A systematic review of nonpharmacological interventions to reduce procedural anxiety among patients undergoing radiation therapy for cancer

Abstract Procedural anxiety is a concern for a number of patients undergoing radiation therapy. While procedural anxiety is often treated pharmacologically, there is a clinical need for effective alternative strategies for patients who are contraindicated from medication use, and those who prefer not to take unnecessary medications. Objectives The primary objective was to assess the efficacy of nonpharmacological interventions delivered to adults with cancer, in the radiation oncology department, just prior to, or during radiation therapy, in reducing levels of self‐reported procedural anxiety. The secondary objectives were to assess the efficacy of these interventions in reducing physiological symptoms of procedural anxiety and anxiety‐related treatment disruptions. Design Systematic review. Data Sources Electronic databases (MEDLINE, CINAHL, EMBASE, PsycINFO and Cochrane Central Register of Controlled Trials) were searched from inception up until February 2022. Inclusion Criteria Population: Adult patients with cancer undergoing external beam radiation therapy. Intervention: Nonpharmacological interventions delivered within the radiation therapy department. Comparison: standard care controls, or standard care plus an alternative intervention. Outcomes: level of self‐reported procedural anxiety (primary), physiological symptoms of anxiety (secondary) and measures of anxiety‐related treatment disruptions (secondary). Data Extraction and Analysis Two reviewers independently extracted data. A meta‐analysis was originally planned but deemed not feasible as the studies could not be confidently pooled for meta‐analysis, due to the variability in the interventions, study designs and the generally low number of studies. Therefore, a narrative synthesis is presented. Results Screening of 2363 records identified nine studies that met inclusion criteria: six studies of music interventions, two of video‐based patient education and one of aromatherapy. Overall, three studies received a global rating of strong methodological quality and low risk of bias. Three studies reported a significant effect of the intervention on reducing the primary outcome of self‐reported procedural anxiety: two music interventions (both strong methodological quality), and one video‐based patient education (moderate methodological quality). One of the studies (a music intervention) also reported a significant reduction in the secondary outcome of physiological symptoms of procedural anxiety (systolic blood pressure). Conclusions The evidence for nonpharmacological interventions delivered to adults with cancer just prior to, or during radiation therapy, in reducing levels of self‐reported procedural anxiety is limited, with very few well‐designed studies. There is a need for interventions for procedural anxiety during radiation therapy to be evaluated through rigorous randomised controlled trials.


| INTRODUCTION
In 2021, more than 150,000 people were diagnosed with cancer in Australia. 1 Approximately half of those diagnosed with cancer (48%) were expected to receive radiation therapy (RT), 2 with more than 74,000 courses of RT delivered in Australia in the 2018-2019 reporting year. 3T is an effective treatment method that can be used as a complete treatment course for cancer, or in combination with other treatment methods (e.g.surgery and/or chemotherapy). 4 The treatment delivery of external beam RT is noninvasive and painless, 5 although it does have several side effects, such as fatigue, nausea and skin rashes. 4dditionally, many patients experience significant symptoms of anxiety, particularly at the commencement of treatment. 6here are many factors that contribute to patient anxiety in oncology settings, including worry about the disease course, concern about side effects, and of course, fear about mortality. 7As well as this, many patients experience fear or anxiety relating to medical procedures, 7 which is referred to as 'procedural anxiety'.In this paper, we use the term 'procedural anxiety' to refer specifically to the acute anxiety experienced in relation to undergoing a healthcare-related procedure.This is distinct from anxiety related to concerns about what the results of a diagnostic procedure might show, which has been termed 'scanxiety'. 8In some cancer populations (e.g. head and neck cancer), procedural anxiety is seen in as many as 26% of patients. 9The fear or anxiety can occur during, or in anticipation of the procedure, and is generally transient. 10However, it is associated with acute distress and may result in behavioural disruption such as avoiding or terminating medical procedures. 10,11n RT, this is particularly common during the first few treatment sessions, 11 and persists for some patients through the course of treatment. 12cute anxiety, such as procedural anxiety, is primarily managed in RT settings pharmacologically, with benzodiazepines or nonbenzodiazepine anxiolytics. 7owever, pharmaceutical management of procedural anxiety presents its own issues.Delays in waiting for a radiation oncologist to prescribe and administer medication, as well as the time waiting for the medication to take effect (up to 30 min) 13 can be a lengthy process.Additionally, benzodiazepines in particular present problems for longer term use, making them unsuitable for patients experiencing persistent procedural anxiety. 5Further, patients who have been given benzodiazepines are advised not to drive while they are affected by the drug, inconveniencing patients and carers.There is also a subset of patients who have contraindications for benzodiazepines, including those with alcohol use disorder, those who are using opioids, as well as elderly interventions, study designs and the generally low number of studies.Therefore, a narrative synthesis is presented.
Results: Screening of 2363 records identified nine studies that met inclusion criteria: six studies of music interventions, two of video-based patient education and one of aromatherapy.Overall, three studies received a global rating of strong methodological quality and low risk of bias.Three studies reported a significant effect of the intervention on reducing the primary outcome of self-reported procedural anxiety: two music interventions (both strong methodological quality), and one video-based patient education (moderate methodological quality).One of the studies (a music intervention) also reported a significant reduction in the secondary outcome of physiological symptoms of procedural anxiety (systolic blood pressure).

Conclusions:
The evidence for nonpharmacological interventions delivered to adults with cancer just prior to, or during radiation therapy, in reducing levels of self-reported procedural anxiety is limited, with very few well-designed studies.
There is a need for interventions for procedural anxiety during radiation therapy to be evaluated through rigorous randomised controlled trials.

K E Y W O R D S
anxiety, oncology, procedural anxiety, procedure-related anxiety, psycho-oncology, radiation oncology, radiation therapy, radiotherapy patients. 7,13Furthermore, many patients report a strong preference to avoid medication where possible. 9n addition to pharmacological treatment, studies have investigated nonpharmacological interventions, including music listening, [14][15][16][17][18][19] education 20,21 and aromatherapy, 22 to reduce procedural anxiety in patients undergoing RT.Radiation oncology staff routinely provide informal support to patients experiencing procedural anxiety, although the nature and prevalence of this support has not, to our knowledge, been formally documented.In 2018, Nunns and colleagues 23 performed a systematic review and metaanalysis of trials of nonpharmacological interventions aiming to reduce procedural anxiety in paediatric patients undergoing treatment for cancer. 23Included in the review were studies trialling hypnosis, [24][25][26][27][28][29][30] distraction (interactive CD-ROM, 31 heated pillow, 32 listening to music, 33 virtual reality, 34 general distraction, 30,34 games or books 35 and an interactive device), 36 cognitive behaviour therapy 27 and music therapy. 37The review reported some promising findings in the studies of hypnosis, [27][28][29]38 however, cautioned that the studies were primarily conducted by a single research group.
We are not aware of any previous systematic review of nonpharmacological interventions aiming to reduce procedural anxiety among adult oncology patients.This is despite the high prevalence of procedural anxiety, its potential implications (e.g.treatment interruption and patient distress) and the increasing number of studies examining nonpharmacological interventions to address procedural anxiety.A summary of the evidence is needed in order to guide research and clinical practice.The purpose of this review is to summarise the evidence of nonpharmacological interventions targeting procedural anxiety in patients undergoing RT.To provide healthcare providers with maximum utility in managing patient anxiety, this review will focus on interventions that can be feasibly integrated into current procedures within the RT department (i.e.interventions that do not require additional appointments, or specialised external staff).It endeavours to address the needs of all cancer care centres, including those lacking access to psycho-oncology services as well as those that face substantial wait-lists. 39

| Objectives
The primary objective of this review was to: i Assess the efficacy of nonpharmacological interventions delivered to adult patients with cancer, in the radiation oncology department, just prior to, or during RT, in reducing levels of self-reported procedural anxiety.
The secondary objectives of this review were to assess the efficacy of nonpharmacological interventions delivered to adult patients with cancer just prior to, or during RT, in reducing additional measures of procedural anxiety: ii Physiological symptoms of procedural anxiety iii Anxiety-related treatment completions and duration

| METHODS AND ANALYSIS
This review is registered within the PROSPERO database (registration number CRD42019112941) and is being reported in accordance with guidance provided in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. 40The methods have been reported in more detail previously. 41

| Inclusion criteria-study characteristics
Due to the lack of clarity and use of the specific term 'procedural anxiety' in the literature, for the purposes of this review, studies were deemed to be addressing procedural anxiety if they aimed to: address anxiety; the intervention occurred during, or just prior to RT, and within the RT department; and a standardised and valid measure of anxiety was used (e.g. the State Trait Anxiety Inventory (STAI) State subscale (STAI-S), which is commonly used to measure procedural anxiety). 23

| Participants
Included studies were those involving adults with cancer currently undergoing, or about to undergo external beam RT, with or without concurrent chemotherapy.Studies involving only paediatric patients (under 18 years) or carers only were excluded.Participants undergoing brachytherapy were also excluded, as the procedure for brachytherapy was considered too different to external beam RT to generalise any findings.

| Interventions
The primary criterion for the inclusion of nonpharmacological interventions was feasibility for implementation in real-world clinical settings.The criteria were based on the clinical expertise of the authorship team (particularly the senior radiation therapist (SO)) and included interventions which could be readily implemented with little notice (as anxiety is often not identified prior to treatment) 11 and without additional departmental resources.These criteria were applied to focus on scalable interventions with maximum potential for translation into routine clinical practice.
Eligible interventions were those that could be delivered within the department of radiation oncology, at an existing appointment and able to be delivered by any healthcare provider within the RT department (i.e. did not require specialised training, such as psychological interventions or alternative therapies).We excluded interventions that: were delivered by licenced mental health providers; were delivered outside of the RT department; required an additional appointment (which included group interventions); and interventions that required extensive additional training for the healthcare provider to deliver (for example yoga, reiki, hypnotherapy and music therapy that included tailored music delivered by a music therapist).Interventions that exceeded 5 min were excluded, unless they were minimally resource intensive.For example, music listening for 15 min in the waiting room, was included, as participants were able to do this alone or alongside treatment (if music was delivered during treatment sessions).However, an education session delivered by a clinician of the same duration was deemed 'not able to be delivered in usual appointments', as this type of intervention requires additional resources from staff.

| Comparators
Studies of usual care with no intervention controls, or usual care plus an alternate intervention comparison group(s) were included.Studies without a comparison group were excluded.

| Outcomes
Studies using any of the following outcome measures were included in this review.Studies that included any of the secondary outcomes did not need to include any of the primary outcomes to be considered eligible for inclusion.

| Study selection
There was no restriction on the year of publication or language.Studies with no full text available were excluded.Articles that did not report original data (reviews, editorials and opinion articles) were also excluded.

| Data extraction
Two authors (EF, and either LM or TH) independently extracted the data using a prepiloted data extraction form, which was based on recommendations by the Cochrane Handbook for Systematic Reviews of Interventions. 48Discrepancies between the reviewers were discussed until a consensus was reached.

| Data items
The following information was extracted:

| Methodological quality and assessment of bias
Two reviewers (EF, and either LM or TH) independently reviewed the included studies and assessed the methodological quality and risk of bias using the Effective Public Health Practice Project Quality Assessment Tool (EPHPP). 49,50The adequacy of six domains were assessed.The EPHPP tool, which is suitable for evaluating randomised and nonrandomised designs (e.g.pre-post and case-control), 49 has been reported to have both content and construct validity, and acceptable inter-rater reliability. 50,51Regular meetings took place to discuss discrepancies until a consensus was reached.
The two reviewers also reviewed the included studies to assess reporting according to the TiDieR checklist.

| Data analysis
Due to the heterogeneity in the intervention types, we did not pool all included studies for a meta-analysis.We also did not pool studies with the same intervention type because of the small number of available studies and heterogeneity in study designs and outcome measurements.Therefore, findings are reported narratively.
In line with recommendations by the Synthesis Without Meta-analysis (SWiM) in systematic reviews: reporting guideline 52 we summarised the findings of both primary and secondary outcomes by intervention type, namely: music (with two subcategories: self-selected music interventions and predetermined music interventions); videobased patient education; and aromatherapy.
Results for eligible studies are also reported in a Forest plot (see Figure 2).As the studies included different outcome measures, standardised mean differences (SMDs) were estimated.

| Grading the strength of evidence
The overall quality of evidence was assessed using the GRADE (Grades of Recommendation, Assessment, Development and Evaluation) approach, 53 as recommended by the Cochrane Handbook for Systematic Reviews of Interventions. 48The primary outcomes were graded by two authors (EF and TH) as 'high', 'moderate', 'low' or 'very low', according to the published definitions.Discrepancies between the reviewers were discussed until a consensus was reached.

| Included studies
After duplicates were removed, 2363 citations were screened for eligibility.Nine papers reporting results of nine studies met inclusion criteria and are included in this review, with a total sample of 995 participants.The full details of papers identified, screened and included and presented in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram 54 (Figure 1).
The primary reasons for exclusion were as follows: interventions that were unable to be delivered in usual appointments (n = 38) or unable to be delivered by any healthcare provider (n = 10); studies reported in conference abstracts that were not published in a subsequent full report (n = 27); ineligible study design (n = 10); ineligible patient population (n = 8); and not delivered in the RT department (n = 8).

| Study characteristics
Study characteristics are presented in Table 1.Most studies were conducted in the United States (three studies), 17,19,21 two studies were conducted in Australia, 16,22 two in Turkey, 15,20 one in Taiwan 14 and one in Italy. 18Included papers were published between 2001 and 2022.

| Participants
Of the included studies, five included patients with any type of cancer who were scheduled to receive a course of RT, 14,16,17,20,22 two studies were constrained to patients with breast cancer, 15,18 one included patients with head and neck malignancies 21 and one included patients with pelvic or abdominal malignancies. 19Participant details including age and sex are included in Table 1.

Music
The music interventions could be broadly divided into those using self-selected (participant chosen) music (n = 4) 14,16,17,19 and those using predetermined (investigator chosen) music selection (n = 2). 15,18Interventions also varied in whether the music intervention was delivered prior to (n = 2) 14,18 or during (n = 4) [15][16][17]19 RT; over one session (n = 3) 14,16,17 or many (n = 3) 15,18,19 ; and via headphones (n = 4) 14,15,18,19 or speakers (n = 2).16,17 Self-selected music interventions.Chen et al 14 conducted a quasi-experimental study (n = 200). Participnts randomised to the music group were given an opportunity to choose from a selection of old songs (soft, slow-paced melodic music, 60-80 beats per minute), and then A further three studies trialled music interventions that involved some level of music self-selection.However, in contrast to Chen and colleagues, 14 all three interventions were delivered in the RT treatment bunker, during treatment delivery.Two of the three studies 16,17 conducted randomised trials of self-selected music during the first RT session only. In both rials the music was delivered via the speakers within the treatment bunker.In the study conducted by O'Steen and colleagues 17 (N = 102), participants randomised to the intervention group selected their preferred genre of music from a directory of prerecorded music, while intervention participants in the study conducted by O'Callaghan and colleagues 16 (N = 100) were asked to bring their own music. The controlgroups in both trials received no intervention.O'Steen and colleagues 17 reported that while there was a 16% reduction in the music group compared to a 10% reduction in selfreported anxiety in the control group, the difference between groups was not significant. Simily, O'Callaghan reported no significant difference in anxiety reduction between the groups, however, the authors reported that participants in the music group indicated a preference for music at future RT appointments statistically significantly more often than control participants.The third study, conducted by Smith and colleagues 19 (experimental, longitudinal, random assignment trial N = 42) trialled a music listening intervention delivered during the RT simulation appointment, as well as daily treatment sessions.Participants in the experimental    group listened to music of their choice from a selection of music provided by the researchers. Paticipants in the control group did not listen to music and received the standard care of the department.In contrast to the trials conducted by O'Steen 17 and O'Callaghan, 16 the music listening intervention was delivered to participants through headphones.The study found no significant effect on anxiety in either group.

Predetermined music interventions.
The two trials of predetermined music selection were conducted by Karadag and colleagues, 15 (randomised controlled trial, N = 60) and Raglio and colleagues 18 (randomised controlled pilot study, N = 60).In the study conducted by Karadag and colleagues, 15 intervention participants received an MP3 player with headphones to listen to 19 Trio Sonatas, a piece composed by Bach, with a duration of 20-40 min, during each radiation treatment for a period of 5 weeks.Participants randomised to the control group received no intervention.The study found a statistically significant reduction in anxiety (p = <0.001) in the music group (mean score decreased by 2.6 on the HADS Anxiety subscale, HADS-A) compared with the control group (mean score increased by 1.0) after the intervention (at 5 weeks) with a moderate effect size of 0.58.The findings also support a clinically meaningful reduction in anxiety of 2.6 on the primary outcome measure (HADS-A), a reduction that exceeds the minimum clinically important difference of 1.7. 56Additionally, the mean score for those in the intervention group was no longer in the clinical range (7 or more 57 ) following the exposure to the intervention (5.20), compared to those in the control group who remained in the clinical range (8.56).Raglio and colleagues 18 trialled two different music interventions: a piece produced by Melomics-Health (which produces music for therapeutic purposes), and the other group received individualised music listening.A third group (control) received no music intervention.However, the individualised music listening intervention was deemed ineligible for this review due to the involvement of a music therapist, who worked with participants to build their personalised playlist.These data were excluded from our review, but the Melomics-Health intervention (and control group) data are reported here, as that part of the study meets inclusion criteria.Participants randomised to the Melomics-Health group were instructed to sit in a quiet room in the RT department prior to RT simulation in the first 5 treatment sessions, where they listened to 15 min of music through headphones.The study found no difference between the Melomics-Health group and the no music control group on self-reported anxiety.

Video-based patient education
Two randomised trials examined the effect of videobased patient education on patients' procedural anxiety.Esen and colleagues 20 (N = 40) trialled a 2-3 min video showing patients what to expect in treatment.The video described the simulation process with explanations about the reason for immobilisers, how the patient may feel and the movements of the treatment machine.Similarly, Koth and colleagues 21 (n = 78) trialled a 5 min educational video with similar content.The video included a demonstration of making the thermoplastic mask, CT simulation, radiation planning and the delivery of radiation treatments, including undergoing a mock treatment delivery and explaining the experience.The control groups in both trials received standard education.Esen and colleagues 20 found a significant reduction (p < 0.001) in anxiety after viewing the video-based patient education, and after the first treatment session when compared to the control group (mean difference of 10 points on the STAI-S).Koth and colleagues 21 reported a significant reduction in one individual survey item of the STAI (adults short form Y-1) ('I am worried') in the intervention group, after viewing the education material.However, there was no significant difference in the total mean anxiety score between the two groups.

Aromatherapy
Graham et al 22 conducted a placebo-controlled doubleblind randomised trial (N = 313) to determine whether aromatherapy reduced anxiety during radiation therapy.Participants were randomised to receive a fragrant placebo, a nonfragrant placebo or pure essential oils (lavender, bergamot and cedarwood).Three drops of oil were applied to a paper bib that participants wore for the duration of their treatment each day.The study found a significant reduction in anxiety (p = 0.04) in the nonfragrant placebo group compared to the essential oils and the fragrant placebo (odds ratios 2.8) and concluded that aromatherapy did not reduce anxiety.

| Provider
Two interventions were delivered by a member of the research team (n = 2) 14,22 ; however, most studies did not report who delivered the intervention (n = 7). 15-213.1.5| Outcomes Primary and secondary outcomes are presented in Table 3, with summary statistics presented in Table 4.A forest plot of the standardised mean differences is presented in Figure 2. Standardised mean differences (SMDs) were estimated for eight of the nine eligible studies.One study 18 was not included as there was insufficient information reported to estimate a SMD.
Anxiety was primarily assessed using the STAI, 42,55,58 with seven studies including the STAI as a primary outcome measure, six of which used the full measure including both the state and trait subscales, 14,[16][17][18][19][20] and one used the 10 item STAI adult short form Y-1. 21 Two studies 15,22 used the anxiety subscale of the HADS-A, 43 one study 20 used the Visual Facial Anxiety Scale (VFAS) 59 and one study 14 reported physiological measures of anxiety, such as blood pressure, heart rate, respiratory rate and fingertip oxygen saturation. 14n studies that used outcome measures which were not validated anxiety scales, two different measures were used: one study 17 used the Symptom Distress Thermometer (SDT), 60 and one study 18 used the Psychological Distress Inventory (PDI), 61 both secondary outcome measures to another anxiety measure.

| TIDieR checklist
Reporting according to the TIDieR checklist (Template for Intervention Description and Replication) 62 for each of the included studies is reported in Table 5.

| Methodological quality assessment and risk of bias
Individual ratings and the global rating of methodological quality and assessment of risk of bias are reported in Table 6.All of the studies received a strong rating for study design (either a randomised controlled trial or controlled clinical trial design), data collection methods and withdrawals and drop-outs.Selection bias was a concern for all nine studies, primarily due to lack of information about recruitment methods.For five of the studies, it was unclear whether confounders were controlled for in the design or the analyses.Finally, for eight of the nine studies it was unclear whether adequate blinding was utilised in the study design.]22

| Quality of evidence
Using the GRADE approach 63 to assess the quality of the evidence, the overall rating of evidence included in this review was deemed very low.The primary outcomes were initially downgraded from high to moderate to reflect the T A B L E 3 Study outcomes.

Music interventions
Chen 14 Primary outcome: Mean change in anxiety score from pretest (immediately prior to intervention), and posttest (immediately after intervention, delivered on the same day) Outcome measure: State-Trait Anxiety Inventory-State and Trait subscales (STAI-S and STAI-T) 11,12 Statistical analysis: Independent two-sample t tests.

Video-based patient education
Esen 20 Primary outcome: Change in anxiety scores before information sessions, after information sessions and after treatment (immediately after treatment delivered on the same day as the intervention) Outcome measure: State-Trait Anxiety Inventory-State and Trait subscales (STAI-S and STAI-T) high risk of bias.Specific issues included the overall lack of blinding in the studies, with only one study (of aromatherapy) reporting double blinding of participants and outcome assessors. 22The evidence was downgraded another level (from moderate to low) due to heterogeneity of interventions and outcomes.While most of the included studies were music interventions, the other two interventions were video-based patient education, and aromatherapy.Additionally, while many of the studies used the same outcome measure, results were inconsistently reported, precluding meaningful comparisons between studies.For example, several studies reported the percentage of participants under or above critical values, while others reported mean differences.Finally, the evidence was downgraded at a final level from low to very low, due to clinical heterogeneity and indirectness.Some studies focused on specific cancer groups (i.e.certain cancer sites, diagnoses or treatment techniques), some studying only females, and others being more widely inclusive.

| DISCUSSION
This review aimed to assess the efficacy of nonpharmacological interventions delivered to adult patients with cancer, just prior to or during RT, in reducing procedural anxiety.As far as we are aware, this is the first systematic review to do so.The review identified nine studies that met the inclusion criteria, of which six reported the results of trials of music, [14][15][16][17][18][19] two reported on trials of video-based patient education 20,21 and one on a trial of aromatherapy. 22Only three of the included studies 14,15,20 reported a significant reduction in the primary outcome of selfreported procedural anxiety for intervention participants, two music interventions 14,15 and one video-based patient education. 20Of these, the two music intervention studies received a methodological quality rating of strong, 14,15 and the video-based patient education study received a moderate rating. 20Only one study, a music intervention, 14 reported a significant reduction in the secondary outcome of physiological symptoms of procedural anxiety (systolic blood pressure).This review identified some promise of music interventions in reducing procedural anxiety in patients undergoing RT.Of the two studies that reported significant findings, 14,15 both received strong methodological quality and low risk of bias ratings.In addition to statistically significant reductions in anxiety, both interventions also reported clinically meaningfully reductions in mean anxiety scores.The two interventions were both delivered via headphones; however, they varied in terms of where (treatment bunker 15 vs. waiting room 14 ), when (during treatment 15 vs. prior to treatment 14 ) and

Aromatherapy
Graham how often the intervention was delivered (over many sessions 15 vs. prior to a single treatment session 14 ) and whether the music was self-selected 14 or not. 15It is possible that the collection of songs in the study conducted by Chen and colleagues 14 shared similarities to the piece of music chosen for the Karadag and colleagues' 15 study; however, the authors are not qualified to comment on the therapeutic benefit of particular song choices.Finally, inclusion criteria in the two studies differed, with Chen and colleagues 14 including all cancer sites, while Karadag and colleagues 15 included only patients with early-stage breast cancer (thus limiting the population by cancer site and also gender).
The remaining studies of music reported no significant effect on procedural anxiety.Like the two aforementioned studies, the delivery of these interventions varied, with no clear differences from those that reported significant findings.However, three [16][17][18] of the four studies [16][17][18][19] reporting nonsignificant findings received an overall quality and risk of bias rating of 'weak'.This suggests that further, high-quality research may be warranted to rigorously explore the effect of music listening on procedural anxiety, and identify the important factors that contribute to a therapeutic effect of music.
This review identified equivocal findings in the studies trialling video-based patient education.While no effect was reported in the study conducted by Koth and colleagues, 21 a significant reduction in procedural anxiety was reported in the study conducted by Esen and colleagues. 20Overall, both video-based patient education studies received a moderate rating for methodological quality and risk of bias, with identical scores on each of the domains.However, the significant findings reported in the Esen study should be interpreted with a reasonable amount of caution.The intervention and control groups in this study differed considerably in preintervention anxiety scores, with the mean anxiety score for the control group exceeding the cut off for clinically significant anxiety (≥ 40).Therefore, it is entirely possible that the intervention may have been more effective for those experiencing anxiety in the nonclinical range, resulting in a greater reduction in the outcome measure for the intervention group.
This review identified only one trial of aromatherapy, 22 that reported no effect of aromatherapy on anxiety.While the trial comprised a large sample size, it received a global rating of 'weak' on the methodological quality and risk of bias tool, due to selection bias and confounders.Therefore, some further exploration of aromatherapy may be warranted.
It is worth noting some limitations of the interventions included in this review.Overall, none of the interventions screened participants for clinically significant anxiety as part of eligibility for study participation.It is possible that samples with preintervention levels of anxiety in the normal range could prevent a measurable reduction in anxiety.Additionally, the study by Karadag and colleagues excluded women with left-sided breast cancer in this study.This criterion was due to the RT treatment protocol for left-sided cancers, which involves a breath-holding technique, precluding the use of a music listening during the treatment delivery.As this intervention was only trialled on women with breast cancer, the exclusion of women with left-sided cancers is quite a significant limitation.While the delivery of the intervention inside the treatment bunker produced a statistically significant and clinically meaningful reduction in anxiety, the suitability of this intervention is dependent on the RT treatment protocol (e.g.precluding those having treatment for left-sided breast cancer) and therefore limits the clinical utility of this intervention.
There is considerable scope for further innovative research in this area, given the prevalence and impact of procedure related anxiety in the RT setting.Further examination of music interventions and video-based patient education is warranted using large, rigorous trials.It is worth noting the clinical heterogeneity among the trials included in this review, between those that reported an effect of their intervention and those that did not.It is possible that some interventions may be more suited to certain populations.Future research should endeavour to identify the key elements of these interventions for different clinical populations.Additionally, future trials should be designed to only provide interventions to patients who report raised anxiety levels.If patients do not report anxiety at baseline there is no scope for an intervention to reduce anxiety scores (floor effects) and this may obscure any beneficial effects of an intervention.
Another useful avenue for further research would be a review of psychological interventions to complement the current review.A systematic review of the literature of psychological interventions for people with head and neck cancer identified 21 intervention studies between 1980 and 2017. 64However none of these targeted procedural anxiety and all but one required multiple sessions.
We note that patients experiencing clinically severe anxiety, for example a specific phobia which is inhibiting their ability to undergo treatment, are likely to require additional specialised assistance from a trained psychologist.Case studies of cognitive behavioural therapy including graded exposure therapy/systematic desensitisation offer a promising avenue in this regard. 65,66A B L E 5 The TIDieR (Template for Intervention Description and Replication) Checklist.

| Review limitations
There was a high degree of heterogeneity between studies and variability in validated tools used, limiting our ability to confidently pool any data.The inconsistency in findings, and the overall very low rating of the evidence precludes any firm conclusions.A key limitation of this review is that it is possible some of the interventions included in the review may be treating generalised anxiety.The term 'procedural anxiety' is not used consistently, so we chose to focus on studies that target anxiety with interventions just prior to or during RT, using standardised measures typically used to measure procedural anxiety.

2 Control
Primary outcome: Mean change in anxiety score from pre-RT (immediately prior to the first treatment session) to post-RT (immediately after the first treatment session).Outcome measure: State-Trait Anxiety Inventory-State and Trait subscales (STAI-S and STAI-T) 11,12 Statistical analysis: ANCOVA STAI-S Intervention: Mean decrease of 4 (9.6 SD) Control: Mean decrease of 5 (7.0 SD) p value: 0.45 STAI-T Only baseline STAI-T scores reported.O'Steen 17 Primary outcome: percent change in mean anxiety score from pretreatment (within 2 h prior to first RT session) and post-treatment (within an hour after their first RT session).Outcome measure: State-Trait Anxiety Inventory-State and Trait subscales (STAI-S and STAI-T)11,12 Statistical analysis: Independent t-test STAI total (both State and Trait subscales) Intervention: 16% reduction in STAI score Control: 10% reduction in STAI score p value = 0.2197 Intervention: 33% change from high (≥40) to low anxiety.Control: 22% change from high (≥40) to low anxiety.p value = 0.6363 Secondary outcome: percent change in mean anxiety score from pretreatment (within 2 h prior to first RT session) and post-treatment (within an hour after their first RT session).Outcome measure: Symptom Distress Thermometer (SDT) 14 Statistical analysis: Independent t-test SDT Intervention: 13% reduction in SDT score Control: 2% increase in SDT score p value = 0.3298 Intervention: 17% change from high (≥4) to low anxiety.Control: 13% change from high (≥4) to low anxiety.pvalue = 0.8567 Raglio 18 Primary outcome: change in proportion of subjects with an outcome score below the critical value (≤ 40), measured at baseline (T0), at the end of treatments (T1) and at follow-up (T2, 2 weeks after the fifth session of RT) Outcome measure: State-Trait Anxiety Inventory-State and Trait subscales (STAI-S and STAI-T) 11,12 Statistical analysis: chi-square test or Fisher exact test STAI-T-Proportion of patients with a STAI-T score below the critical value (40) End of treatments Intervention: 5% increase in patients with a STAI-T score below the critical value (40) Control: No change Follow-up Intervention: 5% increase in patients with a STAI score below the critical value Control: 5.26% increase in patients with a STAI score below the critical value STAI-S-Proportion of patients with a STAI-S score below the critical value End of treatments Intervention: 5% increase of patients with a STAI score below the critical value Control: 15.79% increase of patients with a STAI score below the critical value Follow up Intervention: 5% increase of patients with a STAI score below the critical value Control: 5.26% increase of patients with a STAI score below the critical value T A B L E 3 (Continued) change in proportion of subjects with an outcome score below the critical value (≤ 35), measured at baseline (T0), at the end of treatments (T1) and at follow-up (T2, 2 weeks after the fifth session of RT) Outcome measure: Psychological Distress Inventory (PDI) 15 Statistical analysis: chi-square test or Fisher exact test PDI-Proportion of patients with a Psychological Distress Inventory (PDI) score below the critical value (35) End of treatments Intervention: 16.67% increase of patients with a STAI score below the critical value Control: 18.42% increase of patients with a STAI score below the critical value Follow up Intervention: No change Control: 9.12% decrease (from treatment 1) of patients with a STAI score below the critical value Smith 19 Primary outcome: Mean change in state anxiety from baseline (time of evaluation) pretreatment (postsimulation), during treatment (at the end of the first week of treatment) and post-treatment (at the end of the third week).Outcome measure: State-Trait Anxiety Inventory-State subscale (STAI-S) 11,12 Statistical analysis: Mixed-design, two-way ANOVA STAI-S Intervention: Mean decrease of 5.5 Control: Mean decrease of 3.1 p value = 0.763 Secondary outcome: Mean change in trait anxiety from baseline (time of evaluation) pretreatment (post simulation), during treatment (at the end of the first week of treatment) and post-treatment (at the end of the third week).Outcome measure: State-Trait Anxiety Inventory-Trait subscale (STAI-T) 11,12 Statistical analysis: Mixed-design, two-way ANOVA STAI-T Intervention: Mean decrease of 2.

F I G U R E 2
Forest plot of standardised mean differences and description of methodology.SMD, standardised mean difference; CI, confidence interval; EO, essential oil; FP, Fragrant placebo; STAI-S, Spielberger State-Trait Anxiety Inventory; HADS-A, Hospital Anxiety and Depression Scale-Anxiety.Participant numbers for O'Callaghan (2012) are lower than reported in Study characteristics.
Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram.T A B L E 1 17 11,12Statistical analysis: Friedman test and Wilcoxon test with Bonferroni correction.

Table 1
as three patients withdrew prior to initial radiation therapy.Graham (2003)included 313 patients but did not report how many were randomised to each group.Summary statistics for the studies with continuous outcomes.Participant numbers for O'Callaghan (2012) are lower than reported in Table1as three patients withdrew prior to initial radiation therapy.Graham (2003)included 313 patients but did not report how many were randomised to each group.Abbreviations: CI, confidence interval; HADS-A, Hospital Anxiety and Depression Scale-Anxiety; SMD, standardised mean difference; STAI-S, Spielberger State-Trait Anxiety Inventory.
T A B E 4

4.2 | Clinical implications Overall
23,[67][68][69], there is no clear evidence base on which to guide clinical practice.There are very few well-designed studies of promising interventions, precluding any specific recommendations regarding interventions to reduce procedural anxiety in the RT setting.CONCLUSION These findings highlight the paucity of good quality evidence regarding nonpharmacological interventions targeting procedural anxiety in the RT setting.While there is limited evidence of promise in music and educationbased interventions, it is unclear which elements of these interventions contributed to a reduction in procedural anxiety.There is significant scope for improvement in this area, especially improved methodological quality and reduced heterogeneity.Future research should not only seek to determine which key elements of interventions are important for a therapeutic effect in defined clinical populations, but also explore novel interventions that have shown promise in reducing anxiety in other medical settings.Methodological quality and risk of bias assessment ratings.