Pediatric Distraction Tools for Prehospital Care of Pain and Distress: A Systematic Review

This review adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension statement²⁰  and was registered on the PROSPERO International Prospective Register of Systematic Reviews (CRD42021256097). Post hoc modifications were made to the registered protocol on PROSPERO based on preliminary search findings that identified no relevant prehospital literature. Changes were made to the research question, setting, population, and type of interventions included.

Studies were included if they: (1) reported on distraction techniques, (2) are in pediatrics (⁠$≤$18 years of age) in the ED, (3) with pain and/or distress, and (4) used an interventional or observational study design. Studies using Certified Child Life Specialists (CCLSs), clowns, dogs, or musicians were excluded because they are not feasible in the prehospital setting, requiring an additional person (or dog) on the ambulance for pediatric cases that arise relatively infrequently.¹  An additional study using ice as a distractor was also excluded given the inability to maintain frozen products on ambulances.²¹  Decisions regarding intervention suitability were made through consultation between an intensive care paramedic (K.E.), a CCLS, pediatric emergency physician (A.S.), and emergency physician (P.C.). Case reports, case series, letters, and reviews were excluded.

The Embase, Ovid Medline, CINAHL and Cochrane Library were searched from their beginning to ninth May 2022. Subject headings and keywords relevant to distraction techniques, pediatric or pediatric(s), pain and distress in the ED were searched (Supplemental Table 5). Gray literature searches included the World Health Organization Clinical Trials Registry and Google Scholar.

Titles, abstracts, and eligible full-texts were independently screened against eligibility criteria by 3 authors (N.R., A.D., K.E.). Discrepancies were resolved through discussion. Reference lists of included studies were screened for additional studies and full-text versions of conference abstracts were obtained.

Three authors (N.R., A.D., K.E.) independently extracted study variables using a standardized data extraction form including: study location, participant numbers, setting, distraction intervention, usual care provided, study design, treatment procedure, pain and distress outcome variables, measurement tools, cost, and parent or healthcare provider satisfaction. Extraction was ratified by another author (S.H.) and conflicts were resolved through group discussion. A Royal Children's Hospital CCLS in Melbourne, Australia, classified the interventions as active, passive, or active and passive combined.

The primary outcome measures were self-reported pediatric pain and fear or anxiety. Although fear is defined as an alarm response to real or perceived danger, and anxiety as a mood-state orientated toward potentially negative future events,^22,23  differentiating between the 2 was beyond the scope of this study. Nonetheless, to retain transparency, where possible fear and anxiety are reported separately, however, given many studies have used the same scale to assess these constructs and few have distinguished between the 2,^24–26  most findings will be described narratively in aggregate as “distress.”

Self-reported pain was only provided in studies including children 3 years and over. Therefore, secondary outcomes included parental and/or researcher or clinician proxy-reported pain and distress scores to include younger children. Other secondary outcomes to investigate the suitability of various tools for the prehospital setting included parental anxiety, cost and parent and healthcare provider satisfaction. Further analysis included classifying the distractors as active or passive, and then as digital, nondigital, or environmental modifications.

Two authors (N.R., K.E.) independently performed quality assessments. The Cochrane Risk of Bias 2.0 (RoB 2.0) tool was used to assess randomized control trials (RCTs).²⁷  Nonrandomized control trials (non-RCTs) were assessed using the Cochrane Risk of Bias in Nonrandomized Studies of Intervention tool.²⁸  Any discrepancies were discussed and resolved.

Synthesis of findings was guided by the Synthesis Without Meta-Analysis (SWiM) reporting guidelines, given a meta-analysis was inhibited by poor study quality and heterogeneity in study design, population, distraction tools, measurement tools. and the timing of pain or distress assessment.^29,30  Therefore, in alignment with the SWiM guidelines, we have synthesized our findings using vote counting based on the direction of effect.^29,30 

Pain and distress scores were reported as either means with standard deviations or medians with interquartile ranges. When synthesizing findings, medians and interquartile ranges were converted into means and standard deviations using methods proposed by Wan et al.³¹  Three studies compared multiple treatment arms to a single control group, creating unit of analysis issues that were resolved by pooling across treatment groups in accordance with Cochrane recommendations to provide a single pairwise comparison for each study.³²  Treatment effects were summarized as Hedges’ standardized mean differences (SMD) with 95% confidence intervals (CI). Values of 0.2 to 0.5 are considered small, 0.5 to 0.8 are considered medium, and values >0.8 are considered large.³³ 

The I-squared statistic describes the percentage of variability in effect estimates that is caused by heterogeneity rather than chance.³⁴  An I-squared of 30% to 60% represents moderate; 50% to 90% substantial; and 75% to 100% considerable heterogeneity.³⁴  Preliminary analyses indicated considerable heterogeneity for postprocedural, self-reported child distress scores (I-squared = 88.5%). This was also observed in the subgroup analysis with the I-squared statistic for nondigital interventions being 91.3%. Under these circumstances alongside poor quality of evidence, a pooled treatment effect is not likely to be particularly robust, therefore child distress scores were synthesized narratively.

The search yielded 3122 nonduplicate titles, of which 142 full-text articles were assessed for eligibility following abstract screening (Supplemental Table 6). Of these, 24 studies met the eligibility criteria, and 5 further studies were identified from gray literature and bibliographic screening (Fig 1).

There were 22 RCTs^{16,24,25,35–53}  and 7 non-RCTs^26,54–59  with sample sizes between 15 to 216 children (Table 1). Ages ranged between 3 months to 18 years old and 1 study did not specify the age range.⁵⁶  Twenty-two studies reported their ED was a pediatric ED and 20 studies included participants undergoing needle-related procedures (i.e., blood sampling, venipuncture, intravenous cannulation, and injections). Most studies excluded critically ill children, those with cognitive impairment, or who did not speak the native study location language. Virtual reality was the most common distraction intervention, occurring in 6 studies.^{42,48–50,52,53}  Overall, there were 12 pain and 15 distress measurement tools used to collect ratings from children, parents and guardians, and clinicians and researchers (Tables 2 and 3).

Overall, the quality of evidence was poor (Table 4). Of the 22 RCTs (Supplemental Table 7), only 1 study was at low risk of bias,⁴⁴  13 studies were of some concerns,^{16,24,36,38–40,43,46,48,49,51,53}  and 8 at high risk.^{25,35,37,42,45,47,50,52}  Of the 7 non-RCTs (Supplemental Table 8),²⁹  only 2 studies had moderate bias risk,^57,58  2 were at serious risk,^55,56  and 3 at critical.^26,54,59  The majority of bias present in non-RCTs was because of a lack of true control.^26,54,58  Several studies did not balance nonprotocol and/or cointerventions between control and intervention arms.^{35,39,40,46,48,49,53,57–59}  Given the nature of the interventions in this review, blinding was limited and was therefore a dominant source of bias.^{16,25,35,36,38,40–43,45–50,52,53}  Two studies that recorded participant interactions blinded their assessors by adding music to recordings of both arms^39,52  and strategically positioning the camera to eliminate the assessors view of the virtual reality equipment.⁵²  Numerous additional factors may have biased the results, including reliance on subjective data types, variations or omissions in reasons for ED presentation, differences in diagnostic or medical procedures performed, and little accounting for previous exposure to the ED and/or medical procedures. Therefore, given the opportunities for bias, the results below should be interpreted with caution.

Twenty-three studies measured self-reported pain,^{16,25,35–46,48–50,53–56,59}  with 16 utilizing digital distractors,^{16,35,37–40,42,44,45,48–50,53–55,59}  7 using nondigital distractors,^{25,36,41,43,46,54,59}  and 1 utilizing environmental modification (Table 4).⁵⁶  Ten of these studies reported significant findings all favoring the intervention.^{36,40,41,43,46,48,54–56,59}  A pooled estimate of postprocedural pain scores demonstrated a small effect favoring the intervention (SMD = 0.40; 95% CI:0.19 to 0.61).^{16,25,35,37–39,41–43,48,50,53,55,56}  This estimate should be viewed with caution, given the I-squared statistic of 65.4% represents substantial heterogeneity across the studies.³⁴  The pooled effect for digital interventions was small and likewise favored the intervention group (SMD = 0.26; 95% CI:0.12–0.41) as did the pooled effect for nondigital interventions (SMD = 0.73; 95% CI:0.05 to 1.41). Although this effect size was moderate, it only included 3 studies, and the I-squared statistic (88.7%) indicated considerable heterogeneity across studies using nondigital techniques.

Among the digital distraction studies, 10 different distractors were used. As shown in Fig 2, only 2 studies significantly favored the intervention groups, with 1 study using animation with sing-along (SMD = 0.70; 95% CI:0.06 to 1.34)⁵⁵  and the other using virtual reality (SMD = 0.37, 95% CI:0.03 to 0.71).⁴⁸ 

Six of the 7 studies using nondigital distractors reported significant findings favoring the intervention.^{36,41,43,46,54,59}  Three of these studies (containing 7 intervention arms) were included in an unpooled estimate of effect and 2 studies (containing 6 intervention arms) significantly favored the intervention (Fig 2).^41,43  It was noted that participants in the only nondigital distraction study to report insignificant findings were undergoing a medical examination in the setting of “severe condition,”²⁵  whereas the children in studies citing significant findings in pain scores were undergoing needle-related procedures.

Finally, only 1 study implemented environmental modifications, which was an ambient environment of lighting displays in the waiting and patient care rooms.⁵⁶  This study identified a significant difference in pain postprocedural scores when compared with the control group (SMD = 0.61; 95% CI:0.12 to 1.10), which comprised patients attending another hospital’s pediatric ED.⁵⁶ 

Twelve studies measured this outcome,^{24–26,40,43,46,48–50,53,57,59}  6 used digital distractors,^{26,40,48–50,53}  6 trialed nondigital distractors,^{24,25,43,46,57,59}  and 1 contained an environmental modifications arm.⁵⁷  Overall, 9 studies reported significant findings for distress, all favoring the intervention.^{24–26,40,43,46,48,53,57}  Conflicting results were seen in 3 studies that measured anxiety and fear separately.^24–26  One reported reductions in both,²⁶  whereas the remaining 2 favored the intervention for either fear²⁵  or anxiety²⁴  in isolation.

Of the 4 studies that used virtual reality as the distractor, Chen (P < .05)⁴⁸  and Butt (P < .001)⁵³  reported significantly lower postprocedural distress scores in the intervention group as shown in Fig 3 (SMD = 0.35: 95% CI:0.01 to 0.69; SMD = 0.28: 95% CI:−0.09 to 0.66). The remaining digital distractor studies used videos and audiobooks, both recording significant findings favoring the intervention in child self-reported distress levels.^49,50 

Five of the 6 studies utilizing nondigital distractors produced significant findings favoring the intervention for self-reported distress.^{24,25,43,46,57}  Additionally, 1 study introduced environmental modifications in the form of cartoon-patterned clothing on nursing staff. This study reported a significant difference in postprocedural self-reported distress (Median 2.00, variant 1.29 versus Median 3.00, variant 0.89; P = .031).⁵⁷ 

Clinician and researcher assessment of pediatric pain was measured in 14 studies,^{16,35,38,41–44,47,50,52,55–58}  with 8 adopting digital distractors,^{16,35,38,42,48,50,52,55}  5 nondigital,^{41,43,46,56,57}  and 2 implementing environmental modifications.^56,57  Overall, 6 studies reported significant findings demonstrating reductions in pain favoring the intervention.

Only 1 of the 4 digital intervention studies using virtual reality produced significantly lower clinician and researcher-reported pain scores in the distraction group.⁴⁸  Conversely, all but 1 nondigital study⁵⁸  significantly favored the intervention for clinician and researcher-reported pain.^41,43,57  The study reporting insignificant findings included the youngest cohort with half of the participants being between 3 months and 2 years old.⁵⁸  They also did not incorporate a control arm as their aim was to investigate the feasibility of using distraction kits during needle-related procedures rather than the impact of distraction on pain or distress.⁵⁸ 

Adaptations to the environment were implemented in 2 studies.^57,58  Neither the ambient ED environment nor cartoon outfits worn by ED staff reported significant results for clinician and researcher-reported pain scores.

Eleven studies contained both clinician and researcher pain assessment and patient self-reported pain scores^{16,35,38,41–43,46,48,50,55,56} ; only 1 that used environmental distraction reported discordant findings between the proxy and self-reported pain assessments.⁵⁶ 

Parental-assessed patient pain was measured in 10 studies,^{16,35,36,38,41,43,46,48,49,52}  6 of which used digital distractors^{16,35,38,48,49,52}  and 4 nondigital.^36,41,43,46  No study measured the effect of environment modifications for this outcome. Overall, 7 of the 10 studies significantly favored the intervention,^{16,36,38,41,43,46,48}  whereas the remainder reported no significant difference between groups.

Only half of the digital studies reported significant reductions in parental-assessed pain.^16,38,48  Of the 3 studies utilizing virtual reality, 2 failed to produce significant findings according to parental assessment.^49,52  Like the clinician and researcher-reported pain scores, all the nondigital distractors significantly favored the intervention groups when parents assessed pain.^36,41,43,46 

Nine studies included both parent-reported and child self-reported pain scores.^{16,35,36,38,41,43,46,48,49}  Of these, 2 studies reported discordant findings.^16,38  Children reported no significant difference in pain, whereas parents significantly favored distraction.^16,38  Both of these studies included digital interventions and, in both cases, parents used a different self-reported pain assessment to provide a proxy rating of their child’s pain. In the remaining studies, parents used the same self-reporting tool as their child.

This outcome was measured in 11 studies, 7 reporting anxiety^{36,39,41,44–47}  and 4 fear.^24,43,48,51  Six studies used digital distractors^{39,44,45,47,48,51}  and 6 used nondigital.^{24,36,41,43,46,51}  No studies measured environmental modifications. Only 4 studies reported significant results, all of which favor the intervention.^36,43,46,48 

A significantly lower clinician and researcher-reported distress score was derived favoring distraction in 2 of the 6 studies using a digital distractor^48,51  and 4 of the 6 nondigital distractor studies.^36,43,46,51 

Four studies contained both clinician and researcher-reported child distress and child self-reported distress.^24,43,46,48  Both groups agreed in all studies with 3 reporting significant findings favoring the distraction tool used.^43,46,48 

Parental-reported patient distress was measured in 9 studies, 5 reporting anxiety^{41,46,47,49,52}  and 4 reporting fear.^24,43,48,51  Five used digital distractors,^{47–49,51,52}  5 nondigital,^{24,39,41,43,46,51}  and no study measured environmental modifications for this outcome. Half of these studies reported significant findings favoring the use of distraction.^{43,44,46,48,51} 

Two studies using digital distractors reported significantly lower parental-reported patient distress score in the intervention groups.^48,51  The parental-reported patient distress scores in 3 of the 5 studies involving a nondigital distractor also favored the intervention group.^43,46,51  Five studies included both a parent-reported proxy distress score and a child self-reported distress score. All studies produced concordant findings between parents and their children.^{24,43,48,49,51} 

Six studies measured parental anxiety.^{39,44,47,52,56,59}  Four had children use digital distractors^39,44,47,52  and 2 of these significantly reduced parental anxiety.^44,47  One study using nondigital distractors⁵⁹  and another comparing ambient lighting in a pediatric ED to a standard pediatric ED also significantly reduced parental anxiety.⁵⁶ 

Across the studies investigating self-reported outcomes, 7 used active distraction tools, 6 of which were nondigital,^{25,41,43,51,54,57}  (1 of which included an environmental distraction arm)⁵⁷  and 1 was digital.³⁵  Pain was investigated in 6 of the 7 active distraction studies,^{25,35,41,43,51,54}  4 utilizing nondigital distractors significantly favored the intervention.^41,43,51,54  Self-reported distress was measured in 4 nondigital distraction studies.^25,43,51,57  All favored the intervention, however 1 study that measured both anxiety and fear reported conflicting findings for these outcomes.²⁵ 

Passive distractors were used in 14 studies, of which 7 were digital,^{29,37–40,51,54,55}  6 were nondigital,^{27,36,41,43,51,59}  and 1 was environmental.⁵⁶  Twelve studies investigated self-reported pain, with all 5 of the nondigital studies^{36,41,43,51,59}  and the environmental study favoring the intervention.⁵⁶  Three of the 6 studies using digital distractors favored the intervention,^40,54,55  whereas the remainder found no significant difference between the digital distraction and control arm.^37–39  Four passive distractor studies using nondigital interventions^24,43,51,59  and 2 using digital interventions^26,40  investigated self-reported distress levels. Only 2 nondigital studies reported no significant difference in fear,^24,59  however 1 of these noted a reduction in anxiety.²⁴  The remaining studies significantly favored the intervention.

Finally, 9 studies involved a combination of active and passive distractors.^{16,42,44,45,48–50,53,59}  The distractors in 8 of these studies were digital, and the final study used various distractors that included both digital and nondigital options.⁵⁹  All 9 studies investigated self-reported pain and only 2 reported significant findings favoring the interventions.^48,59  Five studies investigated distress,^{48–50,53,59}  and again only 2 significantly favored the intervention.^48,53 

This systematic review identified 29 studies exploring pediatric pain and/or distress distraction techniques used in the ED that would be suitable for the prehospital setting. Our findings are consistent with previous studies across multiple pediatric settings,^12,60,61  demonstrating distraction therapy to be beneficial in managing pediatric pain and/or distress. Importantly, no significant negative effects were reported, however findings should be interpreted carefully given the low quality of evidence and small self-reported pain effect sizes. Ultimately, further research is required to determine the clinical value of distraction in the field. Some aspects of the studies (e.g., needle-related procedures) are reflected in the prehospital setting, although other aspects, i.e., critically unwell children and transportation in a moving vehicle, mean the generalizability of these results to the prehospital setting is limited.

Overall, nondigital distractors more consistently resulted in significant improvements in pain and/or distress compared with digital distractors. This finding supports a recent meta-analysis that identified no obvious advantage using digital over nondigital distractors in pediatric pain management.⁶¹  This study also found digital distractors were potentially more effective in anxiety management,⁶¹  and although we had too few studies to confirm this, our results appear to support this. The reason why nondigital distractors were superior is unclear; it may be that nondigital distractors engaged the children in more ways, such as through cognitive and multiple sensory modalities (eg, visual, auditory, and tactile stimulation).⁶²  A review of the sensory modalities of distraction stated that pain is a stimulus consuming an individual’s finite attentional resources.⁶²  Therefore, by consuming these resources with distraction rather than pain, the pain should be reduced.⁶²  Although beyond the scope of this review to rationalize why nondigital distractors were more effective at reducing pain, it could be postulated that these distractors consumed more attentional resources in children than digital distractors. Further research into the effectiveness of distractors should consider the sensory and cognitive demands. it could be postulated that these distractors consumed more attentional resources in children than digital distractors. Further research into the effectiveness of distractors should consider the sensory and cognitive demands.

A recently updated meta-analysis identified moderate agreement between child and parent-rated pain, but weaker agreement between child and clinician-rated pain.⁶³  Although the present review did not measure the magnitude of difference between child and proxy-rated assessment, it does appear to support this agreement for both digital or nondigital distraction.

Environmental distractors were investigated in too few studies to synthesize any strong evidence, however the importance of a low-stimulus emergency care environment for children should not be underestimated.⁶⁴  Hospital-based studies in adults have shown that music, lighting level, natural light, views of nature, and noise can all affect distress and pain.^65–68  These environmental factors can be manipulated in both the ED and the prehospital environment and should be considered for their potential contribution to distraction in future research.

Few studies identified whether their distractors were active, passive, or a combination of both. Like previous research, the included studies did not definitively favor active or passive distraction,⁶¹  however generally those using a combination of active and passive distraction performed poorly.^{16,42,44,45,47,49,50,52}  The effectiveness of passive versus active distraction may be relative to the group under study,^61,69–72  e.g., developmental differences in selective attention skills have highlighted passive distraction as more effective in preschool aged children,⁷¹  and this may explain the lack of significant findings supporting either classification. Despite this, only 4 studies stratified for age according to which pain or distress assessment tool was suitable.^37,40,56,58  Furthermore, often no consideration was given to distraction tool suitability, evidenced by the fact many studies used only 1 distraction tool.^{24–26,35,36,38–40,42,44–47,49,50,52,55,56}  Thus, future research should also consider developmental age in terms of the distraction tools used, their type (active, passive, or both), and not only the suitability of pain and distress measurement tools.

Heterogeneity was seen in pain and/or distress measurement tool usage, time-points at which scores were taken and timing proximity of pain and/or distress measurement with noxious stimuli. Some studies also used self-reported pain tools to make proxy pain assessments,^{16,35,36,38,41,43,46,48,49,55}  which has been reported to result in an underestimation of the child’s pain and severity.^72–76  As mentioned earlier, this study did not ascertain the magnitude of difference between self- and proxy-ratings. However, only 3 of 12 studies obtaining both self- and proxy-ratings reported a disagreement and all were for pain.^16,38,56  Twelve studies included both clinician and researcher and parent-rated proxy pain and/or distress scores for children; only 2 studies were discordant with parents noting a reduction in pain where clinician and researchers did not.^16,38  Although our study did not specifically seek to investigate the alignment of self- and proxy-rated pain assessment, our findings do not support those suggesting an underestimation of pain.

When conducting further research, consideration should be given to factors that potentially altered the effectiveness of these tools in the ED. This includes age or cognitive development, presenting problem and severity, nonprotocol interventions like analgesia and the nature of distraction tools themselves, including whether they are digital, nondigital, or environmental distractors and active, passive, or a combination of both. The consistent use of validated and reliable pain or distress assessment tools should occur, and they should be applied in the way they were intended and on the population for whom they were developed. Qualitative studies investigating potential barriers to the implementation of basic distraction techniques may also contribute to the body of knowledge around the effectiveness of distraction techniques for pediatric patients in the ED. Research to determine the feasibility, efficacy, and impact of these distraction tools in the prehospital environment and how this may vary from other settings is also required. Finally, more consistency in the methodological application and reporting of findings in future studies would also benefit the synthesis of high certainty evidence, allowing for strong recommendations for translation into practice and the development of strategies for the prehospital setting.

Despite a broad search, it is possible that some studies were missed. Similarly, notwithstanding the presence of 22 RCTs, the quality of evidence was poor and this may have influenced the robustness of the pooled effect. One of the main drivers of poor study quality was the fact that blinding participants to their intervention was not possible. Methodologically, the rigor of the included studies varied and their methodological heterogeneity meant definitive findings regarding the evidence was difficult and meta-analysis was impossible. Our inclusion criteria were limited to distraction tools that would be suitable for the prehospital environment. Therefore, our findings are partially limited when attempting to generalize to all settings. In-depth consideration of injury and condition and/or procedures as well as factors unique to the prehospital setting, such as patient management in a moving vehicle and motion sickness, have not been included. Finally, too few studies included an investigation of cost or parent and healthcare provider satisfaction to meaningfully contribution to the study aim, therefore these outcomes were not included in this paper.

Despite high heterogeneity and low quality of evidence, a range of distraction tools exist that effectively reduce pain and/or distress in children in the ED. It is possible some of these distraction tools could be adapted to the prehospital environment, however further research is required to determine which would be most feasible and effective.

We thank Anne-Maree Leahy (Certified Child Life Therapist; Bachelor of Behavioral Science, Diploma of Primary Education; Graduate Diploma of Psychology) at the Royal Children’s Hospital in Melbourne, Australia, for her work in classifying the distraction techniques and providing her ongoing expertise and advice throughout this review.

CCLS

certified child life specialists

ED

emergency department

non-RCTs

nonrandomized control trials

RCTs

randomized-control trials

RoB 2.0

Cochrane Risk of Bias 2.0

SWiM

Synthesis Without Meta-Analysis reporting guidelines