, to be more effective. There is a need for randomized controlled trials with active control groups to research treatments for MPA. Furthermore, there should be no parallel treatments directly prior to or during the study period and comorbidities should be assessed and reported.
It was tried to derive some statements regarding prevalence, risk factors and treatment of MPA. Prevalence rate ranged from 16.5% to 60% and was mainly calculated by self-reports of musicians without any third-party assessment. When looking at those reports indicating MPA to be a severe problem for musicians and thus making it more possible to be clinically relevant, about one-third of the musicians seem to suffer from MPA. This goes in line with anxiety disorders being the most prevalent psychiatric disorders in Europe (Wittchen et al., Reference Wittchen, Jacobi, Rehm, Gustavsson, Svensson, Jonsson, Olesen, Allgulander, Alonso, Faravelli, Fratiglioni, Jennum, Lieb, Maercker, van Os, Preisig, Salvador-Carulla, Simon and Steinhausen2011).
Regarding risk factors, most studies reported women to be more affected than men, like it is the case with other anxiety disorders like agoraphobia, panic disorder, generalized anxiety disorder, specific phobia or social anxiety disorder (Wittchen et al., Reference Wittchen, Jacobi, Rehm, Gustavsson, Svensson, Jonsson, Olesen, Allgulander, Alonso, Faravelli, Fratiglioni, Jennum, Lieb, Maercker, van Os, Preisig, Salvador-Carulla, Simon and Steinhausen2011), indicating some greater vulnerability of women for anxiety disorders. Concerning age, the majority of studies reported older musicians to be less affected by MPA. The reason for that might be that very anxious musicians end their career due to MPA and engage in other professions. A further reason for this finding might be some kind of adaptation to the stressing factors of performances, making it easier to deal with those, when musicians have more experience. Younger musicians are more frequently exposed to uncertain situations (vulnerabilities) when being confronted with puberty, career entry or financial uncertainties. Compared to other anxiety disorders, the same pattern is visible. Older people show lower prevalence rates of anxiety disorders (Bandelow and Michaelis, Reference Bandelow and Michaelis2015). Other risk factors should be systematically investigated, such as the type of orchestra or the position (solo/tutti) within the instrument group. There was no study examining the risk factors in an appropriate way. Most findings were surveyed with cross-sectional studies and correlations. To receive more information about the risk factors, there is a need for longitudinal studies recording influencing factors and the presence of MPA.
Concerning treatment, the majority of studies examined CBT. Different CBT techniques were investigated (ACT, cognitive therapy, virtual reality exposure, systematic desensitization) and all resulted in the reductions of MPA after interventions. This goes in line with the positive effects of CBT in other anxiety disorders (Bandelow et al., Reference Bandelow, Lichte, Rudolf, Wiltink and Beutel2014). For future studies, it is important to compare CBT to active control groups (instead of waitlist or no control group) to better determine its efficacy as a treatment for MPA. Bandelow et al. (Reference Bandelow, Reitt, Röver, Michaelis, Görlich and Wedekind2015) showed that waitlists used as control group are less effective than a psychological placebo. Therefore, it is important to use an adequate (active) control group. An effective pharmacological treatment option only for vegetative symptoms of MPA was β-blockers. β-Blockers reduced the physiological symptoms of MPA, such as heart rate and tremor, but anxiety, negative cognitions and behaviour were not affected. The effects for music therapy improvisation and desensitization seem to be promising, but there is a need for further investigations in this field. A few studies examined the effects of yoga and relaxation on MPA, but mostly the effects were not better than the results of control groups, with the exception of one study (Sweeney and Horan, Reference Sweeney and Horan1982). For other interventions (psychodynamic therapy, Alexander technique or hypnotherapy), more research is needed to clarify if weak evidence for positive effects on MPA can be confirmed.
This review focuses on performance anxiety in musicians. In fact, there are other performing groups suffering from performance anxiety such as dancers or athletes. Nevertheless, we decided to focus on the musicians in order to not mix up different aspects and to gain specific results for this subgroup of interest.
Strengths and limitations
For several reasons, the results of the present review should be interpreted with caution. First, included studies used no consistent definition of MPA. Second, the criteria for the evaluation of studies changed in the past decades. In the present systematic review, the most recent update was used to develop quality assessment tools. Thereby, older studies were evaluated with probably stricter criteria than applicable by the time of publication of the studies. This might have resulted in an evaluation of those studies, being too strict. Third, many included studies showed methodological weakness limiting their informative value.
A strength of the present review is that a systematic and comprehensive search of the literature was carried out resulting in an update of published studies regarding performance anxiety. Furthermore, all studies were evaluated with a quality rating, making it possible to determine the methodological power of each study. In the field of MPA, it is the first time that the systematic quality ratings were applied.
Conclusion
Statements regarding prevalence, risk factors and treatment of MPA are limited. It is mostly unclear which criteria were used to diagnose MPA. A definition of the disease, consistent terminology and use of validated measurement instruments are essential for further research. Diagnostic uncertainty may explain the wide range of prevalence rates. Age and gender may be identified as risk factors and there is some evidence for effective treatments of MPA (especially for CBT and regarding vegetative symptoms also β-blockers). Cross-sectional studies, cohort studies and randomized controlled trials with clear diagnostic inclusion criteria and larger samples are needed in order to address a number of outstanding issues in this area of research.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0033291719001910.
Financial support
This work was supported by grants from the Friede Springer Stiftung (for Mumm) and the Bundesministerium für Bildung und Forschung (grant number: 01PL16032) and grants from Musicboard Berlin (for Professor Dr Schmidt).
Conflict of interest
None.
