Virtual reality distraction for acute pain in children

Summary of findings 1. Virtual reality distraction compared to no distraction

Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Certainty of the evidence (GRADE)	Comment
Virtual reality distraction compared to no distraction
Patient or population: children (0 to 18 years) Setting: inpatient and outpatient paediatric healthcare setting Intervention: virtual reality distraction Comparison: no distraction
Outcomes	Risk with no distraction	Risk with virtual reality distraction	Relative effect (95% CI)	No of Participants (studies)	Certainty of the evidence (GRADE)	Comment
Acute pain intensity: self‐report (during procedure)				42 (1 RCT)	⊕⊝⊝⊝ VERY LOW ^a,b	One study found no evidence of beneficial effect of non‐immersive VR compared to no distraction.
Acute pain intensity: observer‐report (during procedure)	No study of this comparison reported this outcome.		‐	‐	‐
Acute pain intensity: behavioural measurement (observer‐report) (during procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	62 (2 RCTs)	⊕⊕⊝⊝ LOW ^a,c	One study found no evidence of beneficial effect of non‐immersive VR and another study found evidence of beneficial effect favouring immersive VR compared to no distraction.
Acute pain intensity: self‐report (post‐procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	461 (10 RCTs)	⊕⊝⊝⊝ VERY LOW ^a,d	Four studies found no evidence of a beneficial effect of immersive, semi‐immersive or non‐immersive VR. Five other studies found evidence of beneficial effect favouring immersive VR compared to no distraction. Another study reported less pain in the VR group.
Acute pain intensity: observer‐report (post‐procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	216 (2 RCTs)	⊕⊕⊝⊝ LOW ^a,e	Two studies found evidence of a beneficial effect favouring immersive VR compared to no distraction for primary caregiver/parent and nurse observer‐reports. One of the studies also found evidence of a beneficial effect favouring immersive VR compared to no distraction for researcher observer‐reports.
Acute pain intensity: behavioural measurement (observer‐report) (post‐procedure)				42 (1 RCT)	⊕⊝⊝⊝ VERY LOW ^a,b	One study found no evidence of beneficial effect of non‐immersive VR compared to no distraction.
Adverse effects (related to engagement with VR)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	154 (5 RCTs)	⊕⊝⊝⊝ VERY LOW ^a,d	Three studies reported no adverse effects. Two studies reported mild adverse effects in the VR group.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RCT: Randomised controlled trial; VR: virtual reality
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
^aUnclear risk of selection bias; high risk of performance and detection bias. Downgraded one level for serious study limitations. ^bSmall sample size with a wide 95% CI. Downgraded two levels for very serious imprecision. ^cDifferent populations in terms of age, conditions and settings. Downgraded one level for serious indirectness. ^dDifferent populations in terms of age, conditions and settings; differences in how the intervention was delivered; and differences in way outcomes measured. Downgraded two levels for very serious indirectness. ^eDifferent populations in terms of conditions and settings. Downgraded one level for serious indirectness.

Summary of findings 2. Virtual reality distraction compared to non‐VR distraction

Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comment
Virtual reality distraction compared to non‐VR distraction
Patient or population: children (0 to 18 years) Setting: inpatient and outpatient paediatric healthcare setting Intervention: virtual reality distraction Comparison: other non‐VR distraction
Outcomes	Risk with other non‐VR distraction	Risk with virtual reality distraction	Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comment
Acute pain intensity: self‐report (during procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	106 (2 RCTs)	⊕⊝⊝⊝ VERY LOW ^a,b	One study found evidence of beneficial effect favouring immersive VR and another study found no evidence of a difference in mean pain change scores between immersive VR and non‐VR distraction.
Acute pain intensity: observer‐report (during procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	106 (2 RCTs)	⊕⊕⊝⊝ LOW ^a,c	One study found evidence of beneficial effect favouring immersive VR and another study found no evidence of a difference in mean pain change scores between immersive VR and non‐VR distraction.
Acute pain intensity: behavioural measurement (observer‐report) (during procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	106 (2 RCTs)	⊕⊕⊝⊝ LOW ^a,c	One study found beneficial effect favouring immersive VR and another study reported evidence of a difference in mean pain change scores between immersive VR and non‐VR distraction with less pain behaviours observed for the VR group.
Acute pain intensity: self‐report (post‐procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	575 (8 RCTs)	⊕⊝⊝⊝ VERY LOW ^a,b	Two studies found evidence of beneficial effect favouring immersive VR and another two studies reported a between group difference favouring immersive VR. Two studies found no evidence of beneficial effect for immersive and semi‐immersive VR and another two studies reported no evidence of a difference in mean pain changes scores between immersive VR and non‐VR.
Acute pain intensity: observer‐report (post‐procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	187 (3 RCTs)	⊕⊕⊝⊝ LOW ^a,c	One study found evidence of beneficial effect favouring immersive VR for parent, nurse and researcher reports and another study also found evidence of beneficial effect favouring immersive VR for caregiver observed report. Another study reported no evidence of a difference in mean pain change scores between immersive VR and non‐VR distraction.
Acute pain intensity: behavioural measurement (observer‐report) (post‐procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	106 (2 RCTs)	⊕⊕⊝⊝ LOW ^a,c	One study found evidence of beneficial effect favouring immersive VR and another study found no evidence of a difference in mean pain change scores between immersive VR and non‐VR distraction.
Adverse effects (related to engagement with VR)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	429 (6 RCTs)	⊕⊝⊝⊝ VERY LOW ^a,b	Three studies found no evidence of a difference between immersive VR and non‐VR distraction for adverse effects. Another two studies reported no adverse effects in the VR group. One study reported that the change in estimated cybersickness before and after VR immersion was not significant.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RCT: Randomised controlled trial; VR: virtual reality
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
^aUnclear risk of selection bias; high risk of performance and detection bias. Downgraded one level for serious study limitations. ^bDifferent population in terms of age, conditions and settings; differences in non‐VR distraction comparisons; and differences in way outcomes measured. Downgraded two levels for very serious indirectness. ^cDifferent population in terms of age, conditions and settings; and differences in non‐VR distraction comparisons. Downgraded one level for serious indirectness.

Summary of findings 3. Virtual reality distraction compared to other VR distraction

Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)
Virtual reality distraction compared to other VR distraction
Patient or population: children (0‐18 years) Setting: inpatient and outpatient paediatric healthcare setting Intervention: virtual reality distraction Comparison: other VR distraction
Outcomes	Risk with other VR distraction	Risk with Virtual reality distraction	Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)
Acute pain intensity: self‐report (during procedure)	No study of this comparison reported this outcome.		‐	‐	‐
Acute pain intensity: observer‐report (during procedure)	No study of this comparison reported this outcome.		‐	‐	‐
Acute pain intensity: behavioural measurement (observer‐report) (during procedure)	No study of this comparison reported this outcome.		‐	‐	‐
Acute pain intensity: self‐report (post‐procedure)	No study of this comparison reported this outcome.		‐	‐	‐
Acute pain intensity: observer‐report (post‐procedure)	No study of this comparison reported this outcome.		‐	‐	‐
Acute pain intensity: behavioural measurement (observer‐report) (post‐procedure)	No study of this comparison reported this outcome.		‐	‐	‐
Adverse effects (related to engagement with VR)	No study of this comparison reported this outcome.		‐	‐	‐
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Background

Description of the condition

Healthcare examinations, treatments, procedures and interventions are typical extreme stressors that can lead to pain for children (Clift 2007; Fox 2016; Horstman 2002; Melnyk 2000; Racine 2016; Rassin 2004; Wollin 2004). A recent definition describes pain as "a distressing experience associated with actual or potential tissue damage with sensory, emotional, cognitive, and social components” (Williams 2016). Similar to the International Association for the Study of Pain (IASP) original definition of pain (IASP 2011), this definition highlights that pain has both distressing sensory (i.e. pain intensity) and emotional (i.e. any negative affect secondary to pain such as distress; including anxiety, fear and/or stress) features, associated with actual or potential tissue damage. The sensory and emotional correlates of pain are ‘subjective psychological states’ (Aydede 2017) and can sometimes be difficult to distinguish between (Brown 2012; Curtis 2012; Goodenough 1999; McGrath 2008). This may be especially the case for children under eight years of age, who by virtue of their developmental abilities may be unable to differentiate pain meaningfully from other unpleasant emotions such as fear and anxiety (Blount 2006; Goodenough 1999). These two dimensions of pain (i.e. pain intensity and pain‐related distress) are important to consider in ensuring pain management strategies reduce not only pain intensity but also the distress, anxiety and/or fear associated with medical treatment‐related pain (Goodenough 1999).

Pain impacts on child and parent satisfaction with healthcare delivery and services. In an investigation of the views and experiences of children in Council of Europe member states, Kilkelly 2011 found that 60.1% of child participants rated ‘not being in pain’ as an important element of health care. Yet evidence suggests that acute pain management in children is not always optimal (Cummings 1996; Groenewald 2012; Taylor 2008). Figures estimate that 27% of children experience moderate to severe pain in hospital, with teenagers and infants experiencing higher prevalence rates of 38% and 32% respectively (Groenewald 2012). This can impact on children’s physiological, psychological and emotional well‐being, in both the short and long term.

It is inevitable that children admitted to healthcare settings will likely be exposed to potentially painful procedures on a daily basis. For instance, Stevens 2011 reported that more than three‐quarters (78.2%) of child participants (n = 3822) in their study had undergone at least one painful procedure in a 24‐hour period preceding data collection. While each child was exposed to an average of 6.3 (range 1 to 50) procedures, only a small portion (28.3%) of children had interventions specifically linked to the painful procedure. While acknowledging that certain procedures are essential for routine medical and surgical care, these procedures/treatments can cause pain for the child. Children can feel "threatened by the monster of medical care" where they fear being hurt, forced and violated by the adults delivering that care (Forsner 2009). Pain results in anxiety and stress, which, in turn, negatively impacts not only on a child’s ability to cope with the treatment/intervention but also on their recovery (Li 2009). Inadequate relief of pain during childhood treatments may have long‐term negative effects on future pain tolerance and pain responses (Young 2005).

Non‐pharmacological techniques (e.g. imagery, hypnosis, story‐telling, play, music) have long been promoted as useful adjuncts to pharmacological analgesics (Butler 2005; Klassen 2008; Landier 2010). Yet, aside from distraction and hypnosis, and more recently combined cognitive behavioural therapy and breathing, there is limited evidence to support the efficacy of many of these conventional psychological interventions (e.g. relaxation, guided imagery, music) for reducing procedure‐related pain in children (Birnie 2018; Stinson 2008). In addition, it has been documented that children may benefit more from interactive (e.g. playing a video game) as opposed to passive (e.g. watching a video game) distraction strategies (Wohlheiter 2013). One such recent ‘active’ adjunctive analgesic technique gaining momentum is virtual reality (VR) (Hoffman 2011).

With the use of technology becoming increasingly prevalent in children’s daily lives, alongside the drive towards e‐health and the empowered patient, it seems reasonable to propose that interactive technologies, if proven effective, should be considered as vital intervention vehicles for enhancing health outcomes for children. The use of VR during healthcare procedures/treatments can create a child‐friendly and developmentally sensitive environment, thereby contributing to the European campaign for a child‐friendly approach to health care (Council of Europe 2011).

Description of the intervention

VR, also referred to as virtuality, is defined as a computer technology that creates a simulated environment/world that users perceive as comparable to real world objects/events (Aguinis 2001; Chan 2007; Hoffman 2004a; Weiss 2003). The user’s attention is drawn away from real world visual, auditory and tactile stimuli, and into the virtual world by the multi‐sensory (i.e. sight, sound, touch) nature of the virtual environment (Gold 2006). VR interventions can vary considerably in terms of three core aspects: types of equipment used; content and nature of the virtual world; and levels of engagement users might have. VR draws the user’s attention to a virtual world/environment using real‐time computer graphics and various inputs (e.g. position trackers, mouse and data glove) and output (e.g. shutter glasses, head‐mounted displays, haptic and audio‐visual) devices that make the person an active participant within a computer‐generated three dimensional world. Active interaction, navigation and immersion are key characteristics of VR systems (Aguinis 2001).

The content of some VR interventions has been developed specifically for certain types of procedures (e.g. Snow World and Ice Cream Factory, devised for burn wound dressings) (Chan 2007; Hoffman 2004a), whereas other VR interventions (e.g. Virtual Gorilla) are selected for convenience to engage children at the time of invasive medical procedures (Gershon 2003; Wolitzky 2005). All VR systems are categorised according to how immersive or non‐immersive they are. With non‐immersive systems, the user is connected to the virtual world (by an external monitor) but can still communicate with the real world (e.g. the healthcare environment; Nilsson 2009). With full immersion, the user’s visual and auditory perception and haptics of stimuli in the outside world is blocked as they become fully enveloped in the computer‐generated virtual environment through the use of a head‐mounted display and a tracker position sensor (e.g. a helmet and headphones which exclude visual and auditory inputs from the healthcare environment; Gold 2006; Weiss 2003). It is this sense of presence and immersive attention (i.e. the ability to give users the sense they are somewhere else) that sets VR apart from other technological interventions such as watching television or video movies, or playing simulated or interactive video games (Chan 2007; Gorini 2011; Hoffman 2004a; Nilsson 2009; Steele 2003; Weiss 2003).

How the intervention might work

VR has been used in many contexts (e.g. treating phobias and post‐traumatic stress disorders; training military and medical personnel). For the purposes of this review, the focus is on the use of VR in the reduction of acute pain intensity and pain‐related distress associated with medical treatments/interventions in any healthcare setting. The theory of how VR works in such instances is as a form of distraction; where distraction is referred to as purposefully directing attention away from undesirable sensations (Mobily 1993). Distraction is a common coping mechanism used by school aged children and adolescents for enduring unpleasant situations (Schneider 2000). Distraction interventions function by diverting the child’s attention away from the stimulus producing the pain and refocusing the child’s attention towards a more pleasant and positive stimulus (i.e. the virtual environment; McCaul 1984; Schneider 2000). VR interventions are thought to manifest analgesic effects by altering pain perception through distracting user attention away from the painful procedure, in addition to changing the way a person interprets incoming pain signals, consequently reducing the amount of pain‐related brain activity (as seen on MRI imagery) (Morris 2009). VR exposure can target cognitive and affective pain pathways, thereby decreasing pain intensity, distress, and anxiety by altering how pain signals are processed in the central nervous system. This is achieved by a number of mechanisms including attentional distraction, conditioning of VR imagery and reduced pain.

VR distraction has been used, for example, to minimise children’s anxiety associated with chemotherapy (Ahmadi 2001; Schneider 1999), to reduce children’s pain during burn wound care (Hoffman 2000; Hoffman 2001; Hoffman 2004a), to access intravenous ports in paediatric oncology patients (Wolitzky 2005), to alleviate pain/anxiety for invasive medical procedures such as venipuncture, lumbar puncture, and bone marrow aspirates (Gershon 2003; Gold 2006; Nilsson 2009; Wint 2002), to help adolescents with cerebral palsy as they endure physiotherapy (Steele 2003), and to reduce children’s preoperative anxiety using handheld video games or films (Low 2008; Patel 2006). Together with pharmacological interventions, distraction is thought to be an effective pain management strategy by cognitively redirecting attention away from pain to a more pleasant stimulus, thereby assisting children to cope with the distress of medical treatments. Long‐term benefits include advantages for later adult life, as pain experienced during medical treatments in childhood is predictive of pain during subsequent medical procedures and avoidance of medical care during young adulthood (Blount 2006).

Why it is important to do this review

VR is a recent technological advancement with the potential to modulate children’s pain when they are undergoing healthcare treatments (e.g. intravenous cannulation, lumbar puncture, wound dressings, chemotherapy, bone marrow aspirates). For instance, Gold 2006 reported that children who underwent intravenous cannulae placement without distraction reported a fourfold increase in affective pain when compared to children immersed in a VR intervention. Additionally, children who received a VR intervention were twice as satisfied with their pain management as compared to children not exposed to a VR intervention. Schneider 2000 found 82% of children indicated that their chemotherapy treatment was better with VR as compared to previous chemotherapy treatment without VR. Parents were also satisfied with the use of VR interventions and believed such interventions did reduce children’s pain and enhance children’s cooperation during medical treatments (Gold 2006). In a review, Hoffman 2011 reported a 35% to 50% reduction in procedural pain in burn patients when in a distracting immersive VR.

Despite these positive evaluations and reports of pain reduction, there remains uncertainty over the effectiveness of VR interventions (Dahlquist 2010; Garrett 2014; Kenney 2016; Malloy 2010; Morris 2009). In addition, in comparison to other simpler forms of non‐pharmacological distraction interventions (e.g. imaginary, breathing, positive thinking), there have been some common criticisms levelled at VR such as high costs, bulky equipment, the need for specialist technological skills and the potential for cyber‐sickness, all of which may threaten the widespread implementation of VR for therapeutic healthcare interventions (Bohil 2011). It is important to conduct this systematic review to evaluate the effectiveness of VR as a pain distractor during healthcare treatments. As few psychological interventions incorporate, or evaluate the effectiveness of, modern and novel interactive technologies such as VR, this review complements other Cochrane Reviews (e.g. Birnie 2018) that evaluate the effectiveness of non‐pharmacological distraction‐based interventions for minimising pain in children when undergoing medical treatments.

Objectives

To assess the effectiveness and adverse effects of VR distraction interventions for children (0 to 18 years) with acute pain in any healthcare setting.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised controlled trials (RCTs), including cross‐over and cluster‐RCTs.

Types of participants

We included children aged from birth up to and including 18 years, with acute pain in any healthcare setting.

Types of interventions

Any technology aimed at creating a virtual environment/world, including immersive and/or non‐immersive VR of any intensity or duration, with the purpose of reducing acute pain intensity. These interventions may be used with or without pharmacological support. We included interventions that used any combination of input and output devices (e.g. mouse and shutter glasses; position tracker and head‐mounted display). For the intervention to be VR the participant must be actively interacting with the virtual environment which responds to their actions. We excluded interventions where the user was a passive observer, such as watching a virtually‐simulated movie as opposed to actively engaging in a virtual environment through physical movement.

Interventions of interest were:

VR distraction compared to no distraction;
VR distraction compared to non‐VR distraction;
VR distraction compared to other VR distraction (grouped by level of immersion which takes account of type of technical device, VR environment and level of user interaction).

Types of outcome measures

Primary outcomes

Acute pain intensity:

1. during the procedure, measured using:
- self‐report;
- observer‐report;
- behavioural measurements (observer‐report).
2. post‐procedure (up to one hour), measured using:
- self‐report;
- observer‐report;
- behavioural measurements (observer‐report).

Secondary outcomes

Adverse effects related to engagement with VR. These may include motion sickness, ocular problems (e.g. eye strain, blurred vision), balance disturbances, headaches, fatigue and repetitive strain injuries.
Child satisfaction with VR.
Child pain‐related distress, for example, self‐report, observer‐report or behavioural measurements of child distress, anxiety, fear and/or stress.
Parent anxiety using parent self‐reported anxiety scales or inventories.
Administration of rescue analgesia (i.e. administration of additional analgesic medications to treat acute pain not controlled by child’s scheduled analgesic regimen).
Cost, which may include cost of the VR intervention or duration of child’s treatment (measured in original currency).

Search methods for identification of studies

Electronic searches

With assistance from the Cochrane Pain, Palliative and Supportive Care (PaPaS) Review Group, we searched the following electronic databases up to October 2019.

The Cochrane Central Register of Controlled Trials (CENTRAL) via CRSO searched on 17/10/2019
MEDLINE (OvidSP) (1989 to 17 October 2019)
Embase (OvidSP) (1989 to 17 October 2019)
CINAHL (EbscoHOST) (1989 to 17 October 2019)
PsycINFO (Proquest) (1989 to 17 October 2019)

We used a combination of controlled vocabulary under the existing databases' organisational systems (e.g. MeSH and EMTREE) and free text terms. We did not apply language restrictions. Search dates start at 1989 as the intervention did not exist before this time. Please see the appendices for the search strategies and terms used for each of the databases: CENTRAL (Appendix 1), MEDLINE (Appendix 2), Embase (Appendix 3), CINAHL (Appendix 4), and PsycINFO (Appendix 5).

Searching other resources

We searched for grey literature and ongoing trials using the following methods.

Proceedings from conferences
ProQuest Digital Dissertations and Theses
Index to Theses (Ireland and UK)
TrialsCentralTM (www.trialscentral.org)
Clinical trials register (Clinicaltrials.gov)
WHO Clinical Trial Search Portal (www.who.int/trialsearch)
Current Controlled Trials (www.controlled-trials.com)

We searched reference lists from retrieved eligible studies for other studies potentially eligible for inclusion. We handsearched relevant journals including Virtual Reality (from inception in 1995 to October 2019) and The International Journal of Virtual Reality (from inception in 1998 to October 2019). We contacted experts in the field and authors of included studies about other potentially relevant studies.

Data collection and analysis

Selection of studies

Two review authors (VL, AM) independently assessed each title and abstract retrieved from the electronic searches for relevance using Covidence 2018. We resolved any discrepancies through discussion with a third review author (PB, PH or DD) acting as arbiter as required. We sourced and assessed the full paper if no abstract was available. We obtained and read full texts of the studies that potentially met our inclusion criteria. Two review authors (VL, AM) independently assessed these full texts against the inclusion criteria before a final decision regarding inclusion/exclusion was confirmed. We resolved any discrepancies by consensus or discussion with a third review author (PB, PH or DD) acting as arbiter where necessary. We listed all potentially relevant papers excluded from the review at full‐text stage in the Characteristics of excluded studies table, noting reasons for exclusion. We listed publications in abstract form only in excluded studies. We collated and reported multiple details of the same study/duplicate publications, to ensure that each study (rather than each report) was the unit of interest in the review. We used an adapted PRISMA flow chart to report the screening and selection process.

Data extraction and management

We designed, piloted and amended as necessary a data extraction form based on the Cochrane Pain, Palliative and Supportive Care Review Group template. Two review authors (VL, AM) independently extracted and managed data from each included RCT using the tailored data extraction form. Data extracted from included studies comprised of the following items.

Methods: aim of study, study design, method of participant recruitment, funding source, declaration of interests for primary investigators, statistical methods and consumer involvement.
Risk of bias: as specified under Assessment of risk of bias in included studies.
Participants: description, participant inclusion and exclusion criteria, geographical location, setting, number, age, gender, ethnicity, principal and stage of diagnosis, type of procedure/treatment receiving.
Intervention: details of intervention (including aim, content, format, source, setting) and control/usual care, delivery of intervention (including timing, frequency, duration), providers of the intervention and intervention fidelity/integrity.
Outcomes: primary and secondary outcome measures (as detailed under Types of outcome measures), timing of assessment and methods of assessing outcome measures, follow up for non‐respondents and adverse events.

We resolved any discrepancies in data extraction between the two review authors through discussion or if required, consultation with a third review author (PB, PH or DD). The first review author (VL) entered the data into Review Manager 5 software (RevMan 2014), with the second review author (AM) checking the accuracy of data entry. We attempted to obtain any missing, unclear or incomplete data by contacting the study authors.

Assessment of risk of bias in included studies

Two authors (VL, AM) independently assessed risk of bias for each study, using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We resolved any disagreements by discussion and consensus with a third review author (DD) acting as arbiter as required. We completed a 'Risk of bias' table for each included study using the 'Risk of bias' tool in Review Manager 5.3 (RevMan 2014). We assessed the following for each study.

Random sequence generation (checking for possible selection bias)

We assessed the method used to generate the allocation sequence for each included study as having:

low risk of bias (any truly random process, e.g. random number table; computer random number generator);
high risk of bias (any non‐random process, e.g. odd or even date of birth, hospital or clinic record number); or
unclear risk of bias (insufficient information about the sequence generation process to permit judgement).

For cross‐over studies we also assessed for period effects (i.e. systematic differences between responses in the second period compared to the first period not due to different interventions) as having:

low risk of bias (1:1 allocation ratio where any general trends in outcomes over time will cancel; period effects included in analysis);
high risk of bias (unequal proportions of participants randomised to the different intervention sequences where a general trend in outcomes over time may lead to bias; no period effects included in analysis); or
unclear risk of bias (insufficient or no information available to permit judgement on period effects).

For random sequence generation, studies assessed as high risk of bias were excluded.

Allocation concealment (checking for possible selection bias)

The method used to conceal allocation to interventions prior to assignment determines whether intervention allocation could have been foreseen in advance of assignmet, during recruitment, or changed after assignment.

We assessed the allocation concealment methods for each included study as having:

low risk of bias (e.g. telephone or central randomisation; sequentially numbered sealed opaque envelopes);
high risk of bias (e.g. open random allocation; unsealed or non‐opaque envelopes; alternation; date of birth); or
unclear risk of bias (method of concealment not clearly described or not described in sufficient detail to allow for a definite judgement).

For allocation concealment, we excluded studies assessed as having high risk of bias.

Blinding of participants and personnel (checking for possible performance bias)

We assessed the methods used to blind study participants and personnel from knowledge of which intervention a participant received. We assessed methods as having:

low risk of bias (blinding of participants and study personnel ensured, and unlikely that the blinding could have been broken);
high risk of bias (no blinding, incomplete or attempted blinding of participants and study personnel, and possible that non‐blinding of others likely to introduce bias; or attempted blinding could have been broken; or reported as not blinded due to nature of the intervention); or
unclear risk of bias (insufficient information on blinding of participants and study personnel to permit a judgement).

Blinding of outcome assessment (checking for possible detection bias)

We assessed the methods used to blind study participants and outcome assessors from knowledge of which intervention a participant received. We assessed the methods as having:

low risk of bias (clear statement that outcome assessors were unaware of treatment allocation, and describes how this was achieved);
high risk of bias (outcome assessment not blinded or or reported as not blinded due to nature of the intervention); or
unclear risk of bias (states outcome assessors were blind to treatment allocation but lacks a clear statement on how it was achieved).

For cross‐over studies we also assessed for carry‐over effects (i.e. effects of an intervention given in one period continue into a subsequent period, thereby interfering with the effects of the second intervention) as:

low risk of bias (sufficient time for carry‐over effects to disappear before outcome assessment in second period);
high risk of bias (insufficient time for carry‐over effects to disappear before outcome assessment in second period); or
unclear risk of bias (insufficient or no information available to permit judgement on carry‐over effects).

Incomplete outcome data (checking for possible attrition bias due to the amount, nature and handling of incomplete outcome data)

We assessed the methods used to deal with incomplete data as having:

low risk of bias (no missing outcome data, less than 10% missing data, missing data balanced in numbers across intervention groups with similar reasons for missing data across groups);
high risk of bias (used 'completer' analysis, more than 10% missing data); or
unclear risk of bias (no or insufficient information provided to permit judgement).

Selective reporting (checking for reporting bias)

We assessed whether primary and secondary outcome measures were pre‐specified and whether these were consistent with those reported. We assessed the methods as having:

low risk of bias (e.g. all of the study’s pre‐specified outcomes and all expected outcomes of interest to the review have been reported);
high risk of bias (e.g. not all of the study’s pre‐specified outcomes have been reported and outcomes of interest are reported incompletely, and so cannot be used); or
unclear risk of bias (insufficient information available to permit judgement).

For cross‐over studies we also assessed for first period data on a basis of a test for carry‐over as having:

low risk of bias (results from both periods reported);
high risk of bias (results from first period only reported); or
unclear risk of bias (insufficient or no information available to permit judgement on reporting bias).

Size of study (checking for possible biases confounded by small size based on number of participants in each study arm)

We evaluated risk of bias for each included study according to the Cochrane PaPaS Review Group guidance on sample size based on the number of participants included in each study arm. We assessed studies to be at:

low risk of bias (≥ 200 participants per treatment arm);
high risk of bias (< 50 participants per treatment arm); or
unclear risk of bias (50 to 199 participants per treatment arm).

Other potential sources of bias

We assessed whether each trial was free of other problems that could put it at a risk of bias as having:

low risk of bias (appears to be free of other sources of bias);
high risk of bias (has at least one important risk of bias e.g. potential source of bias related to the specific study design used, extreme baseline imbalance etc.); or
unclear risk of bias (there may be risk of bias but there is insufficient information to assess whether an important risk of bias exists, or insufficient rationale or evidence that an identified problem will introduce bias).

We had planned to also extend the risk of bias to include specific questions for cluster trials (e.g. selective recruitment of cluster participants; baseline reporting of comparability of clusters) (Higgins 2011; Ryan 2011), but this was not necessary as we did not identify any cluster‐RCTs for inclusion in the review. We will address this in future updates if applicable.

Measures of treatment effect

We planned to analyze the data using RevMan 5.3 (RevMan 2014). For dichotomous outcomes, we planned to report risk ratios (RR) and 95% confidence intervals (CI). For continuous outcomes, we planned to report mean differences (MD) (if outcomes were measured in the same way between trials) and 95% CI. For trials that used different methods to measure the same outcome, we planned to use standardized mean differences (SMD) and 95% CI. We planned to undertake a meta‐analysis if studies were sufficiently similar in design, population, interventions and outcomes, but this was not possible. We will undertake a meta‐analysis in future updates, should more data become available.

Unit of analysis issues

We acknowledged that issues could arise from the inclusion of cross‐over designs and cluster‐RCTs. We did not identify any eligible cluster‐RCTs. In future updates of this review, if we identify any cluster‐RCTs, we will use effect size estimates and standard errors, adjusted in the analysis for clustering, and combine the studies using the generic inverse‐variance method (Higgins 2011). We will adjust sample sizes using the methods described in the Cochrane Handbook using an estimate of the intra‐cluster correlation co‐efficient (ICC) derived from the trial (if possible), from a similar trial or from a study of a similar population. If we use ICCs from other sources, we will report this and conduct sensitivity analyses to investigate the effect of variation in the ICC.

We identified one within‐subject (cross‐over) study as eligible for inclusion in the review. We analysed the data according to recommendations in the Cochrane Handbook (Higgins 2011) for cross‐over trials. We used the effect estimate and standard deviation based on a paired t‐test. If we identify additional cross‐over studies in future updates, we will combine the studies using the generic inverse‐variance method (Higgins 2011, section 16.4) and seek statistical advice for this part of the analysis.

Dealing with missing data

We contacted study authors for missing data. We noted levels of attrition in the included studies. We were unable to perform any meta‐analysis in this review due to clinical heterogeneity in population and procedural conditions. In future updates of this review, we will conduct analysis of outcomes on an intention‐to‐treat basis (i.e. by including all randomised participants in the group to which they were randomised regardless of whether or not they received the allocated intervention). Where this is not possible (i.e. data were not provided by study authors), we will conduct an analysis based on the number of participants for whom outcome data are known. As part of our 'Risk of bias' assessment, we reported the number of participants lost to follow‐up and the levels of, and reasons for, attrition in each trial. We had intended to investigate the impact of including studies with high levels of missing data in the overall assessment of treatment effect by using sensitivity analysis; however, there were not enough data included in the review to conduct this analysis.

Assessment of heterogeneity

We considered whether studies were sufficiently similar (based on consideration of populations, interventions, settings or methodological features) to allow pooling of data using meta‐analysis and we assessed the degree of statistical heterogeneity by visual inspection of forest plots and by examining the Tau² (tau‐squared), I², and Chi² statistics. We regarded heterogeneity as substantial if:

the I² value was 50% or higher; and either
- there was inconsistency between trials in direction or magnitude of effects (judged visually), or a low (P < 0.10) P value in the Chi² test for heterogeneity; or
- the estimate of between‐study heterogeneity (Tau²) was above zero.

We had intended to investigate the presence of substantial heterogeneity using subgroup and sensitivity analyses however there were insufficient data included in the review to conduct these analyses. We did consider whether an overall summary was meaningful, and if it had been, we would have used a random‐effects analysis to produce it.

We detected substantial clinical heterogeneity across included studies and therefore do not report pooled results from meta‐analysis but instead provide a narrative description of data.

Assessment of reporting biases

We did not identify sufficient trials (at least 10) to evaluate reporting biases graphically using funnel plots. In future updates of this review, if 10 or more studies are included, we will conduct formal tests for funnel plot asymmetry; for continuous outcomes we will use the test proposed by Egger 1997, and for dichotomous outcomes we will use the test proposed by Harbord 2006. We will perform exploratory analyses if asymmetry is detected in any of these tests or is suggested by a visual assessment to investigate it. Where we suspect reporting bias we will attempt to contact study authors asking them to provide missing outcome data. Where this is not possible, and the missing data are thought to introduce serious bias, we will explore the impact of including such studies in the overall assessment of results by a sensitivity analysis.

Data synthesis

We planned to use a random‐effects model for meta‐analysis to produce a summary of effect based on the anticipated variability in the populations and interventions of included studies. However, we did not conduct meta‐analysis due to a lack of suitable studies. We judged that the heterogeneity in interventions and comparisons, participants, settings, and outcomes in our included studies would not contribute to meaningful conclusions from a statistically‐pooled result. Therefore, we present the data in additional tables (Table 1 and Table 2), and narratively describe the results. Where possible we calculated effect estimates for each study using Review Manager 5.3 software (RevMan 2014). In future updates of this review, if we identify enough studies suitable to be combined and undergo quantitative analysis, we will conduct meta‐analysis.

Table 1. Comparison 1: Virtual reality distraction compared to no distraction

Study	Measurement Tool	Data VR Intervention	Data No Distraction	P value
Outcome: Acute pain intensity: self‐report (during the procedure)
Nilsson 2009	Colour Analogue Scale (scale range 0 to 10; 0 = no pain, 10 = most pain)	Mean (SD) During procedure (retrospective recorded after procedure) 2.20 (2.78); N = 21	Mean (SD) During procedure (retrospective recorded after procedure) 1.64 (2.17); N = 21
Outcome: Acute pain intensity: behavioural measurement (observer‐report) (during the procedure)
Nilsson 2009	Face, Legs, Activity, Cry, Consolability Scale (total score out of maximum 10; higher score indicates pain)	Mean (SD) During procedure (retrospective recording after procedure) 1.10 (2.00); N = 21	Mean (SD) During procedure (retrospective recording after procedure) 1.67 (1.59); N = 21
Outcome: Acute pain intensity: self‐report (post‐procedureup to one hour)
Jeffs 2014	Adolescent Pediatric Pain Tool (scale range 0 to 100; no pain to worst pain)	Between group difference Difference: 9.7 mm, 95% CI: ‐9.5 to 28.9 VR group reported non‐statistical significant less procedural pain compared to no distraction group		P = 0.32
Jeffs 2014		Mean (SD) 58.20 (31.7); N = 8	Mean (SD) 37 (31.80); N = 10
Nilsson 2009	Colour Analogue Scale (scale range 0 to 10; 0 = no pain, 10 = most pain)	Mean (SD) 1.05 (1.74); N = 21	Mean (SD) 0.51 (0.99); N = 21
Schmitt 2011	Graphic Rating Scale (scale range 0 to 100; no pain to worst pain)	Mean (SD) After treatment condition 40.15 (30.87); N = 52	Mean (SD) After treatment condition 54.48 (26.68); N = 52
Outcome: Acute pain intensity: behavioural measurement (observer‐report) (post‐procedureup to one hour)
Nilsson 2009	Face, Legs, Activity, Cry, Consolability Scale (total score out of maximum 10; higher score indicates pain)	Mean (SD) 0.29 (0.56); N = 21	Mean (SD) 0.19 (0.51); N = 21
Outcome: Child pain related distress (incl. anxiety, fear and distress)
Nilsson 2009	Facial Affective Scale (scale range 0 to 1; from least to most distressed)	Median During procedure (retrospective recording after procedure) 0.47 (N = 21)	Median During procedure (retrospective recording after procedure) 0.47 (N = 21)	NS
Nilsson 2009		Mean (SD) During procedure (retrospective recording after procedure) 0.41 (0.29)	Mean (SD) During procedure (retrospective recording after procedure) 0.47 (0.21)
Wolitzky 2005	Visual Analogue Scale (scale range 0 to 100; none to maximum pain or anxiety)	Composite score (mean of VAS anxiety and pain scores) During procedure (retrospective recorded after procedure) 12.00 (16.36)	Composite score (mean of VAS anxiety and pain scores) During procedure (retrospective recorded after procedure) 34.45 (41.80)

SD: standard deviation
N: number

Table 2. Comparison 2: Virtual reality distraction compared to non‐VR distraction

Study

Measurement Tool

Data VR Intervention

Data Non‐VR Distraction

P Value

Outcome: Acute pain intensity: self‐report (during the procedure)

Visual Analogue Scale

(scale range 0 to 10; from no pain to pain as bad as it could possibly be)

Mean change scores

After dressing removal (taken as during procedure)

2.9 (2.3); N = 20

Mean change scores

After Dressing removal (taken as during procedure)

4.2 (3.2); N = 21

P = 0.16

Outcome: Acute pain intensity: observer‐report (during the procedure)

Visual Analogue Scale

(scale range 0 to 10; from no pain to worst pain)

Mean change scores

After dressing removal (during procedure)

3.5 (2.5); N = 20

Mean change scores

After dressing removal (during procedure)

3.8 (3.2); N = 21

P = 0.71

Outcome: Acute pain intensity: behavioural measurements (observer‐report) (during the procedure)

Face, Legs, Activity, Cry, Consolability Scale

(total score out of maximum 10; higher score indicates pain)

Mean change scores

After dressing removal (during procedure)

2.9 (2.4); N = 20

Mean change scores

After dressing removal (during procedure)

4.7 (2.5); N = 21

P = 0.02

Outcome: Acute pain intensity: self‐report (post‐procedureup to one hour)

Chan 2019

Emergency department

Faces Pain Scale‐Revised

(scale range 0 to 10; no pain to very much pain)

Between group difference: ‐1.78; 95% CI, ‐3.24 to ‐0.32

P = 0.018

Change in FPS‐R from baseline

‐1.39; 95% CI ‐2.42 to ‐0.36; N = 64

Change in FPS‐R from baseline

0.39; 95%CI ‐1.45 to 0.67; N = 59

VR intervention

P = 0.009

Non‐VR distraction

P = 0.47

Chan 2019

Outpatient pathology

Faces Pain Scale‐Revised

(scale range 0 to 10; no pain to very much pain)

Between group difference: ‐1.39; 95% CI, ‐2.68 to ‐0.11

P = 0.034

Change in FPS‐R from baseline

1.37; 95% CI, 0.50 to 2.23; N = 63

Change in FPS‐R from baseline

2.76; 95% CI, 1.79 to 3.72; N = 66

VR intervention

P = 0.003

Non‐VR distraction

P < 0.001

Visual Analogue Scale

(scale range 0 to 10; no pain to pain as bad as it could possibly be)

Mean change scores

After dressing application (post‐procedure)

2.33 (3.4); N = 20

Mean change scores

Dressing application (post‐procedure)

3.8 (3.6); N = 21

P = 0.40

Jeffs 2014

Adolescent Pediatric Pain Tool

(scale range 0 to 100; higher score indicates more pain)

Between group difference

Difference: 23.7 mm, 95% CI: 2.4 to 45.0

VR group reported significant less procedural pain than PD group

P = 0.029

Means (SD)

After procedure

58.2 (31.7); N = 8

Means (SD)

After procedure

30.2 (29.2); N = 10

Walther‐Larsen 2019

Visual Analogue Scale

(scale range 0 to 100; higher scores indicates more pain)

Mean Difference 5; 95% CI, ‐3 to 13

P = 0.23

Median (IQR)

27 (8 to 33)

Median (IQR)

15 (5 to 30)

Outcome: Acute pain intensity: observer‐report (post‐procedureup to one hour)

Visual Analogue Scale

(scale range 0 to 10; no pain to worst pain)

Mean change scores

After dressing application (post‐procedure)

2.6 (3.5); N = 20

Mean change scores

After dressing application (post‐procedure)

2.2 (4.0); N = 21

P = 0.75

Outcome: Acute pain intensity: behavioural measurements (observer‐report) (post‐procedureup to one hour)

Face, Legs, Activity, Cry, Consolability Scale

(total score out of maximum 10; higher score indicates pain)

Mean change scores

After dressing application (post‐procedure)

1.9 (2.8); N = 20

Mean change scores

After dressing application (post‐procedure)

3.0 (2.8); N = 21

P = 0.23

Outcome: Adverse effects related to engagement with VR

Visual Analogue Scale

(scale range 0 to 10; 0 = no sick tummy (i.e., nausea), 10 = sick tummy as bad as it could possibly be)

Mean change scores

Nausea dressing removal (during procedure)

‐0.7 (1.1); N = 20

Nausea dressing application (after procedure)

‐0.3 (1.0); N = 20

Mean change scores

Nausea dressing removal (during procedure)

‐0.3 (1.5); N = 21

Nausea dressing application (after procedure)

‐0.5 (1.3); N = 21

P = 0.27

P = 0.65

Outcome: Administration of rescue analgesia

Frequency of rescue does of Entonox prescribed after commencement of procedure

Number (%)

3 (15%); N = 20

Number (%)

9 (43%); N = 21

P = 0.05

N: number
SD: standard deviation

Certainty of the evidence

Two review authors (VL and AM) independently rated the certainty of evidence for each outcome. We used the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) system to rank the certainty of the evidence using the GRADEprofiler Guideline Development Tool software (GRADEpro GDT 2015), and the guidelines provided in Chapter 12.2 of the Cochrane Handbook (Higgins 2011).

The GRADE approach uses five considerations (study limitations, consistency of effect, imprecision, indirectness, and publication bias) to assess the certainty of the body of evidence for each outcome. The GRADE system uses the following criteria for assigning grades of evidence:

high: we are very confident that the true effect lies close to the estimate of effect;
moderate: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different;
low: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect;
very low: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

We considered evidence from RCTs as high certainty but we downgraded the evidence by one (‐1) or two (‐2) where we identified:

serious (‐1) or very serious (‐2) limitation to study quality (risk of bias);
serious (‐1) or very serious (‐2) inconsistency across studies;
some (‐1) or major (‐ 2) uncertainty about directness of evidence;
serious (‐1) or very serious (‐2) imprecise data; and
high probability of the presence of publication bias (‐1).

We reported our judgement on the certainty of the evidence in the 'Summary of findings' tables.

'Summary of findings' table

We included three ‘Summary of findings' tables:

VR distraction compared to no distraction;
VR distraction compared to non‐VR distraction;
VR distraction compared to other VR distraction.

We used the methods described in Chapter 11 of the Cochrane Handbook (Schunemann 2011) to prepare the ‘Summary of findings’ tables. For each table, we presented the results for the primary outcomes of acute pain intensity self‐report, observer‐report and behavioural measurements (observer‐report) measured during procedure and up to one hour post‐procedure and the secondary outcome adverse effects, as outlined in the section on Types of outcome measures. As meta‐analysis was not possible in this review, we presented results in a narrative 'Summary of findings' table format. We used the GRADE system to judge the certainty of the evidence using GRADEprofiler software (GRADEpro GDT 2015).

Subgroup analysis and investigation of heterogeneity

We had planned to conduct the following subgroup analyses for the primary outcome:

(i) younger (0 to 9 years) versus older (10 to 18 years) aged children;
(ii) immersive versus non‐immersive VR interventions.

However, we did not do so because of a limited number of heterogeneous small trials for each intervention type and the fact that data were not reported separately for different ages. In future updates of this review, if sufficient data, we will conduct our planned subgroup analyses.

Sensitivity analysis

We had planned to conduct a sensitivity analysis based on trial quality but did not do so as we did not pool data statistically. In future updates of this review, if sufficient data are available, we will perform sensitivity analysis by separating high from low quality. We define ‘high quality’ as a trial having low risk of bias for sequence generation and adequate allocation concealment and low risk of bias for loss to follow up, classified as less than 10% for primary outcome data.

We will perform a sensitivity analysis for plausible variations in estimated intra‐cluster correlation coefficients (ICCs) if unit‐of‐analysis errors arise in future included cluster‐randomised trials and the ICCs had been estimated for those studies. We will limit sensitivity analyses to primary outcomes. We will conduct a sensitivity analysis to determine the influence of validated versus non‐validated scales on the effects of intervention on outcome.

Consumer participation

The editorial process of the Cochrane Pain, Palliative and Supportive Care Review Group included feedback on the review from one consumer referee in addition to health professionals. The first author of the review continues to engage in research in the field of children’s health care and services with particular emphasis on the voice and visibility of children themselves. These child perspectives were drawn upon in the review, alongside children’s perspectives of VR use in health care. We also obtained feedback from the consumer representative organisation Children in Hospital Ireland who recommended that for future trials it would be worth considering how children and young people are involved in the planning and design stage of the research as well as participants. Additionally, for future research it would be worth examining whether capacity/ability to understand/communicate or other aspects of a specific disability or condition would have an important influence on the effectiveness of the distraction stimulus.

Results

Description of studies

Results of the search

We identified 1827 citations in the database search and a further 19 citations through searching other resources. There were 1254 citations for screening after removal of duplicates. We excluded 1131 citations not meeting the review inclusion criteria on initial screening of titles and abstracts. We assessed 123 citations for eligibility at full‐text screening and excluded 70 citations (representing 65 studies) that did not meet the review selection criteria. Seventeen studies (reported in 18 citations) met the inclusion criteria and four studies (reported in five citations) are awaiting classification. Thirty studies (reported in 32 citations) are ongoing. See Figure 1.

Figure 1

Study flow diagram.

Included studies

See Characteristics of included studies.

We identified 17 RCTs (reported in 18 citations). Thirteen of these used a parallel design (Chan 2019; reporting two RCTs); Chen 2019; Dumoulin 2019; Gerceker 2018; Gold 2006; Hua 2015; Jeffs 2014; Kipping 2012; Koushali 2017; Nilsson 2009; Walther‐Larsen 2019; Wolitzky 2005). Of these 13 studies, 10 had two groups (Chan 2019; reporting two RCTs); Chen 2019; Gold 2006; Hua 2015; Kipping 2012; Koushali 2017; Nilsson 2009; Walther‐Larsen 2019; Wolitzky 2005) and three had three groups (Dumoulin 2019; Gerceker 2018; Jeffs 2014). Four studies used a cross‐over (within‐subjects) design (Atzori 2018; Das 2005; Hoffman 2019; Schmitt 2011).

We contacted the authors of seven studies to obtain non‐reported mean and standard deviation data (Gold 2006; Jeffs 2014; Kipping 2012; Nilsson 2009; Schmitt 2011; Walther‐Larsen 2019; Wolitzky 2005). Three (Jeffs 2014; Nilsson 2009;Schmitt 2011) responded with data (recorded in Table 1 and Table 2), and one responded stating they no longer had access to the data (Wolitzky 2005).

Study population

The 17 included studies (see Table 3 for PICOs of included studies) had a total of 1008 participants undergoing needle‐related procedures for venepuncture, port access, intravenous placement or intravenous injection (Atzori 2018; Chan 2019; Chen 2019; Dumoulin 2019; Gerceker 2018; Gold 2006; Nilsson 2009; Walther‐Larsen 2019; Wolitzky 2005); wound dressing procedures for chronic wounds and burns (Das 2005; Hoffman 2019; Hua 2015; Jeffs 2014; Kipping 2012; Koushali 2017); and active‐assisted range‐of‐motion physical therapy sessions post burn injuries (Schmitt 2011). Individual study sample sizes ranged from nine (Das 2005) to 136 (Chen 2019) participants.

See Characteristics of studies awaiting classification.

Table 3. PICOs of included studies

STUDY	POPULATION	INTERVENTION	COMPARISON	OUTCOME	STUDY DESIGN
Comparison 1: Virtual reality distraction compared to no distraction Primary Outcome: acute pain intensity
During the procedure: self‐report
Nilsson 2009	5 to 18 years Cancer Venous puncture or venous port device access procedure	Non‐immersive VR distraction	No distraction	Colour Analogue Scale (0 to 10; no pain to most pain)	Parallel group design
During the procedure: behavioural measurement (observer‐report)
Nilsson 2009	5 to 18 years Cancer Venous puncture or subcutaneous venous port device access procedure	Non‐immersive VR distraction	No distraction	FLACC (nurse‐reported) (total score maximum of 10; higher score indicates more pain)	Parallel group design
Wolitzky 2005	7 to 14 years Cancer Port access procedure	Immersive VR distraction	No distraction	CHEOPS (researcher‐reported) (minimun score 4 = no pain and maximun score 13 = worst pain)	Parallel group design
*Post‐procedure: self‐report*
Atzori 2018	7 to 17 years Onco‐haematological disease Venipuncture for IV placement during chemotherapy, transfusions, MRI or blood analysis	Immersive VR distraction	No distraction	Visual Analogue Scale (scale 0 to 10; no pain to worst pain)	Within‐subject design
Chen 2019	7 to 12 years Intravenous injections	Immersive VR distraction	No distraction	Wong Baker faces rating scale (0 to 10; 0 no pain, 10 excruciating pain	Parallel group design
Das 2005	5 to 16 years Acute burn injuries Wound dressing change	Immersive VR distraction	No distraction	Modified self‐report Faces pain (scale 0 to 10; no pain to worst pain)	Within‐subject design
Gerceker 2018	7 to 12 years Phlebotomy	Immersive VR distraction	No distraction	Wong Baker Faces pain scale (scale 0 to 10; no pain to worst pain)	Parallel group design
Gold 2006	8 to 12 years IV placement for MRI/CT scan	Immersive VR distraction	No distraction	Faces Pain Scale‐Revised (scale 0 to 10; no pain to very much pain)	Parallel group design
Hoffman 2019	6 to 17 years Burn injuries Hydrotank	Immersive VR distraction	No distraction	Graphic Rating Scale (scale 0 to 10; no pain to worst pain)	Within‐subject design
Jeffs 2014	10 to 17 years Burn injuries Wound dressing changes	Semi‐immersive VR distraction	No distraction	Adolescent Pediatric Pain Tool (scale range 0 to 100; no pain to worst pain)	Parallel group design
Koushali 2017	7 to 12 years Burn injuries Wound dressing changes	Immersive VR distraction	No distraction	Wong Baker Faces pain scale (scale 0 to 10; no pain to worst pain)	Parallel group design
Nilsson 2009	5 to 18 years Cancer Venous punctures or subcutaneous venous port device access	Non‐immersive VR distraction	No distraction	Colour Analogue Scale (scale 0 to 10; no pain to most pain)	Parallel group design
Schmitt 2011	6 to 18 years Post burn injuries Active‐assistive range‐of‐motion physical therapy	Immersive VR distraction	No distraction	Graphic Rating Scale (scale 0 to 100; no pain to worst pain)	Within‐subject design
Post‐procedure: observer‐report
Chen 2019	7 to 12 years Intravenous injections	Immersive VR distraction	No distraction	Wong Baker faces rating scale (scale 0 to 10; 0 no pain, 10 excruciating pain)	Parallel group design
Gerceker 2018	7 to 12 years Phlebotomy	Immersive VR distraction	No distraction	Visual Analogue Scale (scale 0 to 10; no pain to worst pain)	Parallel group design
Post‐procedure: behavioural measurement (observer‐report)
Nilsson 2009	5 to 18 years Cancer Venous punctures or subcutaneous venous port device access	Non‐immersive VR distraction	No distraction	FLACC (nurse‐reported) (total score maximum of 10; higher score indicates more pain)	Parallel group design
Comparison 2: Virtual reality distraction compared to non‐VR distraction Primary Outcome: acute pain intensity
During the procedure: self‐report
Hua 2015	4 to 16 years Chronic lower limb wounds Wound dressing changes	Immersive VR distraction	Non‐VR distraction Toys, television, books, parental comforting	Wong‐Baker faces pain scale (scale 0 to 5; no pain to worst pain)	Parallel group design
Kipping 2012	11 to 17 years Burn injuries Wound dressing changes	Immersive VR distraction	Non‐VR distraction Television, stories, music or caregivers and child preference for no distraction	Visual Analogue Scale (scale 0 to 10; no pain to as painful as it could possibly be)	Parallel group design
During the procedure: observer‐report
Hua 2015	4 to 16 years Chronic lower limb wounds Wound dressing changes	Immersive VR distraction	Non‐VR distraction Toys, television, books, parental comforting	Visual Analogue Scale – caregiver ratings (scale 0 to 10; no pain to worst pain)	Parallel group design
Kipping 2012	11 to 17 years Burn injuries Wound dressing changes	Immersive VR distraction	Non‐VR distraction Television, stories, music or caregivers and child preference for no distraction	Visual Analogue Scale ‐ caregiver ratings (scale 0 to 10; no pain to as painful as it could possibly be)	Parallel group design
During procedure: behavioural measurements (observer‐report)
Hua 2015	4 to 16 years Chronic lower limb wounds Wound dressing changes	Immersive VR distraction	Non‐VR distraction Toys, television, books, parental comforting	FLACC pain rating scale (total score maximum of 10; higher score indicating more pain)	Parallel group design
Kipping 2012	11 to 17 years Burn injuries Wound dressing changes	Immersive VR distraction	Non‐VR distraction Television, stories, music or caregivers and child preference for no distraction	FLACC pain rating scale (total score maximum of 10; higher score indicating more pain)	Parallel group design
Post‐procedure: self‐report
Chan 2019 Emergency department	4 to 11 years Venipuncture or intravenous cannulation for any indication (ED dept.)	Immersive VR distraction	Non‐VR distraction Age‐appropriate distraction such as child‐life therapy, toys, books, and electronic devices. Distractions allowed at discretion of clinician and standardised distractions not mandated to reflect clinical practice. Note: In the non‐VR group, distraction used in 43 patients; with electronic media (TV, video, phones, tablets) in 32 patients	Faces Pain Scale Revised (scale 0 to 10; no pain to very much pain)	Parallel group design
Chan 2019 Outpatient pathology	4 to 11 years Venipuncture or intravenous cannulation for any indication (Outpatients)	Immersive VR distraction	Non‐VR distraction Age appropriate distraction such as child‐life therapy, toys, books, and electronic devices. Distractions allowed at discretion of clinician and standardised distractions not mandated to reflect clinical practice. Note: In the non‐VR group distraction used in 57 patients with electronic media in 55 patients	Faces Pain Scale Revised (scale 0 to 10; no pain to very much pain)	Parallel group design
Dumoulin 2019	8 to 17 years Needle‐related procedures (blood work, IV, or both) for pending or known diagnostic and visiting ED	Immersive VR distraction	Non‐VR distraction Two groups 1. Watching TV 2. Child Life program	Visual Analogue Scale (scale 0 to 100; higher scores indicate more pain)	Parallel group design
Gerceker 2018	7 to 12 years Phlebotomy	Immersive VR distraction	Non‐VR distraction External cold and vibration i.e. Buzzy	Wong‐Baker faces rating scale (scale 0 to 10; very happy/no pain to hurts worst)	Parallel group design
Hua 2015	4 to 16 years Chronic lower limb wounds Wound dressing changes	Immersive VR distraction	Non‐VR distraction Toys, television, books, parental comforting	Wong‐Baker faces pain rating scale (scale 0 to 5; higher scores indicate more pain)	Parallel group design
Jeffs 2014	10 to 17 years Burn injuries Wound dressing changes	Semi‐immersive VR distraction	Non‐VR distraction Watching an age‐appropriate movie	Adolescent Paediatric Pain Tool (scale range 0 to 100; no pain to worst pain)	Parallel group design
Kipping 2012	11 to 17 years Burn injuries Wound dressing changes	Immersive VR distraction	Non‐VR distraction Television, stories, music or caregivers and child preference for no distraction	Visual Analogue Scale (scale range 0 to 10; no pain to worst pain)	Parallel group design
Walther‐Larsen 2019	7 to 16 years Venous cannulation before planned anaesthetic induction	Immersive VR distraction	Non‐VR distraction Tablet or smartphone to play two‐dimensional game of own choice	Visual Analogue Scale (scale range 0 to 100; higher score indicates more pain)	Parallel group design
Post‐procedure: observer‐report
Gerceker 2018	7 to 12 years Phlebotomy	Immersive VR distraction	Non‐VR distraction External cold and vibration i.e. Buzzy	Visual Analogue Scale (scale range 0 to 10; no pain to worst pain)	Parallel group design
Hua 2015	4 to 16 years Chronic lower limb wounds Wound dressing changes	Immersive VR distraction	Non‐VR distraction Toys, television, books, parental comforting	Visual Analogue Scale (scale range 0 to 10; no pain to worst pain)	Parallel group design
Kipping 2012	11 to 17 years Burn injuries Wound dressing changes	Immersive VR distraction	Non‐VR distraction Television, stories, music or caregivers and child preference for no distraction	Visual Analogue Scale (scale range 0 to 10; no pain to worst pain)	Parallel group design
Post‐procedure: behavioural measurements (observer‐report)
Hua 2015	4 to 16 years Chronic lower limb wounds Wound dressing changes	Immersive VR distraction	Non‐VR distraction Toys, television, books, parental comforting	FLACC pain rating scale (total score maximum of 10; higher score indicating more pain)	Parallel group design
Kipping 2012	11 to 17 years Burn injuries Wound dressing changes	Immersive VR distraction	Non‐VR distraction Television, stories, music or caregivers and child preference for no distraction	FLACC pain rating scale (total score maximum of 10; higher score indicating more pain)	Parallel group design

CHEOPS: Children's Hospital of Eastern Ontario Pain Scale
ED: Emergency Department
FLACC: Faces, Legs, Activity, Cry, Consolability
VR: virtual reality

The studies varied in terms of the age of the included participants ranging from four to 18 years. Age ranges in individual studies were: seven to 17 years (Atzori 2018); four to 11 years (Chan 2019); five to 16 years (Das 2005); eight to 17 years (Dumoulin 2019); seven to 12 years (Chen 2019; Gerceker 2018; Koushali 2017); eight to 12 years (Gold 2006); six to 17 years (Hoffman 2019); four to 16 years (Hua 2015); 10 to 17 years (Jeffs 2014); 11 to 17 years (Kipping 2012); five to 18 years (Nilsson 2009); six to 18 years (Schmitt 2011; one participant was older than 18 years of age and the author shared the data set, which enabled us to extract the results for children age six to 18 years separately); seven to 16 years (Walther‐Larsen 2019); and seven to 14 years (Wolitzky 2005). Across the 17 included studies the total proportion of male participants was 61%; ranging from 40% (Hua 2015) to 88% (Walther‐Larsen 2019).

Of the six studies that included wound dressing change procedures, all studies reported the cause of the burn or wound injury (Das 2005; Hoffman 2019; Hua 2015; Jeffs 2014; Kipping 2012; Koushali 2017), five reported the total body surface area of the burn or wound size (Das 2005; Hua 2015; Jeffs 2014; Kipping 2012), four reported the wound depth (Hoffman 2019; Jeffs 2014; Kipping 2012; Koushali 2017), three reported the wound or burn site (Hoffman 2019; Hua 2015; Koushali 2017), and two reported type of dressing used (Hua 2015; Kipping 2012). Two studies reported recruiting participants who presented for their first, or first conscious, dressing change (Jeffs 2014; Kipping 2012). One study stated that to be included participants had to have had a minimum of one dressing change (Koushali 2017). One study reported that participants enrolled into the study were at either second or third dressing change (Das 2005). Whether participants were recruited for their first dressing change or multiple dressing changes was unclear in one study (Hua 2015). One study stated that pain was measured during at least one day of wound care and for up to 10 study days the patient used VR (Hoffman 2019).

For one study, eligible participants were those who required post‐burn active‐assisted range‐of‐motion physical therapy at least once during their stay in hospital (Schmitt 2011).

Of the 10 studies that included needle‐related procedures, eight reported on participant previous exposure to needle‐related procedures (Atzori 2018; Chan 2019 (reporting two RCTs); Dumoulin 2019; Gerceker 2018; Gold 2006; Nilsson 2009; Wolitzky 2005).

The main exclusion criteria for participants for the included studies were as follows.

Cognitive impairment or other condition which might impact on self‐report outcome measures (Chan 2019; Chen 2019; Das 2005; Dumoulin 2019; Gerceker 2018; Gold 2006; Hoffman 2019; Jeffs 2014; Kipping 2012; Nilsson 2009; Walther‐Larsen 2019).
Factors which might impact on the effective use of the VR intervention such as medical conditions, visual/auditory impairment, wound location/treatment and impaired range of movement (Chan 2019; Chen 2019; Das 2005; Gold 2006; Hoffman 2019; Hua 2015; Jeffs 2014; Kipping 2012; Schmitt 2011; Walther‐Larsen 2019).
Venous access already inserted or cases where topical anaesthesia not properly applied (Atzori 2018; Walther‐Larsen 2019).
No wound cleaning sessions required or burn size less than 10% TBSA (Hoffman 2019).
Receiving pain or sedative (Gold 2006; Hua 2015; Koushali 2017; Walther‐Larsen 2019).
Having a chronic, genetic, critical or secondary illness, or other conditions contributing to pain (Chan 2019; Chen 2019; Gerceker 2018; Hoffman 2019; Hua 2015).
Receiving prophylaxis for alcohol or drug withdrawal (Hoffman 2019); undergoing chemotherapy (Chen 2019); requiring blood transfusions and blood preparation (Chen 2019).
History of psychiatric disorder or demonstrating delirium, psychosis, or organic brain disorder (Hoffman 2019; Walther‐Larsen 2019).
Past history of epilepsy, seizure disorders, motion sickness, headache, dizziness or recent head injury (Atzori 2018; Chen 2019; Das 2005; Dumoulin 2019; Hoffman 2019; Jeffs 2014; Schmitt 2011; Walther‐Larsen 2019).
Unaccompanied by legal guardian (Atzori 2018) or incarcerated minors and minors in foster care (Jeffs 2014).
Communication or language deficiency (Dumoulin 2019; Gerceker 2018; Hoffman 2019; Hua 2015; Kipping 2012; Koushali 2017; Nilsson 2009; Walther‐Larsen 2019).
Child safety and protection issues (Kipping 2012).
Inability to consent/assent (Chan 2019) or refusal to participate (Gerceker 2018; Koushali 2017).
Wanting to use own distraction tool (Atzori 2018).
Medical procedure lasting more than 15 minutes (Dumoulin 2019).

One study did not report any exclusion criteria (Wolitzky 2005).

Setting

The procedures took place in the following settings.

Oncology/haematology units in Italy (Atzori 2018), Sweden (Nilsson 2009), and the US (Wolitzky 2005).
Emergency departments in Australia (Chan 2019), Taiwan (Chen 2019), and Canada (Dumoulin 2019).
Outpatient pathology in Australia (Chan 2019) and phlebotomy unit in Turkey (Gerceker 2018).
Inpatient, outpatient or intensive care burn units in the US (Hoffman 2019; Jeffs 2014; Schmitt 2011), Australia (Das 2005; Kipping 2012), and Iran (Koushali 2017).
Outpatient radiology department in the US (Gold 2006).
A paediatric centre in a tertiary hospital in China (Hua 2015).
Anaesthetic department in Denmark (Walther‐Larsen 2019).

Interventions and comparators

Types of intervention were immersive VR distraction (15 studies; Atzori 2018; Chan 2019 (reporting two RCTs); Chen 2019; Das 2005; Dumoulin 2019; Gerceker 2018; Gold 2006; Hoffman 2019; Hua 2015; Kipping 2012; Koushali 2017; Schmitt 2011; Walther‐Larsen 2019; Wolitzky 2005), semi‐immersive VR distraction (one study; Jeffs 2014), and non‐immersive VR distraction (one study; Nilsson 2009).

The following virtual environments were used.

Snow World originally designed and created at the University of Washington specifically for patients receiving burn wound procedural care (Atzori 2018; Hoffman 2019; Jeffs 2014; Schmitt 2011).
Aquatic environment developed by two of the authors (Chan 2019).
Choice of four applications i.e. roller coasters, space exploration, wildlife park, travel destinations (Chen 2019).
Game software, developed at the University of South Australia based on the game Quake (Das 2005).
Game of shooting flies flying around an apartment developed by Cyberpsychology Lab. University of Quebec Outaouais (Dumoulin 2019).
Choice of three cartoon videos (i.e. Magic English Disney Family, Princess Sofia’s Secret Library, Dinosaur Animation Cartoon) (Gerceker 2018).
Street Luge (skateboard), by Fifth Dimension Technologies (Gold 2006).
Chinese version of Ice Age 2: Meltdown game (Hua 2015)
Game Chicken Little^TM for ages 11 to 13 years and Need for Speed^TM for ages 14 to 17 years (Kipping 2012).
Smash Hit designed by Mediocre Company in musical arcade style (Koushali 2017).
The Hunt of the diamonds developed by Digital Content Creation software and Adobe Photoshop (Nilsson 2009).
Seagull Splash custom made for needle procedures in collaboration with Khora VR Denmark (Walther‐Larsen 2019).
Educational supplement for children visiting gorilla habitat at Zoo Atlanta (Wolitzky 2005).

Nine studies included no distraction as a comparison group (Atzori 2018; Chen 2019; Das 2005; Gold 2006; Hoffman 2019; Koushali 2017; Nilsson 2009; Schmitt 2011; Wolitzky 2005).

Six studies included non‐VR distractions as comparator groups (Chan 2019 (reporting two RCTs); Dumoulin 2019; Hua 2015; Kipping 2012; Walther‐Larsen 2019). Non‐VR distractions included age‐appropriate distractions such as child‐life therapy, toys, books, and electronic devices allowed at discretion of clinician (Chan 2019); toys, television, books, parental comforting (Hua 2015); television, stories, music or caregivers and child preference for no distraction (Kipping 2012); tablet or smart phone playing a two‐dimensional game of own choice (Walther‐Larsen 2019). One study included two non‐VR distraction comparison groups: one group watched TV via a portable DVD player with video of their choice and a second group received a Child Life program where three preselected activities were provided (Dumoulin 2019),

Two studies included both no distraction and non‐VR distraction as two comparison groups with non‐VR distraction defined as Buzzy a reusable bee which provides both cold compression and vibration in one study (Gerceker 2018) and as watching an age appropriate movie in one study (Jeffs 2014).

Funding sources

Of the 17 included studies, four were supported by funding from research foundations/institutes and donations (Gold 2006; Jeffs 2014; Nilsson 2009; Schmitt 2011), three received support from universities (Chen 2019; Gerceker 2018; Koushali 2017); two received funding from combined sources (i.e. research institute grants, hospital funds, donations, university and foundations) (Atzori 2018; Hoffman 2019); two were was supported by government department of industry, innovation and science (Chan 2019); one was supported by grants from Diversionary Therapy Technologies (Kipping 2012) and one was funded by Khore Virtual Reality Denmark and Kristian Bluff (Motivates) and received donations for software development (Walther‐Larsen 2019). Four studies reported no source of funding (Das 2005; Dumoulin 2019; Hua 2015; Wolitzky 2005).

Declarations of interest were not reported in seven studies (Atzori 2018; Gold 2006; Hoffman 2019; Hua 2015; Jeffs 2014; Koushali 2017; Wolitzky 2005). Ten studies reported declarations of interest statements (Chan 2019 (two RCTs); Chen 2019; Das 2005; Dumoulin 2019; Gerceker 2018; Kipping 2012; Nilsson 2009; Schmitt 2011; Walther‐Larsen 2019).

Excluded studies

See Characteristics of excluded studies.

We excluded 65 studies (reported in 70 citations) after reading full‐texts for the following reasons.

Not a randomised controlled trial (seven studies; Chan 2007; Diaz‐Hennessey 2019; Fonseca 2018; Gold 2005; LeMay 2017a; Shah 2018; Sharar 2008).
Not VR (immersive or non‐immersive) distraction according to our protocol criteria (24 studies; Altmann 2017; Babaie 2019; Blunt 1998; Brown 2014; Burns‐Nader 2013; Celikol 2019; Dosani 2019; Eijlers 2017; Feng 2018; Ferullo 2017; Inan 2019; Kumar 2017; Lozano 2018; MacLaren 2005; Miller 2010; Miller 2011; Miller 2016; Minute 2012; Mott 2008; Newell 2018; Nilsson 2013; Tschiedel 2019; vanRooijen 2017; Windlich‐Biermeier 2007; Zelmar 2015).
Not possible to isolate the effects of VR from a multi‐component intervention (one study; Wolyniez 2013).
Included participants older than 18 years; for published papers that included both children and adults we contacted the authors of these papers with requests for separate data for children 18 years or younger and in cases where we received no response from authors we excluded the paper (15 studies; Canares 2018; Gershon 2004; Gold 2018; Hoffman 2001; Hoffman 2004b; Hoffman 2008; Marteau 2018; Leibovici 2009; Mosso 2008; Parker 2016; Patterson 2010; Sharer 2007; Small 2015; vanGoor 2019; Wint 2002).
Included healthy volunteers (one study; Hoffman 2015).
Did not report pain as an outcome (three studies; Beale 2006, Goldman 2018a; Li 2011).
Focus was experimental pain not in a healthcare setting (one study; Dahlquist 2010).
Abstract form only available with insufficient study detail and results to include in the review (nine studies; Gilbert 2014; Gonzalez 2016; Hoffman 2012; Jeffs 2013; Miller 2009; Nilsson 2007; Tropez‐Arceneaux 2011; Ver Lee 2012; Yuan 2018).
Trial was not completed; confirmed through email correspondence with the authors (three studies; Faucher 2007; Jeffs 2010; LeMay 2016a).

Ongoing studies

See Characteristics of ongoing studies.

Thirty studies (reported in 32 citations) are ongoing. Twenty‐seven of the ongoing studies are registered with ClinicalTrials.gov (NCT03584776; NCT03521076; NCT03480724; NCT03985930; NCT03903510; NCT04040036; NCT03681730; NCT03681717; NCT03681743; NCT03693469; NCT03155607; NCT02986464; LeMay 2016b; NCT03680625; NCT03888690; NCT03784352; NCT03750578; NCT03913897; NCT02646787; NCT03239743; NCT03353584; NCT03464955; NCT03304769; NCT03518346; NCT03645213; NCT02929771; NCT03435367), two are registered with the Australian New Zealand Clinical Trials Registry (ACTRN12617000285358; ACTRN12618001363279) and one is registered with the Chinese Clinical Trial Registry (Wong 2019). Eleven of these studies are being conducted in the USA, eleven in Canada, three in Turkey, two in Australia, one in France, one in Hong Kong and one in Colombia. These ongoing studies are listed below.

Thirteen studies of VR for needle related procedures (i.e. intravenous cannulation, venipuncture, intramuscular medication, immunisation) (ACTRN12617000285358; NCT03985930; ACTRN12618001363279; NCT04040036; NCT03693469; NCT03681730; NCT03750578; NCT03913897; NCT03304769; NCT03518346; NCT03645213; NCT03435367; Wong 2019).
Four studies of VR for wound dressing care or physiotherapy and hydrotherapy for burns or other injuries (NCT03155607; NCT02986464; LeMay 2016b; NCT02646787).
Two studies of VR for post‐operative pain management (NCT03584776; NCT03239743).
Two studies of VR for botulinum toxin injections for spasticity management (NCT03521076; NCT03480724).
One study of VR for pain management of vaso‐occlusive crisis with sickle cell disease (NCT03353584).
One study of VR for pain management for non‐invasive surgical sub‐specialty procedures (NCT03464955).
One study of VR for procedural pain in children with cancer (NCT02929771).
One study of VR for invasive procedures in onco‐haematology unit (NCT03888690).
One study for VR for cast removal (NCT03903510)
Two studies for VR for orthopaedic outpatient procedures (NCT03680625; NCT03784352)
One study for VR for laceration repair (NCT03681717)
One study for VR for minor plastic surgery procedures (NCT03681743)

Studies awaiting classification

We identified four studies as awaiting classification. All four studies are registered with ClinicalTrials.gov. Our initial attempts to contact authors, or to identify subsequent publications, were not successful (Festini 2016; Meyer 2014; NCH 2016; Patterson 2009). If we identify further reports from these studies, we will reassess eligibility.

Risk of bias in included studies

We have summarised ‘Risk of bias’ results in Figure 2 and Figure 3.

Figure 2

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Figure 3

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Random sequence generation

We judged thirteen studies to be at low risk of bias as methods used to generate the random sequence were adequately described (Chan 2019 (two RCTs); Chen 2019; Das 2005; Dumoulin 2019; Gerceker 2018; Hoffman 2019; Hua 2015; Jeffs 2014; Kipping 2012; Koushali 2017; Schmitt 2011; Walther‐Larsen 2019), and four to be at unclear risk of bias because insufficient information was provided on the process for random sequence generation (Atzori 2018; Gold 2006; Nilsson 2009; Wolitzky 2005).

Allocation concealment

We judged six studies to be at low risk of bias because the methods of how allocation was concealed were adequately described (Chan 2019 (two RCTs); Chen 2019; Dumoulin 2019; Gerceker 2018; Kipping 2012), and eleven to be at unclear risk of bias because insufficient information was provided about how allocation was concealed (Atzori 2018; Das 2005; Gold 2006; Hoffman 2019; Hua 2015; Jeffs 2014; Koushali 2017; Nilsson 2009; Schmitt 2011; Walther‐Larsen 2019; Wolitzky 2005).

Blinding

Performance bias

We judged sixteen studies to be at high risk of bias because there was no blinding, blinding was not described (but due to the nature of the intervention was highly unlikely), or reported as not blinded, due to the nature of the intervention (Atzori 2018; Chan 2019 (two RCTs); Chen 2019; Das 2005; Dumoulin 2019; Gerceker 2018; Gold 2006; Hoffman 2019; Hua 2015; Jeffs 2014; Kipping 2012; Nilsson 2009; Schmitt 2011; Walther‐Larsen 2019; Wolitzky 2005), and one study to be at unclear risk of bias because insufficient information was available (Koushali 2017).

Detection bias

We judged sixteen studies to be at high risk of bias because there was no blinding, blinding was not described (but due to the nature of the intervention was highly unlikely), or reported as not blinded, due to the nature of the intervention (Atzori 2018; Chan 2019 (two RCTs); Chen 2019; Das 2005; Dumoulin 2019; Gold 2006; Hoffman 2019; Hua 2015; Jeffs 2014; Kipping 2012; Koushali 2017; Nilsson 2009; Schmitt 2011; Walther‐Larsen 2019; Wolitzky 2005), and one study to be at unclear risk of bias because insufficient information was available (Gerceker 2018).

Incomplete outcome data

We judged fifteen studies to be at low risk of bias because all participant numbers were accounted for (Chan 2019 (two RCTs); Chen 2019; Dumoulin 2019; Gerceker 2018; Gold 2006; Hoffman 2019; Hua 2015; Jeffs 2014; Kipping 2012; Koushali 2017; Nilsson 2009; Schmitt 2011; Wolitzky 2005), and two to be at high risk of bias because of greater than 10% missing data (Atzori 2018; Das 2005).

Selective reporting

We judged thirteen studies to be at low risk of bias as pre‐specified outcomes were reported (Atzori 2018; Chan 2019 (two RCTs); Chen 2019; Das 2005; Dumoulin 2019; Gerceker 2018; Gold 2006; Hoffman 2019; Hua 2015; Kipping 2012; Koushali 2017; Nilsson 2009), and four to be at high risk of bias because not all outcomes were pre‐specified and data could not be used (Jeffs 2014; Schmitt 2011; Walther‐Larsen 2019; Wolitzky 2005).

Other potential sources of bias

Size of study

Following PaPaS review group guidance, we assessed the size of the trial as a potential other source of bias. We judged thirteen studies to be at high risk of bias because these studies had less than 50 participants per treatment arm (Atzori 2018; Das 2005; Dumoulin 2019; Gerceker 2018; Gold 2006; Hoffman 2019; Hua 2015; Jeffs 2014; Kipping 2012; Koushali 2017; Nilsson 2009; Walther‐Larsen 2019; Wolitzky 2005), and four to be at unclear risk of bias because these studies had 50 to 199 participants per treatment arm (Chan 2019 (two RCTs); Chen 2019; Schmitt 2011).

Other bias

We judged fourteen studies to be at low risk of other bias because no other bias was identified (Atzori 2018; Chan 2019 (two RCTs); Chen 2019; Dumoulin 2019; Gerceker 2018; Gold 2006; Hua 2015; Jeffs 2014; Kipping 2012; Koushali 2017; Schmitt 2011; Walther‐Larsen 2019; Wolitzky 2005). We judged one study to be at high risk of bias because the trial design (i.e. exposure to VR and no VR was repeatedly alternated every five minutes) made it difficult to be conclusive about the results (Hoffman 2019); and two studies to be at unclear risk of bias because there was insufficient information to assess whether a risk of bias existed due to the trial design (i.e. the order in which children had wound dressings done; some children had multiple dressings performed) (Das 2005); or differences in the procedures (i.e. insertion of subcutaneous venous port devices or venous punctures) received by participants in the control and intervention groups (Nilsson 2009).

Effects of interventions

See: Summary of findings 1 Virtual reality distraction compared to no distraction; Summary of findings 2 Virtual reality distraction compared to non‐VR distraction; Summary of findings 3 Virtual reality distraction compared to other VR distraction

Comparison 1: VR distraction compared to no distraction

Primary outcome: acute pain intensity

1. During the procedure

a. Acute pain intensity: self‐report

This was reported in one study (42 participants; Nilsson 2009). The study found no evidence of a beneficial effect of non‐immersive VR compared to no distraction as measured using the Colour Analogue Scale (scale range 0 to 10; 0 = no pain, 10 = most pain) (Mean Difference (MD) 0.56, 95% Confidence Interval (CI) ‐0.95 to 2.07).

We judged this to be very low‐certainty evidence, downgraded for serious study limitations, and very serious imprecision (summary of findings Table 1).

b. Acute pain intensity: observer‐report

We did not find any studies that reported this.

c. Acute pain intensity: behavioural measurement (observer‐report)

This was reported in two studies (62 participants; Nilsson 2009; Wolitzky 2005). One study used the FLACC observer‐reported pain rating scale (total score out of maximum 10; higher score indicates more pain) (Nilsson 2009). One study used the CHEOPS observer‐reported pain rating scale (minimum possible score of 4 = no pain, maximum possible score 13 = worst pain) (Wolitzky 2005).

Nilsson 2009 (n = 42) found no evidence of a beneficial effect of non‐immersive VR compared to no distraction for nurse observer‐reported pain (MD ‐0.57, 95% CI ‐1.66 to 0.52). Wolitzky 2005 (n = 20) found evidence of a beneficial effect favouring immersive VR compared to no distraction for researcher observer‐reported pain (MD ‐3.40, 95% CI ‐5.01 to ‐1.79).

We judged this to be low‐certainty evidence, downgraded for serious study limitations, and serious indirectness.

2. Post‐procedure (up to one hour)

a. Acute pain intensity: self‐report

This was reported in ten studies (461 participants; Atzori 2018; Chen 2019; Das 2005; Gerceker 2018; Gold 2006; Hoffman 2019; Jeffs 2014; Koushali 2017; Nilsson 2009; Schmitt 2011).

Atzori 2018 used a Visual Analogue Scale (scale range 0 to 10; no pain to worst pain). Chen 2019, Gerceker 2018 and Koushali 2017 used the Wong Baker Faces pain scale (scale range 0 to 10; no pain to worst/excruciating pain). Das 2005 used a modified Faces pain scale in combination with a visual analogue scale of 0 to 10 (higher score indicates more pain).Gold 2006 used both the Wong‐Baker Faces and the Faces Pain Scale‐Revised rating scales. We report the FPS‐R scale (scale range 0 to 10; no pain to very much pain) in this review because this scale is recommended for assessing self‐report pain intensity in clinical trials for children (Stinson 2006; McGrath 2008). Hoffman 2019 used a Graphic Rating Scale (scale range 0 to 10; no pain to worst pain). Jeffs 2014 used the Adolescent Pediatric Pain Tool (scale range 0 to 100; no pain to worst pain). Nilsson 2009 used the Colour Analogue Scale (scale range 0 to 10; no pain to most pain). Schmitt 2011 used a Graphic Rating Scale (scale range 0 to 100; no pain to worst pain).

Four studies (n = 95) found no evidence of a beneficial effect of immersive (MD ‐1.60, 95% CI ‐3.24 to 0.04; Atzori 2018) and (MD ‐0.60, 95% CI ‐2.47 to 1.27; Gold 2006), semi‐immersive (MD 21.20, 95% CI ‐8.31 to 50.71; Jeffs 2014) or non‐immersive VR (MD 0.54, 95% CI ‐0.32 to 1.40; Nilsson 2009) compared to no distraction.

Five studies (n = 357) found evidence of a beneficial effect favouring immersive VR compared to no distraction (MD ‐1.00, 95% CI ‐1.90 to ‐0.10; Chen 2019) and (MD ‐3.60, 95% CI ‐3.74 to ‐3.46; Gerceker 2018) and (MD ‐3.42, 95% CI ‐4.47 to ‐2.37; Hoffman 2019) and (MD ‐2.90, 95% CI ‐3.57 to ‐2.23; Koushali 2017) and (MD ‐14.33, 95% CI ‐25.42 to ‐3.24; Schmitt 2011). Another study (n = 9) reported less pain in the immersive VR group (Mean 1.3, SD 1.8) than the no distraction group (Mean 4.1, SD 2.9) (Das 2005).

We judged this to be very low‐certainty evidence, downgraded for serious study limitations, and very serious indirectness.

b. Acute pain intensity: observer‐report

This was reported in two studies (216 participants; Chen 2019; Gerceker 2018). Chen 2019 used the Wong Baker Faces pain scale (scale range 0 to 10; no pain to worst pain). Gerceker 2018 used a Visual Analogue Scale (scale range 0 to 10; no pain to worst pain).

Two studies (n = 216) found evidence of a beneficial effect of immersive VR compared to no distraction for primary caregiver/parent (MD ‐1.03, 95% CI ‐1.88 to ‐0.18; Chen 2019 and MD ‐3.20, 95% CI ‐3.34 to ‐3.06; Gerceker 2018) and nurse (MD ‐1.00, 95% CI ‐1.77 to ‐0.23; Chen 2019 and MD ‐2.70, 95% CI ‐2.81 to ‐2.59; Gerceker 2018) observer‐reports of child pain. One study (n = 80) found evidence of a beneficial effect of immersive VR compared to no distraction for researcher observer‐reports of child pain (MD ‐4.10, 95% CI ‐4.24 to ‐3.96; Gerceker 2018).

We judged this to be low‐certainty evidence, downgraded for serious study limitations, and serious indirectness.

c. Acute pain intensity: behavioural measurement (observer‐report)

This was reported in one study (42 participants; Nilsson 2009). The study found no evidence of a beneficial effect of non‐immersive VR compared to no distraction for nurse observer‐reported pain (MD 0.10, 95% CI ‐0.22 to 0.42) as measured used the FLACC observer‐reported pain rating scale (total score out of maximum 10; higher score indicates pain).

We judged this to be very low‐certainty evidence, downgraded for serious study limitations, and very serious imprecision.

Secondary outcomes

Adverse effects related to engagement with VR

Adverse effects related to VR were reported in five studies (154 participants; Atzori 2018; Gold 2006; Hoffman 2019; Jeffs 2014; Schmitt 2011).

Atzori 2018 assessed nausea in both VR and no VR conditions using a Visual Analogue Scale (scale range 0 to 10; none to worst nausea). Gold 2006 used the Child Simulator Sickness Questionnaire to determine whether the child felt sick as a result of the intervention. Hoffman 2019 assessed nausea as a result of VR using a Graphic Rating Scale (scale range 0 to 10; no nausea to vomit). Jeffs 2014 did not report what measure was used to assess adverse effects. Schmitt 2011 used a Graphic Rating Scale (scale range 0 to 100; no nausea to severe nausea) to assess nausea in the VR condition.

Two studies (n = 38) reported no adverse effects (Gold 2006; Jeffs 2014). One study (Atzori 2018; n = 15) found evidence of no evidence of a difference for nausea levels between immersive VR and non‐VR distraction conditions. Hoffman 2019 (n = 48) reported that VR nausea was less than one on a 10 point scale. Schmitt 2011 (n = 53) reported some mild nausea symptoms in the VR treatment group.

We judged this to be very low‐certainty evidence, downgraded for serious study limitations, and very serious indirectness.

Child satisfaction with VR

We did not find any studies that reported child satisfaction with VR.

Child pain‐related distress (including anxiety, fear and/or stress)

Child pain‐related distress was reported in three studies (198 participants; Chen 2019; Nilsson 2009; Wolitzky 2005). Chen 2019 used the Children Fear Scale (scale range 0 to 4; no fear to extreme fear). Nilsson 2009 used the Facial Affective Scale (FAS; scale range 0 to 1 from least to most distressed). Wolitzky 2005 used Visual Analogue Scales (scale range 0 to 100; none to maximum pain or anxiety) to report a composite distress score (i.e. mean of self‐report of anxiety and pain).

Chen 2019 found evidence of a beneficial effect of immersive VR compared to no distraction for reducing child fear (MD ‐0.46, 95% CI ‐0.90 to ‐0.02). Nilsson 2009 found no evidence of a beneficial effect for non‐immersive VR compared to no distraction for reducing pain‐related distress (MD ‐0.06, 95% CI ‐0.21 to 0.09). Wolitzky 2005 found no evidence of a difference in distress between those receiving immersive VR and no distraction (Table 1).

We judged this to be low‐certainty evidence, downgraded for serious study limitations, and serious indirectness.

Parent anxiety

We did not find any studies that reported parent anxiety.

Administration of rescue analgesia

We did not find any studies that reported administration of rescue analgesia.

Cost

We did not find any studies that reported cost.

Subgroup analyses

We did not perform meta‐analysis on any data and were not able to conduct subgroup analyses on data for this comparison.

Sensitivity analyses

We did not perform meta‐analysis on any data and were not able to conduct sensitivity analyses on data for this comparison.

Comparison 2: VR distraction compared to non‐VR distraction

Primary outcome: acute pain intensity

1. During the procedure

a. Acute pain intensity: self‐report

This was reported in two studies (106 participants; Hua 2015; Kipping 2012). Hua 2015 used the Wong‐Baker faces pain rating scale (scale range 0 to 5; no pain to worst pain). Kipping 2012 used a Visual Analogue Scale (scale range 0 to 10; from no pain to pain as bad as it could possibly be).

Hua 2015 (n = 65) found evidence of a beneficial effect favouring immersive VR compared to non‐VR distraction (MD ‐1.77, 95% CI ‐2.74 to ‐0.80). Kipping 2012 (n = 41) found no evidence of a difference in mean pain change scores between immersive VR and non‐VR distraction (Table 2).

We judged this to be very low‐certainty evidence, downgraded for serious study limitations, and very serious indirectness (summary of findings Table 2).

b. Acute pain intensity: observer‐report

This was reported in two studies (106 participants; Hua 2015; Kipping 2012). The two studies used a Visual Analogue Scale to record caregiver ratings (scale ranges 0 to 10; no pain to worst pain).

Hua 2015 (n = 65) found evidence of a beneficial effect favouring immersive VR compared to non‐VR distraction (MD ‐1.90, 95% CI ‐3.23 to ‐0.57). Kipping 2012 (n = 41) found no evidence of a difference in mean pain change score between immersive VR and non‐VR distraction (Table 2).

We judged this to be low‐certainty evidence, downgraded for serious study limitations, and serious indirectness.

c. Acute pain intensity: behavioural measurements (observer‐report)

This was reported in two studies (106 participants; Hua 2015; Kipping 2012). The two studies used the FLACC pain rating scale to record nurse ratings (total score out of maximum 10; higher score indicates more pain).

Hua 2015 (n = 65) found evidence of a beneficial effect favouring immersive VR compared to non‐VR distraction (MD ‐3.18, 95% CI ‐4.75 to ‐1.61). Kipping 2012 (n = 41) found evidence of a difference in mean pain change scores between immersive VR and non‐VR distraction with fewer pain behaviours observed for the VR group (Table 2).

We judged this to be low‐certainty evidence, downgraded for serious study limitations, and serious indirectness.

2. Post‐procedure (up to one hour)

a. Acute pain intensity: self‐report

This was reported in eight studies (575 participants; Chan 2019 (two RCTs); Dumoulin 2019; Gerceker 2018; Hua 2015; Jeffs 2014; Kipping 2012; Walther‐Larsen 2019).

Chan 2019 used the Faces Pain Scale‐Revised (scale range 0 to 10; no pain to very much pain).Gerceker 2018 used the Wong‐Baker faces rating scale (scale range 0 to 10; 0 = very happy/no pain, 10 = hurts worst). Hua 2015 used the Wong‐Baker faces pain rating scale (scale range 0 to 5; higher score indicates more pain). Dumoulin 2019 and Walther‐Larsen 2019 used a Visual Analogue Scale (scale range 0 to 100; higher score indicates more pain). Kipping 2012 used a Visual Analogue Scale (scale range 0 to 10; 0 = no pain, 10 = pain as bad as it could possibly be). Jeffs 2014 used the Adolescent Pediatric Pain Tool (scale range 0 to 100; higher score indicates more pain).

Two studies (n = 146) found evidence of a beneficial effect favouring immersive VR compared to non‐VR distraction (MD ‐0.90, 95% CI ‐1.70 to ‐0.10; Hua 2015 and MD ‐0.50, 95% CI ‐0.59 to ‐0.41; Gerceker 2018). Chan 2019 (two RCTs; n = 252) reported evidence of a between group difference favouring immersive VR compared to non‐VR distraction (Table 2).

One study (n = 59) found no evidence of a beneficial effect of immersive VR compared with TV (MD ‐13.68, 95% CI ‐29.64 to 2.28) and Child Life (MD ‐3.58, 95% CI ‐19.31 to 12.15) non‐VR distraction (Dumoulin 2019). One study (n = 18) found no evidence of a beneficial effect of semi‐immersive VR compared with non‐VR distraction (MD 28.00, 95% CI ‐0.46 to 56.46; Jeffs 2014). Two studies (Kipping 2012; Walther‐Larsen 2019; n = 100) reported no evidence of a between group difference in pain scores between immersive VR and non‐VR distraction (Table 2).

We judged this to be very low‐certainty evidence, downgraded for serious study limitations, and very serious indirectness.

b. Acute pain intensity: observer‐report

This was reported in three studies (187 participants; Gerceker 2018; Hua 2015; Kipping 2012). The three studies used a Visual Analogue Scale (scale range 0 to 10; no pain to worst pain).

Gerceker 2018 (n = 81) found evidence of a beneficial effect favouring immersive VR compared to non‐VR distraction for parent (MD ‐0.50, 95% CI ‐0.59 to ‐0.41), nurse (MD ‐0.20, 95% CI ‐0.27 to ‐0.13) and researcher (MD ‐0.50, 95% CI ‐0.59 to ‐0.41) observer‐reports. Hua 2015 (n = 65) also found evidence of a beneficial effect favouring immersive VR compared with non‐VR distraction for caregiver observer‐reports (MD ‐3.27, 95% CI ‐4.12 to ‐2.42). Kipping 2012 (n = 41) reported no evidence of a difference in mean pain change scores between immersive VR and non‐VR distraction for caregiver ratings (Table 2).

We judged this to be low‐certainty evidence, downgraded for serious study limitations, and serious indirectness.

c. Acute pain intensity: behavioural measurements (observer‐report)

This was reported in two studies (106 participants; Hua 2015; Kipping 2012), as measured by the FLACC pain rating scale (total score of maximum 10; higher score indicates more pain). Hua 2015 (n = 65) found evidence of a beneficial effect favouring immersive VR compared with non‐VR distraction (MD ‐2.11, 95% CI ‐3.73 to ‐0.49). Kipping 2012 (n = 41) reported no evidence of a difference in mean pain change scores between immersive VR and non‐VR distraction (Table 2).

We judged this to be low‐certainty evidence, downgraded for serious study limitations, and serious indirectness.

Secondary outcomes

Adverse effects related to engagement with VR

Adverse effects were reported in six studies (429 participants; Chan 2019 (two RCTs); Dumoulin 2019; Jeffs 2014; Kipping 2012; Walther‐Larsen 2019). Dumoulin 2019 assessed the side effects of VR immersion using a rating scale of how much nausea was experienced in the VR distraction group. Kipping 2012 used a Visual Analogue Scale (scale range 0 to 10; no sick tummy (i.e., nausea) to sick tummy as bad as it could possibly be) to measure nausea associated with VR simulator sickness. Four studies did not report what measure was used to monitor adverse effects related with VR (Chan 2019 (two RCTs); Jeffs 2014; Walther‐Larsen 2019).

Three studies (n = 229) (Chan 2019; one RCT pathology); Kipping 2012; Walther‐Larsen 2019) found no evidence of a difference between immersive VR and non‐VR distraction groups for adverse effects (Table 2). Chan 2019 (one RCT emergency department; n = 123) reported that four participants in the non‐VR distraction group had adverse effects (i.e. dizziness, nausea, headache, vomiting) but there were no adverse effects in the VR distraction group. Dumoulin 2019 (n = 59) reported that the change in estimated cyber sickness before and after VR immersion was not significant (Table 2). Jeffs 2014 (n = 18) reported that no participants in the VR group reported seizures, nausea or adverse motion effects.

We judged this to be very low‐certainty evidence, downgraded for serious study limitations, and very serious indirectness.

Child satisfaction with VR

This was reported in two studies (118 participants; Dumoulin 2019; Walther‐Larsen 2019). The two studies assessed child satisfaction using a Visual Analogue Scale (scale range 0 to 100; higher scores indicate higher satisfaction).

Dumoulin 2019 found no evidence of a beneficial effect of immersive VR compared with Child Life non‐VR distraction (MD 0.00, 95% CI ‐11.19 to 11.19) and evidence of a beneficial effect of immersive VR compared with TV non‐VR distraction (MD 18.43, 95% CI 7.54 to 29.32). Walther‐Larsen 2019 reported no evidence of a difference between immersive VR and non‐VR groups for satisfaction with distraction measures (Table 2).

We judged this to be low‐certainty evidence, downgraded for serious study limitations, and serious indirectness.

Child pain‐related distress (including anxiety, fear and/or stress)

This was reported in three studies (311 participants; Chan 2019 (two RCTs); Dumoulin 2019). Chan 2019 measured child anxiety using a Visual Analogue Thermometer (scale range 0 to 10; no anxiety to most anxiety). Dumoulin 2019 measured child fear of pain using a Visual Analogue Scale (scale range 0‐100; from least to most fear).

Chan 2019 (for two RCTs) reported evidence of a between group difference in child self‐rated anxiety in the VR group compared to the non‐VR distraction group (Table 2). Dumoulin 2019 found no evidence of a beneficial effect for immersive VR compared to TV (MD ‐15.67, 95% CI ‐32.78 to 1.44) and Child Life (Child Life; MD ‐9.58, 95% CI ‐29.34 to 10.18) non‐VR distraction.

We judged this to be very low‐certainty evidence, downgraded for serious study limitations, and very serious indirectness.

Parent anxiety

We did not find any studies that reported parent anxiety.

Administration of rescue analgesia

Administration of rescue analgesia was reported in one study (41 participants; Kipping 2012). The study measured frequency of rescue analgesia (i.e. Entonox doses prescribed after commencement of procedure) requirement. Less frequent administration of rescue analgesia was observed in the VR intervention group (Table 2).

We judged this to be low‐certainty evidence, downgraded for serious study limitations and serious imprecision.

Cost

We did not find any studies that reported cost.

Subgroup analyses

We did not perform meta‐analysis on any data and were not able to conduct subgroup analyses on data for this comparison.

Sensitivity analyses

We did not perform meta‐analysis on any data and were not able to conduct sensitivity analyses on data for this comparison.

Comparison 3: VR distraction compared to other VR distraction

No studies reported this comparison (summary of findings Table 3).

Discussion

Summary of main results

The aim of this review was to assess the effectiveness and adverse effects of VR (immersive and non‐immersive) distraction interventions for reducing acute pain intensity in children (0 to 18 years) in any healthcare setting. We identified 17 studies with 1008 participants aged between four and 18 years undergoing treatments and procedures for intravenous puncture or port access, wound dressing changes, and physical therapy sessions.

The primary outcomes of the review were acute pain intensity (self‐report, observer‐report and behavioural measurement) assessed during and up to one‐hour post‐procedure. The secondary outcomes were adverse effects related to VR, child satisfaction with VR, child pain‐related distress, parent anxiety, administration of rescue analgesia and cost.

We aimed to include studies of VR distraction compared to no distraction, other non‐VR distraction and other VR distraction.

VR distraction compared to no distraction

For acute pain intensity during the procedure we found no evidence of an effect of VR (non‐immersive) for self‐reported pain compared to no distraction. No data were reported for observer‐reported pain. For behavioural measurement (observer‐reported) of pain we found inconsistent evidence of an effect of VR (immersive and non‐immersive).

For acute pain intensity up to one‐hour post‐procedure we found inconsistent evidence of an effect of VR (immersive, semi‐immersive and non‐immersive) for self‐reported pain compared to no distraction. For observer‐reported pain we found evidence of a beneficial effect of immersive VR compared to no distraction. For behavioural measurement (observer‐reported) of pain we found no evidence of an effect of VR (non‐immersive).

For secondary outcomes, there was insufficient and inconsistent evidence to draw a conclusion on adverse effects related to VR. We found inconsistent evidence of an effect of VR for child pain related distress. No studies assessed child satisfaction with VR, parent anxiety, administration of rescue analgesia, and cost.

The GRADE rating for all outcomes for VR distraction compared to no distraction indicated low to very low certainty of evidence meaning we have little confidence in the results. We downgraded the ratings due to serious limitations of study quality, serious or very serious indirectness; and for some outcomes very serious imprecision.

VR distraction compared to other non‐VR distraction

For acute pain intensity during the procedure we found inconsistent evidence of a beneficial effect favouring VR (immersive) for self‐reported and observer‐reported pain compared to other non‐VR distraction. For behavioural measurement (observer‐reported) of pain we found evidence of a beneficial effect for immersive VR compared to other non‐VR distraction.

For acute pain intensity up to one‐hour post‐procedure we found inconsistent evidence of an effect of VR (semi‐immersive and immersive) for self‐reported pain compared to other non‐VR distraction. For observer‐reported pain and behavioural measurement (observer‐reported) of pain we also found inconsistent evidence of an effect of VR (immersive) compared to other non‐VR distraction.

For secondary outcomes, there was insufficient evidence to draw a conclusion on adverse effects related to VR. We found inconsistent evidence of an effect of VR for child satisfaction with VR and for child pain related distress. One study reported less frequent administration of rescue analgesia in the VR intervention group. No studies assessed parent anxiety or cost.

The GRADE rating for all outcomes for VR distraction compared to other non‐VR distraction indicated low to very low certainty of evidence meaning we have little confidence in the results. We downgraded ratings due to serious limitations of study quality, and serious or very serious indirectness.

VR distraction compared to other VR interventions

We found no studies that reported this comparison.

Overall completeness and applicability of evidence

The limited, low‐ to very low‐certainty evidence in this review means that the review question remains incompletely answered. We were unable to draw conclusions about the effectiveness and adverse effects of VR distraction interventions, compared to no distraction, non‐VR distraction or other VR distraction, for children with acute pain in any healthcare setting. Included studies targeted children aged between four and 18 years undergoing needle‐related procedures for venipuncture, port access or intravenous placement; wound dressing procedures for chronic wounds and burns; and active‐assisted range‐of‐motion physical therapy sessions post burn injuries in specific inpatient and outpatient paediatric healthcare settings (i.e. burn units, oncology units, emergency departments, pathology/phlebotomy units, radiology and anaesthetic departments).

The clinical heterogeneity of the child participant populations in terms of age (i.e. seven to 17 years; four to 11 years; five to 16 years; eight to 17 years; seven to 12 years; eight to 12 years; six to 17 years; four to 16 years; 10 to 17 years; 11 to 17 years; five to 18 years; six to 18 years; seven to 16 years; and seven to 14 years) and variations in procedural and treatment conditions (i.e. venipuncture, port access, intravenous placement, wound dressings, physical therapy sessions) across the included studies made statistical pooling of data impossible. Most of the review primary outcomes were assessed by only a small number of studies with small sample sizes. The following secondary outcomes were not measured in any of the included studies: parent anxiety or cost. We found limited data for the remaining secondary outcomes of adverse effects related to VR, child pain‐related distress and administration of rescue analgesia; therefore, we can draw no conclusions related to these outcomes.

Quality of the evidence

We noted that a number of studies reported insufficient information about sequence generation and allocation concealment and we judged them to be at unclear risk of bias. We judged the majority of studies to be at high risk of bias for blinding participants and study personnel, and blinding of outcome assessment, because blinding did not take place, was broken, was not described but due to the nature of the intervention blinding was highly unlikely or was reported as not blinded, due to the nature of the intervention. We also judged a number of studies to be at high risk of bias for selective reporting and the majority of studies to be at high risk of bias for small sample sizes.

We used the GRADE approach to assess the certainty of the evidence. Across all comparisons and outcomes, we downgraded the certainty of evidence to low or very low due to serious study limitations (i.e. unclear risk of selection bias and high risk of performance and detection bias) and serious or very serious indirectness (i.e. differences in populations in terms of age, conditions and settings; differences in how the intervention was delivered; differences in comparisons and differences in the way outcomes were measured). We also downgraded some of the evidence for very serious imprecision (i.e. small sample size with wide 95% CI). Overall, this level of uncertainty makes it difficult to interpret the benefits or lack of benefits of VR distraction for acute pain in children.

Potential biases in the review process

We designed the methods of the review to minimise the introduction of additional bias. Extensive searches of the literature were performed and we attempted to contact authors for additional information as required. Two review authors independently completed data screening, data extraction, risk of bias and certainty of evidence rating. We acknowledge that such assessments involve subjective judgements. For any discrepancies, we reached consensus through consultation with the other authors.

Agreements and disagreements with other studies or reviews

Our review findings largely concur with other non‐Cochrane reviews of VR distraction interventions for acute pain in children in the literature (Eijlers 2019; Garrett 2014; Kenney 2016; Iannicelli 2019; Malloy 2010; Morris 2009). These reviews mainly present limited and inconsistent evidence of the effectiveness of VR distraction for reducing pain intensity in children undergoing clinical treatments and procedures.

Morris 2009 conducted a systematic review of the effectiveness of VR distraction in conjunction with pharmacologic analgesia, compared with pharmacologic analgesia alone or other forms of distraction, on reducing pain in both children and adults with burn injuries undergoing wound dressing changes and physiotherapy management. Of nine included studies, three were specific to children; one was included in our review (Das 2005) and two were excluded (Chan 2007; Hoffman 2000).

In a narrative review of the effectiveness of VR distraction for relieving pain in children and adults according to five pain types (i.e. experimental pain; chronic pruritus; port access and venous punctures; IV placement; and burn injuries), Malloy 2010 included five studies with child participants; three of these are included in our review (Gold 2006; Nilsson 2009; Wolitzky 2005) and two were excluded (Gershon 2004; Mott 2008). In agreement with this current review, the authors conclude that there is limited conclusive evidence to support the effectiveness of immersive and non‐immersive VR distraction for reducing pain in children associated with port access, venous puncture and IV placement.

Garrett 2014 undertook a rapid evidence review of the effectiveness of VR therapy in pain management in both children and adults undergoing a variety of clinical procedures. Of the included child specific studies, Gold 2006, Kipping 2012, Schmitt 2011, and Wolitzky 2005 are included in our review and Gershon 2004, Chan 2007, and Steele 2003 were excluded because they did not meet the review eligibility criteria. The authors conclude that there is conflicting and limited evidence available and, similar to this current review's conclusions, highlight the need for further studies with greater scientific rigour to establish a better evidence base for VR distraction interventions for acute pain management.

In a meta‐analysis of the effectiveness of VR distraction for reducing pain in children and adults in experimental (e.g. cold pressor, finger pressure) and clinical (e.g. IV placement, burn wound dressings) conditions (14 controlled studies) Kenney 2016 included five studies with child populations; four of which were included in our review (Gold 2006; Jeffs 2014; Kipping 2012; Wolitzky 2005) and one was excluded (Gershon 2004). In the review, Kenney 2016 calculated a pooled effect size within each study for each indicator of pain (i.e. intensity, unpleasantness, and threshold). The results suggest that VR distraction is more effective when used with adults than with children; however, the majority of included studies with adults involved experimentally‐induced pain and it is difficult to draw comparisons here with pain experienced by children when undergoing clinical procedures and treatments.

A recent systematic review (Iannicelli 2019) of the efficacy of VR for pain reduction in children and adolescents undergoing treatments and procedures for vaso‐occlusive pain episodes, inferior alveolar nerve block, immunization, pulp therapy, phlebotomy, dressing changes, burn wound care, IV placement and venipuncture included nine studies, five of which met the inclusion criteria for this current review (Chan 2019 (two RCTs); Dumoulin 2019; Gerceker 2018; Hua 2015; Jeffs 2014). In agreement with our review, inconsistent evidence was reported with unclear overall conclusions drawn about the efficacy of VR distraction for acute pain intensity in children.

In another recent systematic review and meta‐analysis on the effectiveness of VR distraction on pain in paediatric patients undergoing medical procedures (14 studies), Eijlers 2019 conclude that VR is more effective than usual care indicating a clinical benefit however the heterogeneity of study effects was high (I² = 93.3%); as a consequence of outlying and low‐quality studies, and the different medical procedures. Eijlers 2019 further conclude that it is hard to differentiate the benefit of VR over other forms of non‐VR distraction (e.g. watching television), and no distraction, because usual care is often not well defined. Eight of these studies were included in our review (Das 2005; Gold 2006; Gerceker 2018; Hua 2015; Jeffs 2014; Kipping 2012; Schmitt 2011; Wolitzky 2005); we did not deem it appropriate to pool data statistically in this current review due to the clinical heterogeneity of the child participant populations and variations in procedural/treatment conditions across the included studies.

Collectively, these reviews, similar to the current review's conclusions, suggest that evidence is limited and of the evidence that does exist it is of low certainty and very low certainty from a small number of studies, and therefore we cannot be confident in the results. As the majority of review/meta‐analysis authors agree, more high certainty evidence with large sample sizes is needed before we can draw conclusions and consider VR distraction as a potential treatment for consistently reducing acute pain intensity in children. This supports the Birnie 2018 Cochrane systematic review of psychological interventions for needle‐related procedural pain and distress in children and adolescents. The review identified evidence supporting the efficacy of distraction, hypnosis, combined cognitive behavioural therapy and breathing interventions for reducing children’s needle related pain and distress. The quality of the trials and overall evidence remains low to very low underscoring the need for improved methodological rigour and trial reporting.

Figure 1

Study flow diagram.

Navigate to figure in ReviewOpen in new tab

Figure 2

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Navigate to figure in ReviewOpen in new tab

Figure 3

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Navigate to figure in ReviewOpen in new tab

Summary of findings 1. Virtual reality distraction compared to no distraction

Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Certainty of the evidence (GRADE)	Comment
Virtual reality distraction compared to no distraction
Patient or population: children (0 to 18 years) Setting: inpatient and outpatient paediatric healthcare setting Intervention: virtual reality distraction Comparison: no distraction
Outcomes	Risk with no distraction	Risk with virtual reality distraction	Relative effect (95% CI)	No of Participants (studies)	Certainty of the evidence (GRADE)	Comment
Acute pain intensity: self‐report (during procedure)				42 (1 RCT)	⊕⊝⊝⊝ VERY LOW ^a,b	One study found no evidence of beneficial effect of non‐immersive VR compared to no distraction.
Acute pain intensity: observer‐report (during procedure)	No study of this comparison reported this outcome.		‐	‐	‐
Acute pain intensity: behavioural measurement (observer‐report) (during procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	62 (2 RCTs)	⊕⊕⊝⊝ LOW ^a,c	One study found no evidence of beneficial effect of non‐immersive VR and another study found evidence of beneficial effect favouring immersive VR compared to no distraction.
Acute pain intensity: self‐report (post‐procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	461 (10 RCTs)	⊕⊝⊝⊝ VERY LOW ^a,d	Four studies found no evidence of a beneficial effect of immersive, semi‐immersive or non‐immersive VR. Five other studies found evidence of beneficial effect favouring immersive VR compared to no distraction. Another study reported less pain in the VR group.
Acute pain intensity: observer‐report (post‐procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	216 (2 RCTs)	⊕⊕⊝⊝ LOW ^a,e	Two studies found evidence of a beneficial effect favouring immersive VR compared to no distraction for primary caregiver/parent and nurse observer‐reports. One of the studies also found evidence of a beneficial effect favouring immersive VR compared to no distraction for researcher observer‐reports.
Acute pain intensity: behavioural measurement (observer‐report) (post‐procedure)				42 (1 RCT)	⊕⊝⊝⊝ VERY LOW ^a,b	One study found no evidence of beneficial effect of non‐immersive VR compared to no distraction.
Adverse effects (related to engagement with VR)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	154 (5 RCTs)	⊕⊝⊝⊝ VERY LOW ^a,d	Three studies reported no adverse effects. Two studies reported mild adverse effects in the VR group.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RCT: Randomised controlled trial; VR: virtual reality
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
^aUnclear risk of selection bias; high risk of performance and detection bias. Downgraded one level for serious study limitations. ^bSmall sample size with a wide 95% CI. Downgraded two levels for very serious imprecision. ^cDifferent populations in terms of age, conditions and settings. Downgraded one level for serious indirectness. ^dDifferent populations in terms of age, conditions and settings; differences in how the intervention was delivered; and differences in way outcomes measured. Downgraded two levels for very serious indirectness. ^eDifferent populations in terms of conditions and settings. Downgraded one level for serious indirectness.

Summary of findings 1. Virtual reality distraction compared to no distraction

Summary of findings 2. Virtual reality distraction compared to non‐VR distraction

Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comment
Virtual reality distraction compared to non‐VR distraction
Patient or population: children (0 to 18 years) Setting: inpatient and outpatient paediatric healthcare setting Intervention: virtual reality distraction Comparison: other non‐VR distraction
Outcomes	Risk with other non‐VR distraction	Risk with virtual reality distraction	Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comment
Acute pain intensity: self‐report (during procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	106 (2 RCTs)	⊕⊝⊝⊝ VERY LOW ^a,b	One study found evidence of beneficial effect favouring immersive VR and another study found no evidence of a difference in mean pain change scores between immersive VR and non‐VR distraction.
Acute pain intensity: observer‐report (during procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	106 (2 RCTs)	⊕⊕⊝⊝ LOW ^a,c	One study found evidence of beneficial effect favouring immersive VR and another study found no evidence of a difference in mean pain change scores between immersive VR and non‐VR distraction.
Acute pain intensity: behavioural measurement (observer‐report) (during procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	106 (2 RCTs)	⊕⊕⊝⊝ LOW ^a,c	One study found beneficial effect favouring immersive VR and another study reported evidence of a difference in mean pain change scores between immersive VR and non‐VR distraction with less pain behaviours observed for the VR group.
Acute pain intensity: self‐report (post‐procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	575 (8 RCTs)	⊕⊝⊝⊝ VERY LOW ^a,b	Two studies found evidence of beneficial effect favouring immersive VR and another two studies reported a between group difference favouring immersive VR. Two studies found no evidence of beneficial effect for immersive and semi‐immersive VR and another two studies reported no evidence of a difference in mean pain changes scores between immersive VR and non‐VR.
Acute pain intensity: observer‐report (post‐procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	187 (3 RCTs)	⊕⊕⊝⊝ LOW ^a,c	One study found evidence of beneficial effect favouring immersive VR for parent, nurse and researcher reports and another study also found evidence of beneficial effect favouring immersive VR for caregiver observed report. Another study reported no evidence of a difference in mean pain change scores between immersive VR and non‐VR distraction.
Acute pain intensity: behavioural measurement (observer‐report) (post‐procedure)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	106 (2 RCTs)	⊕⊕⊝⊝ LOW ^a,c	One study found evidence of beneficial effect favouring immersive VR and another study found no evidence of a difference in mean pain change scores between immersive VR and non‐VR distraction.
Adverse effects (related to engagement with VR)	Data not pooled due to high heterogeneity in interventions, comparisons, participants, settings, and outcomes.		‐	429 (6 RCTs)	⊕⊝⊝⊝ VERY LOW ^a,b	Three studies found no evidence of a difference between immersive VR and non‐VR distraction for adverse effects. Another two studies reported no adverse effects in the VR group. One study reported that the change in estimated cybersickness before and after VR immersion was not significant.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RCT: Randomised controlled trial; VR: virtual reality
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
^aUnclear risk of selection bias; high risk of performance and detection bias. Downgraded one level for serious study limitations. ^bDifferent population in terms of age, conditions and settings; differences in non‐VR distraction comparisons; and differences in way outcomes measured. Downgraded two levels for very serious indirectness. ^cDifferent population in terms of age, conditions and settings; and differences in non‐VR distraction comparisons. Downgraded one level for serious indirectness.

Summary of findings 2. Virtual reality distraction compared to non‐VR distraction

Summary of findings 3. Virtual reality distraction compared to other VR distraction

Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)
Virtual reality distraction compared to other VR distraction
Patient or population: children (0‐18 years) Setting: inpatient and outpatient paediatric healthcare setting Intervention: virtual reality distraction Comparison: other VR distraction
Outcomes	Risk with other VR distraction	Risk with Virtual reality distraction	Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)
Acute pain intensity: self‐report (during procedure)	No study of this comparison reported this outcome.		‐	‐	‐
Acute pain intensity: observer‐report (during procedure)	No study of this comparison reported this outcome.		‐	‐	‐
Acute pain intensity: behavioural measurement (observer‐report) (during procedure)	No study of this comparison reported this outcome.		‐	‐	‐
Acute pain intensity: self‐report (post‐procedure)	No study of this comparison reported this outcome.		‐	‐	‐
Acute pain intensity: observer‐report (post‐procedure)	No study of this comparison reported this outcome.		‐	‐	‐
Acute pain intensity: behavioural measurement (observer‐report) (post‐procedure)	No study of this comparison reported this outcome.		‐	‐	‐
Adverse effects (related to engagement with VR)	No study of this comparison reported this outcome.		‐	‐	‐
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Summary of findings 3. Virtual reality distraction compared to other VR distraction

Table 1. Comparison 1: Virtual reality distraction compared to no distraction

Study	Measurement Tool	Data VR Intervention	Data No Distraction	P value
Outcome: Acute pain intensity: self‐report (during the procedure)
Nilsson 2009	Colour Analogue Scale (scale range 0 to 10; 0 = no pain, 10 = most pain)	Mean (SD) During procedure (retrospective recorded after procedure) 2.20 (2.78); N = 21	Mean (SD) During procedure (retrospective recorded after procedure) 1.64 (2.17); N = 21
Outcome: Acute pain intensity: behavioural measurement (observer‐report) (during the procedure)
Nilsson 2009	Face, Legs, Activity, Cry, Consolability Scale (total score out of maximum 10; higher score indicates pain)	Mean (SD) During procedure (retrospective recording after procedure) 1.10 (2.00); N = 21	Mean (SD) During procedure (retrospective recording after procedure) 1.67 (1.59); N = 21
Outcome: Acute pain intensity: self‐report (post‐procedureup to one hour)
Jeffs 2014	Adolescent Pediatric Pain Tool (scale range 0 to 100; no pain to worst pain)	Between group difference Difference: 9.7 mm, 95% CI: ‐9.5 to 28.9 VR group reported non‐statistical significant less procedural pain compared to no distraction group		P = 0.32
Jeffs 2014		Mean (SD) 58.20 (31.7); N = 8	Mean (SD) 37 (31.80); N = 10
Nilsson 2009	Colour Analogue Scale (scale range 0 to 10; 0 = no pain, 10 = most pain)	Mean (SD) 1.05 (1.74); N = 21	Mean (SD) 0.51 (0.99); N = 21
Schmitt 2011	Graphic Rating Scale (scale range 0 to 100; no pain to worst pain)	Mean (SD) After treatment condition 40.15 (30.87); N = 52	Mean (SD) After treatment condition 54.48 (26.68); N = 52
Outcome: Acute pain intensity: behavioural measurement (observer‐report) (post‐procedureup to one hour)
Nilsson 2009	Face, Legs, Activity, Cry, Consolability Scale (total score out of maximum 10; higher score indicates pain)	Mean (SD) 0.29 (0.56); N = 21	Mean (SD) 0.19 (0.51); N = 21
Outcome: Child pain related distress (incl. anxiety, fear and distress)
Nilsson 2009	Facial Affective Scale (scale range 0 to 1; from least to most distressed)	Median During procedure (retrospective recording after procedure) 0.47 (N = 21)	Median During procedure (retrospective recording after procedure) 0.47 (N = 21)	NS
Nilsson 2009		Mean (SD) During procedure (retrospective recording after procedure) 0.41 (0.29)	Mean (SD) During procedure (retrospective recording after procedure) 0.47 (0.21)
Wolitzky 2005	Visual Analogue Scale (scale range 0 to 100; none to maximum pain or anxiety)	Composite score (mean of VAS anxiety and pain scores) During procedure (retrospective recorded after procedure) 12.00 (16.36)	Composite score (mean of VAS anxiety and pain scores) During procedure (retrospective recorded after procedure) 34.45 (41.80)
SD: standard deviation N: number

Table 1. Comparison 1: Virtual reality distraction compared to no distraction

Table 2. Comparison 2: Virtual reality distraction compared to non‐VR distraction

Study	Measurement Tool	Data VR Intervention	Data Non‐VR Distraction	P Value
Outcome: Acute pain intensity: self‐report (during the procedure)
Kipping 2012	Visual Analogue Scale (scale range 0 to 10; from no pain to pain as bad as it could possibly be)	Mean change scores After dressing removal (taken as during procedure) 2.9 (2.3); N = 20	Mean change scores After Dressing removal (taken as during procedure) 4.2 (3.2); N = 21	P = 0.16
Outcome: Acute pain intensity: observer‐report (during the procedure)
Kipping 2012	Visual Analogue Scale (scale range 0 to 10; from no pain to worst pain)	Mean change scores After dressing removal (during procedure) 3.5 (2.5); N = 20	Mean change scores After dressing removal (during procedure) 3.8 (3.2); N = 21	P = 0.71
Outcome: Acute pain intensity: behavioural measurements (observer‐report) (during the procedure)
Kipping 2012	Face, Legs, Activity, Cry, Consolability Scale (total score out of maximum 10; higher score indicates pain)	Mean change scores After dressing removal (during procedure) 2.9 (2.4); N = 20	Mean change scores After dressing removal (during procedure) 4.7 (2.5); N = 21	P = 0.02
Outcome: Acute pain intensity: self‐report (post‐procedureup to one hour)
Chan 2019 Emergency department	Faces Pain Scale‐Revised (scale range 0 to 10; no pain to very much pain)	Between group difference: ‐1.78; 95% CI, ‐3.24 to ‐0.32		P = 0.018
Chan 2019 Emergency department		Change in FPS‐R from baseline ‐1.39; 95% CI ‐2.42 to ‐0.36; N = 64	Change in FPS‐R from baseline 0.39; 95%CI ‐1.45 to 0.67; N = 59	VR intervention P = 0.009 Non‐VR distraction P = 0.47
Chan 2019 Outpatient pathology	Faces Pain Scale‐Revised (scale range 0 to 10; no pain to very much pain)	Between group difference: ‐1.39; 95% CI, ‐2.68 to ‐0.11		P = 0.034
Chan 2019 Outpatient pathology		Change in FPS‐R from baseline 1.37; 95% CI, 0.50 to 2.23; N = 63	Change in FPS‐R from baseline 2.76; 95% CI, 1.79 to 3.72; N = 66	VR intervention P = 0.003 Non‐VR distraction P < 0.001
Kipping 2012	Visual Analogue Scale (scale range 0 to 10; no pain to pain as bad as it could possibly be)	Mean change scores After dressing application (post‐procedure) 2.33 (3.4); N = 20	Mean change scores Dressing application (post‐procedure) 3.8 (3.6); N = 21	P = 0.40
Jeffs 2014	Adolescent Pediatric Pain Tool (scale range 0 to 100; higher score indicates more pain)	Between group difference Difference: 23.7 mm, 95% CI: 2.4 to 45.0 VR group reported significant less procedural pain than PD group		P = 0.029
Jeffs 2014		Means (SD) After procedure 58.2 (31.7); N = 8	Means (SD) After procedure 30.2 (29.2); N = 10
Walther‐Larsen 2019	Visual Analogue Scale (scale range 0 to 100; higher scores indicates more pain)	Mean Difference 5; 95% CI, ‐3 to 13		P = 0.23
Walther‐Larsen 2019		Median (IQR) 27 (8 to 33)	Median (IQR) 15 (5 to 30)
Outcome: Acute pain intensity: observer‐report (post‐procedureup to one hour)
Kipping 2012	Visual Analogue Scale (scale range 0 to 10; no pain to worst pain)	Mean change scores After dressing application (post‐procedure) 2.6 (3.5); N = 20	Mean change scores After dressing application (post‐procedure) 2.2 (4.0); N = 21	P = 0.75
Outcome: Acute pain intensity: behavioural measurements (observer‐report) (post‐procedureup to one hour)
Kipping 2012	Face, Legs, Activity, Cry, Consolability Scale (total score out of maximum 10; higher score indicates pain)	Mean change scores After dressing application (post‐procedure) 1.9 (2.8); N = 20	Mean change scores After dressing application (post‐procedure) 3.0 (2.8); N = 21	P = 0.23
Outcome: Adverse effects related to engagement with VR
Kipping 2012	Visual Analogue Scale (scale range 0 to 10; 0 = no sick tummy (i.e., nausea), 10 = sick tummy as bad as it could possibly be)	Mean change scores Nausea dressing removal (during procedure) ‐0.7 (1.1); N = 20 Nausea dressing application (after procedure) ‐0.3 (1.0); N = 20	Mean change scores Nausea dressing removal (during procedure) ‐0.3 (1.5); N = 21 Nausea dressing application (after procedure) ‐0.5 (1.3); N = 21	P = 0.27 P = 0.65
Outcome: Administration of rescue analgesia
Kipping 2012	Frequency of rescue does of Entonox prescribed after commencement of procedure	Number (%) 3 (15%); N = 20	Number (%) 9 (43%); N = 21	P = 0.05
N: number SD: standard deviation

Table 2. Comparison 2: Virtual reality distraction compared to non‐VR distraction

Table 3. PICOs of included studies

STUDY	POPULATION	INTERVENTION	COMPARISON	OUTCOME	STUDY DESIGN
Comparison 1: Virtual reality distraction compared to no distraction Primary Outcome: acute pain intensity
During the procedure: self‐report
Nilsson 2009	5 to 18 years Cancer Venous puncture or venous port device access procedure	Non‐immersive VR distraction	No distraction	Colour Analogue Scale (0 to 10; no pain to most pain)	Parallel group design
During the procedure: behavioural measurement (observer‐report)
Nilsson 2009	5 to 18 years Cancer Venous puncture or subcutaneous venous port device access procedure	Non‐immersive VR distraction	No distraction	FLACC (nurse‐reported) (total score maximum of 10; higher score indicates more pain)	Parallel group design
Wolitzky 2005	7 to 14 years Cancer Port access procedure	Immersive VR distraction	No distraction	CHEOPS (researcher‐reported) (minimun score 4 = no pain and maximun score 13 = worst pain)	Parallel group design
*Post‐procedure: self‐report*
Atzori 2018	7 to 17 years Onco‐haematological disease Venipuncture for IV placement during chemotherapy, transfusions, MRI or blood analysis	Immersive VR distraction	No distraction	Visual Analogue Scale (scale 0 to 10; no pain to worst pain)	Within‐subject design
Chen 2019	7 to 12 years Intravenous injections	Immersive VR distraction	No distraction	Wong Baker faces rating scale (0 to 10; 0 no pain, 10 excruciating pain	Parallel group design
Das 2005	5 to 16 years Acute burn injuries Wound dressing change	Immersive VR distraction	No distraction	Modified self‐report Faces pain (scale 0 to 10; no pain to worst pain)	Within‐subject design
Gerceker 2018	7 to 12 years Phlebotomy	Immersive VR distraction	No distraction	Wong Baker Faces pain scale (scale 0 to 10; no pain to worst pain)	Parallel group design
Gold 2006	8 to 12 years IV placement for MRI/CT scan	Immersive VR distraction	No distraction	Faces Pain Scale‐Revised (scale 0 to 10; no pain to very much pain)	Parallel group design
Hoffman 2019	6 to 17 years Burn injuries Hydrotank	Immersive VR distraction	No distraction	Graphic Rating Scale (scale 0 to 10; no pain to worst pain)	Within‐subject design
Jeffs 2014	10 to 17 years Burn injuries Wound dressing changes	Semi‐immersive VR distraction	No distraction	Adolescent Pediatric Pain Tool (scale range 0 to 100; no pain to worst pain)	Parallel group design
Koushali 2017	7 to 12 years Burn injuries Wound dressing changes	Immersive VR distraction	No distraction	Wong Baker Faces pain scale (scale 0 to 10; no pain to worst pain)	Parallel group design
Nilsson 2009	5 to 18 years Cancer Venous punctures or subcutaneous venous port device access	Non‐immersive VR distraction	No distraction	Colour Analogue Scale (scale 0 to 10; no pain to most pain)	Parallel group design
Schmitt 2011	6 to 18 years Post burn injuries Active‐assistive range‐of‐motion physical therapy	Immersive VR distraction	No distraction	Graphic Rating Scale (scale 0 to 100; no pain to worst pain)	Within‐subject design
Post‐procedure: observer‐report
Chen 2019	7 to 12 years Intravenous injections	Immersive VR distraction	No distraction	Wong Baker faces rating scale (scale 0 to 10; 0 no pain, 10 excruciating pain)	Parallel group design
Gerceker 2018	7 to 12 years Phlebotomy	Immersive VR distraction	No distraction	Visual Analogue Scale (scale 0 to 10; no pain to worst pain)	Parallel group design
Post‐procedure: behavioural measurement (observer‐report)
Nilsson 2009	5 to 18 years Cancer Venous punctures or subcutaneous venous port device access	Non‐immersive VR distraction	No distraction	FLACC (nurse‐reported) (total score maximum of 10; higher score indicates more pain)	Parallel group design
Comparison 2: Virtual reality distraction compared to non‐VR distraction Primary Outcome: acute pain intensity
During the procedure: self‐report
Hua 2015	4 to 16 years Chronic lower limb wounds Wound dressing changes	Immersive VR distraction	Non‐VR distraction Toys, television, books, parental comforting	Wong‐Baker faces pain scale (scale 0 to 5; no pain to worst pain)	Parallel group design
Kipping 2012	11 to 17 years Burn injuries Wound dressing changes	Immersive VR distraction	Non‐VR distraction Television, stories, music or caregivers and child preference for no distraction	Visual Analogue Scale (scale 0 to 10; no pain to as painful as it could possibly be)	Parallel group design
During the procedure: observer‐report
Hua 2015	4 to 16 years Chronic lower limb wounds Wound dressing changes	Immersive VR distraction	Non‐VR distraction Toys, television, books, parental comforting	Visual Analogue Scale – caregiver ratings (scale 0 to 10; no pain to worst pain)	Parallel group design
Kipping 2012	11 to 17 years Burn injuries Wound dressing changes	Immersive VR distraction	Non‐VR distraction Television, stories, music or caregivers and child preference for no distraction	Visual Analogue Scale ‐ caregiver ratings (scale 0 to 10; no pain to as painful as it could possibly be)	Parallel group design
During procedure: behavioural measurements (observer‐report)
Hua 2015	4 to 16 years Chronic lower limb wounds Wound dressing changes	Immersive VR distraction	Non‐VR distraction Toys, television, books, parental comforting	FLACC pain rating scale (total score maximum of 10; higher score indicating more pain)	Parallel group design
Kipping 2012	11 to 17 years Burn injuries Wound dressing changes	Immersive VR distraction	Non‐VR distraction Television, stories, music or caregivers and child preference for no distraction	FLACC pain rating scale (total score maximum of 10; higher score indicating more pain)	Parallel group design
Post‐procedure: self‐report
Chan 2019 Emergency department	4 to 11 years Venipuncture or intravenous cannulation for any indication (ED dept.)	Immersive VR distraction	Non‐VR distraction Age‐appropriate distraction such as child‐life therapy, toys, books, and electronic devices. Distractions allowed at discretion of clinician and standardised distractions not mandated to reflect clinical practice. Note: In the non‐VR group, distraction used in 43 patients; with electronic media (TV, video, phones, tablets) in 32 patients	Faces Pain Scale Revised (scale 0 to 10; no pain to very much pain)	Parallel group design
Chan 2019 Outpatient pathology	4 to 11 years Venipuncture or intravenous cannulation for any indication (Outpatients)	Immersive VR distraction	Non‐VR distraction Age appropriate distraction such as child‐life therapy, toys, books, and electronic devices. Distractions allowed at discretion of clinician and standardised distractions not mandated to reflect clinical practice. Note: In the non‐VR group distraction used in 57 patients with electronic media in 55 patients	Faces Pain Scale Revised (scale 0 to 10; no pain to very much pain)	Parallel group design
Dumoulin 2019	8 to 17 years Needle‐related procedures (blood work, IV, or both) for pending or known diagnostic and visiting ED	Immersive VR distraction	Non‐VR distraction Two groups 1. Watching TV 2. Child Life program	Visual Analogue Scale (scale 0 to 100; higher scores indicate more pain)	Parallel group design
Gerceker 2018	7 to 12 years Phlebotomy	Immersive VR distraction	Non‐VR distraction External cold and vibration i.e. Buzzy	Wong‐Baker faces rating scale (scale 0 to 10; very happy/no pain to hurts worst)	Parallel group design
Hua 2015	4 to 16 years Chronic lower limb wounds Wound dressing changes	Immersive VR distraction	Non‐VR distraction Toys, television, books, parental comforting	Wong‐Baker faces pain rating scale (scale 0 to 5; higher scores indicate more pain)	Parallel group design
Jeffs 2014	10 to 17 years Burn injuries Wound dressing changes	Semi‐immersive VR distraction	Non‐VR distraction Watching an age‐appropriate movie	Adolescent Paediatric Pain Tool (scale range 0 to 100; no pain to worst pain)	Parallel group design
Kipping 2012	11 to 17 years Burn injuries Wound dressing changes	Immersive VR distraction	Non‐VR distraction Television, stories, music or caregivers and child preference for no distraction	Visual Analogue Scale (scale range 0 to 10; no pain to worst pain)	Parallel group design
Walther‐Larsen 2019	7 to 16 years Venous cannulation before planned anaesthetic induction	Immersive VR distraction	Non‐VR distraction Tablet or smartphone to play two‐dimensional game of own choice	Visual Analogue Scale (scale range 0 to 100; higher score indicates more pain)	Parallel group design
Post‐procedure: observer‐report
Gerceker 2018	7 to 12 years Phlebotomy	Immersive VR distraction	Non‐VR distraction External cold and vibration i.e. Buzzy	Visual Analogue Scale (scale range 0 to 10; no pain to worst pain)	Parallel group design
Hua 2015	4 to 16 years Chronic lower limb wounds Wound dressing changes	Immersive VR distraction	Non‐VR distraction Toys, television, books, parental comforting	Visual Analogue Scale (scale range 0 to 10; no pain to worst pain)	Parallel group design
Kipping 2012	11 to 17 years Burn injuries Wound dressing changes	Immersive VR distraction	Non‐VR distraction Television, stories, music or caregivers and child preference for no distraction	Visual Analogue Scale (scale range 0 to 10; no pain to worst pain)	Parallel group design
Post‐procedure: behavioural measurements (observer‐report)
Hua 2015	4 to 16 years Chronic lower limb wounds Wound dressing changes	Immersive VR distraction	Non‐VR distraction Toys, television, books, parental comforting	FLACC pain rating scale (total score maximum of 10; higher score indicating more pain)	Parallel group design
Kipping 2012	11 to 17 years Burn injuries Wound dressing changes	Immersive VR distraction	Non‐VR distraction Television, stories, music or caregivers and child preference for no distraction	FLACC pain rating scale (total score maximum of 10; higher score indicating more pain)	Parallel group design
CHEOPS: Children's Hospital of Eastern Ontario Pain Scale ED: Emergency Department FLACC: Faces, Legs, Activity, Cry, Consolability VR: virtual reality

Table 3. PICOs of included studies